AFT Impulse Instructions

AFT Impulse Seminar
Week of 22 May, 2017

Introduction
Overview of Seminar
INTRODUCTION
I1. Overview of AFT Impulse
I2. Steady-State Hydraulic Solution Methodology
I3. Waterhammer Solution Methodology
I4. Demonstration Problem - Surge at Valve Closure
I5. AFT Impulse Hands - On Modeling
I6. Pipe and Junction Details
I7. The Five Primary Windows
I8. Time and Event Transients
I9. Steady and Transient Special Conditions
I10. Pump Transients
I11. Pipe Sectioning and Transient Control
I12. Special Topics Including Troubleshooting
I13. Surge Suppression
I14. Calculating Unbalanced Forces
I15. Using the Scenario Manager
I16. Customizing AFT Impulse and Using Databases
I17. Introduction to AFT Impulse SSL Module
I18. More AFT Impulse Hands-On Modeling
About Applied Flow Technology
 Applied Flow Technology (AFT), founded in 1993, is a world

leader in providing high quality software to analyze flows,
pressures and transients in systems with pipes, pumps and
valves
 Customers in 70+ countries
 Representatives in 32 locations around the world
Intro - 1
AFT Fathom™ 9
 Models incompressible network pipe systems

– Liquid and low velocity gas systems
 Models open and closed systems
 Models systems that are pressure, gravity or pump driven
 Models heat transfer and system energy balance
 Offers broad range of innovative reporting features
– Printed output is of report quality
 Offers customizable component and property databases
– Cost calculations
– Rheological data handling to support non-Newtonian fluids
Intro - 2
AFT Fathom™ Add-On Modules
 XTS – eXtended Time Simulation

– Simulate dynamic behavior of systems over time
– Models infinite and open and closed finite tanks of constant and
varying cross section
– Supports user defined time and event transients of pumps, valves
and other components
 GSC – Goal Seek & Control
– Automatically determines input variables that will yield specified
output values
– Extends Fathom’s control simulation capabilities to include remote
sensing
 SSL – Settling Slurry simulation
– Simulates settling slurry behavior
– Simulates pump performance degradation
Intro - 3
AFT Arrow™ 6
 Models compressible network pipe systems

– High to low velocity gas systems
– High to low pressures
 Models open and closed systems
 Accurately models
– Real gases
– Heat transfer
– Highly compressible (sonic and near sonic) systems
 Balances flow and energy throughout the system
 Offers broad range of innovative reporting features
 Offers customizable component and property databases
 Includes high accuracy steam/water properties to ASME
Intro - 4
AFT Arrow™ Add-On Module
 GSC – Goal Seek & Control

– Automatically determines input variables that will yield specified
output values
– Extends Arrow’s control simulation capabilities to include remote
sensing
Intro - 5
AFT Impulse™ 6
 Models waterhammer/surge flow in pipe networks

 Models system transients caused by
– Sudden valve closures
– Pump startups and shutdowns including pump inertia effects
– Relief valve cracking
– Events defined within the system (e.g. flow, pressure, etc.)
 Includes modeling of
– Control and relief valves, vacuum breaker valves, check valves,
pumps, accumulators, turbines, and surge tanks
 Includes a steady-state solver to determine initial conditions
 Calculates unbalanced transient forces
– Forces can be graphed or exported as Force/Time data files
 Can also import AFT Fathom models
Intro - 6
AFT Impulse™ Add-On Module
 SSL – Settling Slurry simulation

– Simulates settling slurry behavior
– Simulates pump performance degradation
Intro - 7
AFT Mercury™ 7
AFT Titan™ 4
 Models and designs network pipe systems
 Combines a powerful hydraulic solver and flexible graphical
interface with an advanced optimization engine
– Automatically selects best pipe and component sizes to minimize
initial or life cycle cost, size or weight using IntelliFlow®
 Ability to apply multiple constraints to pipes and junctions
 Cost optimization may include;
– non-recurring costs (materials and installation)
– recurring costs (energy and maintenance) including time varying
cost (energy costs varying with time)
 Offers customizable engineering and cost databases
 Includes powerful modeling and output capabilities
of AFT Fathom 7 and Arrow 4
Intro - 8
Additional Software Products
 Chempak™ Property Database

– Property database of ~700 fluids
– Ability to define static pre-mixtures
– Dynamic mixing capability in Arrow
 Chempak Viewer™ 2.0 & Chempak Add-in™ (for Excel)
– Viewer allows use of Chempak as a stand alone application
– Add-in makes all of the Chempak functions accessible within an
Excel spreadsheet
 SteamCalc™ 2.0
– High accuracy ASME steam/water library for Windows and Excel
Intro - 9
Product Applications
 AFT products are being successfully applied to a broad range

of industrial systems:
– Power generation systems
– Chemical and petrochemical systems
– Oil and gas production, transportation, refining and delivery
– Automotive systems
– Aerospace systems
– Air conditioning and refrigeration systems
– Pulp and paper processing
– Fire suppression
– Water and Wastewater treatment plant design
– Mining processing and support systems
– Municipal water distribution
Intro - 10
AFT Flow Expert Package™
 Provides consulting services beyond typical technical support

requests on the installation, upgrade assistance, and
functionality of AFT software.
 Access to a consulting engineer assigned as your primary
point of contact.
 Package Options: Blocks of 5 hours, 10 hours and 20 hours
 Typical ways to use your hours:
– Receive online training on specific topics of your choice
– Request help on model results interpretation
– Get a second opinion of your assumptions, modeling choices
and reports
Intro - 11
AFT Flow Expert Package™ (2)
 Additional ways to use your hours:

– Have an expert double check your modeling input and point out
common modeling mistakes or suggest better ways to model the
desired behavior
– Receive guidance in how to model pumps and pump-system
interaction, relief valves and relief systems, surge suppression
equipment, slurry pipelines, system transients, and anything
having to do with flow in pipe systems
– Discuss with an expert alternative solutions for hydraulic
problems
– Help launch AFT software within your company and reduce your
learning curve
– Help new hires get acquainted with AFT software
Intro - 12
Overview of Seminar
INTRODUCTION
Nomenclature
a = wavespeed I = inertia
A = cross-sectional area J = Jacobian matrix
B = impedance K = bulk modulus for liquid
BM = parameter as defined in Section 3 K = loss factor
BP = parameter as defined in Section 3 K = constant for Power Law fluid
CD = discharge coefficient K´ = rigidity of tank
CM = parameter as defined in Section 3 Ksprinkler = fire sprinkler loss factor
CP = parameter as defined in Section 3 L = length
CV = valve coefficient m
 = mass flowrate
CHW = Hazen-Williams factor n = polytropic constant
D = diameter n = number of pipe sections
E = modulus of elasticity n = constant for Power Law fluid
EGL = energy gradeline N = pump speed in rpm
e = pipe wall thickness Ns = specific speed
f = friction factor P = pressure, static
F = error value in steady-state mass balance P = power
g = gravitational acceleration Po = pressure, stagnation
H = head
HGL = hydraulic gradeline
Nomenclature (2)
Q = volumetric flow rate α = angle (slope of pipe)
r = radius ε = roughness
r = relaxation λ = method of characteristics multiplier
R = resistance µ = Poisson's ratio
Rc = coefficient of rigidity µ, θ = angle
Re = Reynolds number ρ = density
s = pump speed ω = rotational speed
SB = parameter as defined in Section 3
Subscripts
SC = parameter as defined in Section 3
exit = exit point from pipe system or junction
Sy = yield stress
i = computing location in pipe
t = time
j = junction
T = torque
new = current time for computation
V = velocity
old = most recent time for computation
V = volume
x = distance
z = elevation
AFT Impulse General Description
 General purpose pipe network waterhammer and surge

analysis
 Drag-and-Drop graphical interface
 Simulates steady and transient behavior in liquid pipe
networks
 For steady flow uses Newton-Raphson matrix techniques to
solve 2 equations: Continuity (Mass) Equation and
Momentum (Bernoulli) Equation
 For transient flow uses explicit Method of Characteristics
technique to solve the 2 equations for each pipe: Continuity
(Mass) Equation and Momentum Equation
I1 - 1
AFT Impulse General Description (2)
 Can model systems in any generalized configuration

– Open or closed systems
– Branching systems
– Looping systems
 Can model non-Newtonian fluids using Power Law and
Bingham Plastic
 Can model variable fluid properties
 English and SI units supported
I1 - 2
Components That Can Be Modeled
 Branching section (up to 25 pipes)

– Can model transient flow source or sink
 Known pressure or flow boundaries
– Conditions can change with time
 Pumps
– Pump speed can change with time
– Pump inertial effects and four-quadrant curves
– Centrifugal pumps and positive displacement pumps
– Pumps with flow controllers
 Pressure and flow control valves
 Valves
– Valve position can change with time
I1 - 3
Components That Can Be Modeled (2)
 Relief valves
– Three different valve types – Internal, Exit, and Inline Exit
– Can model different opening and closing profiles (i.e. time or
pressure)
– Can model overpressure, and blowdown pressure in addition to
opening set pressure
– Rate limits can be applied to opening or closing pressure profiles
 Check valves
– Model Cv vs. Time profiles to avoid instantaneous opening and
closing of check valves
 Gas accumulators, liquid accumulators, surge tanks and
vacuum breaker valves
 Francis hydroelectric turbines
I1 - 4
Engineering Limitations
 No practical software limit to model size

 Flow is liquid, liquid-full and one-dimensional
 Wavespeed stays constant during transient
 Non-condensable gas (e.g. air) release is negligible
 No limit on number of pumps, valves, etc.
 No limit on number of custom components, fluids or pipe
materials
I1 - 5
Impulse 6 Startup Window
I1 - 6
Primary Windows
 The AFT Impulse modeling process flows through five Primary

Windows
– Workspace
– Model Data
– Output
– Visual Report
– Graph Results
 The Primary Windows offer a mixture and graphical and text-
based features to assist in the modeling process
 Tabbed Primary Windows allow for easier navigation
 Robust usage of dual monitors is supported
– Can drag the Primary Window tabs into their own
separate window
I1 - 7
Primary Window Process Flow
Model Data
Visual Report
Workspace Output
Graph Results
I1 - 8
Workspace: Toolbox
This tool is used to
draw new pipes
This tool will add
These icons annotation to the
represent workspace
different
components
I1 - 9
Quick Access Panel – Activate Modules
 Ability to activate SSL Module
I1 - 10
Workspace - Editing Features
 Cut, copy, paste, delete, duplicate and undo features

supported
– Only one level of undo is supported
 Workspace can be sized to fit the model
 You can zoom out to see a larger area
 Objects can be selected as a group in several ways
– Selecting the components by dragging the mouse over them
– Using the SHIFT key while clicking on the objects
– Using Select Flow Path on the Edit menu
– Using the Select Special tool on the Edit menu
– Using Groups / Select on the Edit menu
– Using the Select All feature
I1 - 11
Workspace - Editing Features (2)
 The Reference Flow Direction of a pipe can be changed

 The selected objects can be renumbered in four ways:
– Manually
– Renumber Automatic
– Renumber Wizard
– Renumber Increment
 The Find tool will move the Workspace window to show a pipe
or junction
I1 - 12
Workspace - Platform for Data Entry
 All pipe and junction objects placed onto the Workspace are
interactive
 To open the Properties window for data entry, just double-click
the graphical object
– Alternatively, you can select the object by clicking on it once and
then press the Enter key
– Or you can select the object by clicking on it once and then click
on the Open Pipe/Jct Window button on the Workspace Toolbar
 The Properties windows are the primary manner in which
component data is entered
 The Global Pipe Edit and Global Junction Edit windows can
speed up data entry
I1 - 13
Workspace - Reporting
 The Workspace image can be printed on printers and plotters

 Print Preview allows page customization
– The image can be sized on the page
– A company logo and custom text can be added
I1 - 14
Model Data Window
 Model Data is broken into three sections

– General Data
– Pipe Data
– Junction Data
 Each section can be re-sized or collapsed allowing the user to
focus on any of the sections
 User can select all or portions of the Model Data Window
content for printing
– Print format window allows customizing of content
– User can also select the font
I1 - 15
Model Data Window (2)
 With a Workspace printout and the complete Model Data

printout, the input can be printed in its entirety
 Properties windows for data entry can be opened by double-
clicking the far left column
I1 - 16
Output Window
 The Output window is the primary vehicle for communicating

the results of an analysis in text form
 Output Window is broken into three sections
– General Results
– Pipe Results
– Junction Results
 Each section contains tabs to permit quick viewing of output
by type
 Transient event messages are separated into two lists
– Sorted by junction
– Sorted by time
I1 - 17
Output Window (2)
 User can select all or portions of the Output Window content

for printing
– Print format window allows customizing of content
– User can also select the font
 User can sort output according to any of the columns for quick
review of data extreme maximums and minimums
 Output Window content is specified by Output Control
Window
I1 - 18
Visual Report Window
 Visual Report allows user to display input and output results

together with pipe system image
I1 - 19
Graph Results Window
 The Graph Results Window allows creation of full-featured

Windows graphs
I1 - 20
I2. Steady-State Hydraulic Solution
Methodology
Overview of Seminar
INTRODUCTION
Introduction
 AFT Impulse uses the Newton-Raphson Method to solve the

steady-state flow distribution in a pipe network
– The Newton-Raphson Method for pipe networks is a matrix
method
 This method gained favor with the introduction of the digital
computer
 The technique has been considered standard industry
practice for 40 years
I2 - 1
Pressure Definitions
 Conflicting definitions exist in the literature for total and

stagnation pressures
 AFT Impulse terminology
1
P + ρV 2 + ρgh = Const
2
}
}
}
Static Hydrostatic
Dynamic
(or Velocity)
 Ptotal = Pstagnation = Pstatic + Pdynamic
 AFT Impulse uses the term Pstagnation and not Ptotal
 Input pressures in Impulse, such as assigned pressure
junction, assume Pstagnation as input
I2 - 2
Basic Laws of Pipe Flow
 Mass Conservation
m = ρAV
 Momentum Equation (Bernoulli)

1 1
P1 + ρV12 + ρgh1 = P2 + ρV22 + ρgh2 + ∆Ploss
2 2
 The dynamic pressure and static pressure can be combined

into the stagnation (total) pressure, and the solution is then for
total pressure
– Therefore, the momentum equation becomes
Po ,1 + ρgh1 = Po , 2 + ρgh2 + ∆Ploss
I2 - 3
Law of Friction
 Traditional method of friction loss calculation uses the Darcy-

Weisbach friction factor, f
L 1 2
∆Ploss = f  ρV 
D 2 
 The friction factor is not a constant, but a function of the pipe
wall characteristics and the Reynolds number
 AFT Impulse uses the iterative Colebrook-White correlation
for turbulent flow and the traditional laminar flow equation
when laminar −2
 ε 
f = 1.14 − 2 log +
9.35
 D Re f


(Re > 4000)
  
f =
64 (Re < 2300)
Re
 Special friction models available for pulp and
paper stock and crude oil I2 - 4
Modified Form for Law of Friction
 Basic law
L 1 2
∆Ploss = f  ρV 
D 2 
 Substituting mass flow rate definition
2
L  1  m 

∆Ploss = f ρ 
D  2  ρA  

 Defining new term, where R is a pipe resistance
2
∆Ploss = Rm
L 1 
R= f  
D  2 ρA 2 
 Bernoulli’s equation then becomes

2
Po,1 + ρ gh1 = Po,2 + ρ gh2 + Rm
I2 - 5
Balancing Mass at Branches
 Applying law of mass conservation to a branching section

n
∑ m ij = 0
j =1
 Substituting yields the following equation to be solved for

every branch, i 0.5
n (
 P − P + ρg h − h
 o, j ) 
∑ sgn(Po, j − Po,i + ρg (h j − hi ))
o,i j i
  =0
 Rij 
j =1  
 where sgn = ±1 depending on flow direction

 The objective is to find all of the P values that satisfy the
above equation applied to every branch
 We will then have a solution for two unknowns:
– pressure at all junctions
– mass flow rate in all pipes I2 - 6
Solving the Equations
 We need to solve as many equations as there are flow splits

 All of the equations are non-linear
 AFT Impulse uses the Newton-Raphson Method to solve the
system of equations
– Newton-Raphson is an iterative method used to solve for roots of
equations
 Initially the pipe flow rates are not known so an error, F, exists
at each branch 0.5
n (
 P − P + ρg h − h
 o, j ) 
o,i j i
  = Fi
 Rij 
j =1  
 The objective is to use the Newton-Raphson Method to drive
all of the F errors to zero (within some tolerance)
I2 - 7
The Newton-Raphson Method
 The procedure for applying Newton-Raphson to a single

equation is as follows
1. Take a guess at the solution to function F
2. Calculate an improved guess using the following equation:
F (xi )
xi + 1 = xi −
F ′(xi )
3. Substitute the improved guess back into the above

equation until the change in x is small
F(x)
F(xi)
-F'(xi)
x
xi xi+1 I2 - 8
Solving the System
 When applied to a system of equations with P as the

unknown, Newton-Raphson looks as follows
  
Po,new = Po,old − J F −1F
where P is the vector of pressures and JF is the Jacobian matrix of error

function derivatives - both of a size, n, which is the number of branches
(i.e., equations in the system)
 ∂ F1 ∂ F1 ∂ F1 
 ∂P ∂P 
o,1 o, 2 ∂ Po,n 
 
 ∂ F2 ∂ F2 ∂ F2 
  
J F =  ∂Po,1 ∂ Po,2 ∂Po,n 
 
  
 ∂F n ∂F n ∂F n 
  
∂ P
 o,1 o,2∂ P ∂ Po,n 
I2 - 9
Derivative Terms in Jacobian
 The diagonal derivative terms in the Jacobian can be

calculated analytically
n (
 P − P + ρg h − h
 o, j ) 0.5
o,i j i
Fi =  
 Rij 
j =1  
n  
∂F i
= ∑  − 0.5 ( −
 0.5  Po, j Po,i
(
+ ρg h j − hi )) −0.5
∂Pi R
j =1  ij 
 The off-diagonal terms can also be calculated analytically
∂F i  0.5 
= (P − P
∂P j  R 0.5  o, j o,i
+ ρg (
h j − hi )
) −0.5
 ij 
I2 - 10
Solving the Matrix
 Rather than inverting the Jacobian matrix, it is usually faster

to solve a linear system of equations as follows
  
Po,new = Po,old − z
 −1 
z = JF F
 
JF z = F
 We need to solve for the values in vector, z, that satisfy the

above
 Use Gaussian Elimination to solve for z
– By multiple substitutions, we progressively eliminate terms in JF and
eventually obtain the identity matrix, where all terms are zero except the
diagonal, which is unity
– We then have the solution for z, which can be substituted back into the
original equation at the top to improve our guess for all
of the pressures in the pressure vector
I2 - 11
Test Problem #1
1 2
P = 200 psig pipe 1 pipe 2 P = 175 psig
h = 0 feet h = 0 feet
4
h = 0 feet
pipe 3
pipe f L (ft) D (in) Fluid

P = 160 psig
h = 0 feet
3 1 0.0219 100 4 Water @ 70F
2 0.0156 75 4 Water @ 70F
3 0.0180 125 6 Water @ 70F
In this test problem, pipe resistances Jct P (psig) Steel - ANSI pipe, STD schedules
can be calculated based on known 1 200
friction factor (shown in the table) 2 175
3 160
I2 - 12
Test Problem #1 (2)
To start the solution, we need to guess P4, so guess 180 psig
0.5
n  P j − Pi 
[ (
F = ∑ sgn P j − Pi
j =1
)] 
 Rij




0.5 0.5 0.5
 P −P   P − P4   P −P 
F = [sgn(P1 − P4 )] 1 4  + [sgn(P2 − P4 )]  2  +[sgn(P 3 − P4 )] 3 4 
 R1− 4   R2 − 4   R 3− 4 
F = -277.1489
− 0.5 0.5 − 0.5 0.5 − 0.5 0.5

F′= P1 − P4 − + P2 − P4 − + P3 − P4 −
R1− 4 0.5 R2 − 4 0.5 R3− 4 0.5
F ′ = -18.6139
F (Pold )
Pnew = Pold −
F ′(Pold )
I2 - 13
Note: All pressures here are stagnation
Test Problem #1 (3)
We then use Newton-Raphson to get an improved value for P4
F (Pold )
Pnew = Pold −
F ′(Pold )
P (psia), M (lbm/s) Solution shown in Test1.xls

Iteration Pj=4 Mpipe=1 Mpipe=2 Mpipe=3 F F'
1 180.0000 115.3464 -78.9048 -313.5906 -277.1489 -18.6139
2 165.1106 152.3476 110.9691 -158.5207 104.7960 -23.3027
3 169.6078 142.1903 81.9409 -217.3503 6.7810 -21.2485
4 169.9269 141.4418 79.4792 -220.9305 -0.0095 -21.3129
5 169.9265 141.4429 79.4827 -220.9255 0.0000 -21.3128
6 169.9265 141.4429 79.4827 -220.9255 0.0000 -21.3128
7 169.9265 141.4429 79.4827 -220.9255 0.0000 -21.3128
I2 - 14
Note: All pressures here are stagnation
Convergence
 When the change in pressures and flow rates decreases to

some small amount, the calculation is converged
 Different criteria can be applied for identifying convergence
– Percentage change in result
– Absolute change in result
 We will cover convergence in a later section
I2 - 15
Flow Rate Updates
 The previous example focused on the pressure solution

 In practice, after the pressure solution was obtained a new
flow solution would be calculated
 The new flows would then be compared against the old flows
 If the flows changed too much they would be updated and the
pressure solution repeated
 This whole procedure would be repeated until flow and
pressure updates were small
I2 - 16
Steady-State Solver Flow Chart
Update Hydraulic
Start Losses and Pumps
Solve Junction
Pressures
No
No Yes
Converged ? > Max Iterations ? End
Yes
Recalculate Mass
Flow Rates
No No
Converged ? > Max Iterations ?
Yes Yes
Return End
I2 - 17
Known Flow Vs. Known Pressure
Junctions
 At all system boundaries AFT Impulse must solve for either
flow or pressure
 User cannot specify both flow and pressure at the same point
because there would be nothing for AFT Impulse to solve
 Either the flow rate calculation or the pressure calculation
must be available to AFT Impulse
I2 - 18
Overview of Seminar
INTRODUCTION
Overview of Waterhammer
 Waterhammer is a transient phenomenon that occurs in a

liquid piping system when some event causes a departure
from steady state
 Waterhammer is the process the piping system experiences
as it adjusts to the new conditions
 Waterhammer can be caused by many events including
– Valve closure or opening (in full or in part)
– Pump speed change
– Relief valve cracking open
– Tank pressurization
– Periodic pressure or flow conditions
I3 - 1
Terminology
 Waterhammer can occur in any liquid piping system

 The term “waterhammer” confuses some because it implies a
process only in water systems
 Other terms which have been used are
– Fluidhammer
– Hydraulic Transients
– Fluid Transients
– Surge
I3 - 2
Types of Waterhammer
 Waterhammer can be caused by different physical

mechanisms
– There is no universal terminology for these mechanisms so the terminology here
is for discussion purposes
1. “Thermodynamic” waterhammer
– Liquid acceleration caused by local phase change
2. “Slug” waterhammer
– Liquid flows into an evacuated pipe system or when there are distinct liquid slugs
and gas pockets
– When liquid contacts equipment or direction changes (elbows) pressure spikes
can occur
3. “Mechanical” waterhammer
– Caused by equipment or component operational changes
• Pump trips, valves closed, etc.
– This is the type of waterhammer that Impulse can model
I3 - 3
Instantaneous Waterhammer
 The magnitude of a waterhammer transient is dependent on

the wavespeed of the liquid
 The wavespeed (a) is dependent on the:
– liquid acoustic velocity
– liquid density & liquid modulus of elasticity
– pipe modulus of elasticity (E), wall thickness (t), and material
Poisson Ratio (μ)
– pipe restraints
I3 - 4
US
 It can be shown that the maximum theoretical pressure surge

is given by the instantaneous waterhammer equation
P    aV
– Example: Density is 62 lbm/ft3, Wavespeed is 2500 feet per

second, Initial Velocity is 10 feet per second, Static Pressure is
50 psia.
2 2
lbm ft ft 1 ft lbf  s
Pmax  62 3 * 2500 * 10 * 2 *  50 psia
ft s s 144 in 32.2 lbm ‐ ft
Pmax  334  50
Pmax  384 psia
I3 - 5
SI
 It can be shown that the maximum theoretical pressure surge

is given by the instantaneous waterhammer equation
P    aV
– Example: Density is 1000 kg/m3, Wavespeed is 1000 meters per

second, Initial Velocity is 3 meters per second, Static Pressure is
1000 kPa.
kg m m
Pmax = 1000 *1000 *3 + 1000 kPa
m3 s s
Pmax  3000 kPa+1000 kPa
Pmax  4000 kPa
I3 - 6
Communication Time
 The communication time is a helpful concept in understanding

waterhammer
 The communication time is the time it takes for transient
events to communicate their existence to boundaries in the
piping system
 The communication time is given by the following
L
t  2
a
 Any event that occurs in a time frame less than the

communication time is in effect instantaneous
I3 - 7
Waterhammer Sequence
V=Vsteady V= Vsteady
V=0 V=0
a a
a c
b d
V= Vsteady V= Vsteady
V=0 V=0
a a
I3 - 8
Waterhammer Sequence - 0 < t < L/a
Valve closed
instantaneously at t=0
V=Vsteady
V=0
a
P
Pinstantaneous
Psteady
x
V
Vsteady
x
I3 - 9
Waterhammer Sequence - L/a < t < 2L/a
Valve closed
V=Vsteady
V=0
a
P
 Pinstantaneous
Psteady
x
V
x
‐Vsteady
I3 - 10
Waterhammer Sequence - 2L/a < t < 3L/a
Valve closed
V=Vsteady
V=0
P
Psteady
 Pinstantaneous
x
V
x
‐Vsteady
I3 - 11
Waterhammer Sequence - 3L/a < t < 4L/a
Valve closed
V=Vsteady
V=0
P
Psteady
 Pinstantaneous
x
V
Vsteady
x
I3 - 12
Waterhammer Videos
 Instructor – show video files

– Waterhammer Sequence.wmv (1:06)
• Video of waterhammer sequence
• Waterhammer Sequence.imp
– waterhammer.wmv (2:03)
• Video of piping system
• online - http://www.youtube.com/watch?v=oZbguheiVs4&NR=1
– Waterhammer 4-0ms-Bruecke.wmv (1:04)
• Video of piping system
• online - http://www.youtube.com/watch?v=U0XfYCKxZks
– CFD Simulation of the Water Hammer.wmv (2:39)
• Computer simulation showing pressure waves with color
• online - http://www.youtube.com/watch?v=ng__hdZ8yD8
I3 - 13
Fundamental Equations
 Mass / continuity equation

 
a 2 V  P  0
 x t
 Momentum equation
1P V fV V
  g sin(  )  0
 x t 2D
Where :
a = wavespeed
V = velocity
x = distance along pipe
P = pressure
t = time
g = gravitational constant
 = slope of pipe
f = friction factor
D = diameter of pipe
I3 - 14
Method of Characteristics
t = 5 t
t = 4 t
 By combining the mass and
momentum equations linearly and t = 3 t
substituting mass flow rate, m', for t = 2 t

velocity, V, one obtains P
t = t
a f C+ C‐
 dP     gdz 
dm m
m  dx  0
2DA 2 A B
A t = 0
x = 0 x = i‐1 x = i x = i+1 x = L
 Integrating along the “characteristic line” from A to P yields

the positive characteristic
PP 
m z x
a P P f P
 dP   dm   g  dz  2 
m
m  dx  0
PA A mA zA 
2 DA x A
PP  PA  a m P  m A   g z P  z A  fx m P m A  0

A 2 DA 2
 A similar equation can be written for the
negative characteristic I3 - 15
Method of Characteristics (2)
t = 5 t
 Introducing two convenient t = 4 t

parameters
t = 3 t
– Impedance
a Where: t = 2 t
B
A A = cross sectional area
P
t = t
C+ C‐
– Resistance A B
t = 0
f x x = 0 x = i‐1 x = i x = i+1 x = L
R
2 DA 2 
 Note that after the initial calculations the impedance and

resistance have constant property values for each pipe,
except for the friction factor, f
I3 - 16
Moody Chart Showing Friction Variation
I3 - 17
t = 5 t
 The equations can now be written t = 4 t
Pi , new  C P  B P m
 i , new t = 3 t
Pi , new  C M  B M m  i , new
t = 2 t
where: t = t
P
 i 1, old   g zi  zi 1 

C P  Pi 1,old  B m C+ C‐
A B
t = 0
 i 1,old   g zi  zi 1 
C M  Pi 1,old  B m x = 0 x = i‐1 x = i x = i+1 x = L
BP  B  R m
 i 1,old
BM  B  R m
 i 1,old
 Here the subscript new refers to point P

– The subscript i-1, old refers to point A
– The subscript i+1, old refers to point B
I3 - 18
t = 5 t
 These two equations are known as t = 4 t
the compatibility equations t = 3 t
Pi , new  C P  B P m
 i , new t = 2 t
Pi , new  C M  B M m  i ,new P
t = t
C+ C‐
A B
t = 0
 Note that there are two equations x = 0
x = i‐1 x = i x = i+1 x = L
and two unknowns

 For example, to solve for the pressure at an interior pipe
location, apply the compatibility equations and eliminate flow
rate C P BM  C M BP
Pi , new 
BP  BM
 The flow rate can be calculated similarly

I3 - 19
t = 5 t
 At x = 0 and x = L, boundary t = 4 t
conditions must be applied t = 3 t

 These are the AFT Impulse
t = 2 t
junctions
P
t = t
 For example, assume that there C+ C‐
is a reservoir at the upstream end, t = 0 A B
x = 0 x = i‐1 x = i x = i+1 x = L
x=0
– In this case, the pressure is known at x = 0 for all times
– The flow rate is obtained by applying the negative compatibility
equation
Pi , new  C M  B M m
 i ,new
Pi , new  C M
 i , new
m 
BM
I3 - 20
Two Methods for Handling Cavitation
DVCM & DGCM
 Discrete Vapor Cavity Model
– Has been available in AFT Impulse since 1996
– Works well for short lived, minor, and localized cavitation
– Has the potential to return chaotic/non-real results when
cavitation becomes excessive
 Discrete Gas Cavity Model
– Has been available in AFT Impulse since 2013
– Has the potential to returns more stable results than DVCM,
especially in 2nd, 3rd, etc. pressure spikes.
– If cavitation becomes excessive, model has harder time
converging than DVCM
– When cavitation is present, this method has the potential to
significantly increase the transient solver run time
I3 - 21
Discrete Vapor Cavity Model
 For an interior pipe point, the pressure is obtained from the

two compatibility equations
C P BM  C M BP
Pi , new 
BP  BM
 If the solution for pressure, Pi,new, is less than the vapor

pressure, a vapor cavity forms
– This is sometimes referred to as liquid column separation
– The pressure then becomes fixed at the vapor pressure, and
acts as a known pressure boundary (that is, pressure cannot go
below the vapor pressure, but can go above)
I3 - 22
Discrete Vapor Cavity Model (2)
 When vapor pressure is reached, the flow no longer balances

at the computing section
 Since the pressure is known (i.e., Psat), the flow rate into and
out of the cavity can be obtained from the two compatibility
equations
vapor cavity
C  Psat m'up
 up  P
m m'down
BP
P CM
 down  sat
m
BM
I3 - 23
Discrete Vapor Cavity Model (3)
 The vapor cavity volume is obtained from integration over the

time step

Vvapor , new  Vvapor ,old   m
 up, new  m
 up, old  m
 down, new  m
 down, old  t
2
 When the calculation for vapor volume is negative, the cavity

collapses and the normal compatibility equations apply once
again
I3 - 24
Discrete Gas Cavity Model
 A distribution of free gas, dissolved in liquid, is modeled as

discrete gas voids concentrated at computing stations
 Pipe sections are treated as liquid full between computing
stations, so the wave speed may be held constant
 A fixed mass of free gas is assumed within each section
 Cavities are considered a mixture of vapor and gas
I3 - 25
Discrete Gas Cavity Model (2)
 The DGCM is distinguished from the DVCM by its ability to

treat cavity pressure as a variable
 Pressure waves do not propagate through Vapor Cavities
because the pressure is held constant, however, Gas Cavities
relate pressure to cavity volume
 The model is simplified by assuming the gas
expansion/contraction is isothermal and behaves as an ideal
gas
 The vapor pressure may be considered constant because the
expansion/contraction is isothermal
 Using Dalton’s Law, the cavity pressure consists of gas and
vapor partial pressures and
I3 - 26
 According to the ideal gas law: · · ·

 With · · constant, the pressure varies only with volume
 With constant vapor pressure, the gas pressure is related
· · · ·
directly to cavity volume: 
 It is generally understood that wave speed depends heavily
on the size and distribution of cavities in the flow
 Cavities will expand and contract based on static pressure
and therefore the wave speed is pressure dependent
 The gas cavity model is very good at simulating the effects of
wave speed pressure dependence, while being more reliable
and avoiding the complications of truly variable wave speed
I3 - 27
 Gas cavities continuously expand and contract based on a

reference pressure, which is above the vapor pressure of the
fluid
 The cavity size is initially very small
and insignificant
 As the static pressure falls, the gas
cavity expands; before reaching the
vapor pressure
 The expansion and the collapse of gas cavities are more
progressive than cavities consisting of vapor only
 The gas cavity model produces more realistic, less extreme
pressure spikes
I3 - 28
Comparing DVCM & DGCM – Exit Valve
Pressure Profile
120
Published Data
100 Impulse DVCM
Impulse DGCM
80
HGL (meters)
60
40
20
-20
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
Time (seconds)
I3 - 29
Comparing DVCM & DGCM – Pipe Interior
Pressure Profile
I3 - 30
Comparing DVCM & DGCM – Simulated
Pump Transient
I3 - 31
Comparing DVCM & DGCM – Simulated
Pump Transient (2)
I3 - 32
Waterhammer Videos - Cavitation

– waterhammer iihr.wmv (1:04)
• Video of rapid, manual valve closure and column separation in a
clear line
• online version -
http://www.iahrmedialibrary.net/db/i1/waterhammer.htm
– GEFA Water Hammer GB.wmv
• High speed movie of a cavitating valve
• online version - http://www.youtube.com/watch?v=X9UbzcanuDk
– check valve.wmv (0:44)
• Video of check valve
• online version – unavailable
I3 - 33
Waterhammer Videos – Cavitation (2)

– Waterhammer 2-0ms.wmv (0:58)
• Video of column separation in a clear line
• online version - http://www.youtube.com/watch?v=bmcOpuzemRU
– How a Bladder Surge Tank can alleviate column
separation1.wmv (1:38)
• Pump trip and column separation
• online version - http://www.youtube.com/watch?v=E6NIA4LxPPw
I3 - 34
Comments
t = 5 t
t = 4 t
 Note that the C+ and C-
characteristic lines lie within t = 3 t
the characteristics grid t = 2 t
 For each computing section, P

t = t
the following therefore is true C+ C‐
A B
t = 0
x
a x = 0 x = i‐1 x = i x = i+1 x = L
t
 Once a pipe is sectioned and the initial conditions are

specified, the solution is entirely explicit
– No iterations are required, unlike the steady-state Solver
– The solution requires marching in time, with each solution
obtainable from the solution at the previous time step
I3 - 35
The Time Step and Pipe Sectioning
 The time step cannot be chosen arbitrarily

– It is determined by the pipe wavespeed and length
 The maximum time step is derived from shortest pipe in the
system
L
tmax  i
ai
– This assumes that only one section is used to model the pipe
– This also assumes that the time step is short enough to properly
resolve all transient boundary conditions
 Once the time step is determined, all other pipes in the
system can be sectioned
Li
ni 
ai t
I3 - 36
The Time Step and Pipe Sectioning (2)
 The resulting number of sections for each pipe will rarely be

whole numbers
– Fractional pipe sections are not possible
 It is therefore necessary to round off the number of computing
sections in each pipe to the nearest whole number
– This introduces errors into the model
 The traditional approach is to recognize that the wavespeed is
the least certain parameter and to allow an error in
wavespeed to be introduced up to ± 15%
– The wavespeed uncertainty is therefore used as justification to
allow sectioning errors up to ± 15%
 AFT Impulse’s Section Pipes window automates the process
of finding optimal sectioning scenarios
I3 - 37
Numbering Convention in AFT Impulse
 The computing stations are at the pipe section boundaries

Computing 0 1 2 3 4 5 6
stations
Pipe sections 1 2 3 4 5 6
 Example of a valve in a pipe
P1 J2 P2
1,0 1,1 1,2 1,3 1,4 2,0 2,1 2,2 2,3
I3 - 38
Numbering Convention in AFT Impulse (2)
 Example of numbering at a branch
P2 2,0
1,0 P1 2,1
1,1 J2 2,2
1,2 2,3
2,4
1,3
3,0 P3
3,1
3,2
3,3
3,4
I3 - 39
Steady-State Data in Transient Solver
 The Transient Solver requires the following:

– Initial steady-state flow rates in all pipes
– Initial pressures at all junctions
– Initial states of all junctions
• Pumps on or off
• Valve open or closed
• Check valves open or closed
• Etc.
– Pipe resistance (friction factors)
I3 - 40
I4. Demo. Problem - Surge at Valve
Closure
Overview of Seminar
INTRODUCTION
Pipes
 AFT Impulse uses two system constructs: pipes and junctions

 Pipes are conduits for liquid, one-dimensional transient fluid
flow
 The flow rate through the length of the pipe is not constant
during the transient
– The waterhammer pressure waves cause flow transients in each
pipe
 Each pipe must be connected to a junction on each end
 A pipe differs from a junction in that it has a reference positive
flow direction
– To say a pipe has a flow rate of 1 ft3/sec is meaningless unless
the flow direction is specified
I4 - 1
Junctions
 Junctions are connector points for pipes

 Junctions are elements at which transient events are initiated
or responded to and where flow balances are made
 Some junction types can only connect to one pipe while
others can connect with up to twenty-five
 AFT Impulse provides a total of twenty standard junction
types
I4 - 2
Junctions (2)
 In addition to balancing flow, junctions also influence the flow

or pressure behavior of the system
– A reservoir junction applies a pressure head at a location, and
the flow at a reservoir is free to adjust in whatever manner is
consistent with the governing equations
– An assigned flow junction applies a flow rate at its location,
allowing the pressure to adjust to that level dictated by the
governing equations
I4 - 3
Creating Objects
 Pipe and junction objects are created using the Workspace

Toolbox
– New pipes and junctions can also be derived from previous ones
by duplication
 Pipes are drawn on the Workspace
 Junctions are dragged from the Toolbox
 Pipe and junctions have default numbers assigned
– Users can reassign numbers
– Pipes numbers are displayed near the pipe center preceded by a
"P"
– Junction numbers are displayed over the junction icons
preceded by a "J"
– Pipes also have a direction arrow displayed to
indicate the positive flow direction I4 - 4
Moving Objects
 The objects on the Workspace can be moved individually or

as groups
 To move an object, select it, drag it within the Workspace, and
drop it in the desired location
– When an object is dragged off the existing Workspace area, the
Workspace is expanded accordingly
 The pipe object can be stretched by grabbing the handles at
the pipe endpoints and moving an endpoint to a new location
 To prevent accidental movement of objects, lock the objects
on the Workspace
– The Lock feature is accessed from the Edit menu or the lock
button on the Toolbar.
I4 - 5
Moving Objects (2)
 To group multiple objects for movement or other operations,

hold down the SHIFT key when selecting the objects
 Objects can also be selected by using the Selection Tool on
the Workspace toolbar
– Click on the Workspace and drag the mouse to draw a box
around the objects
– Holding down the SHIFT key while drawing multiple boxes
permits multiple sets of grouped selections
I4 - 6
Connecting Pipes and Junctions
 Pipes and junction objects can be placed anywhere on the

Workspace
 Connectivity ONLY exists between junctions and pipes
– There are no junctions that connect to junctions, and no pipes
that connect to pipes
 The model connectivity you establish on the Workspace
remains only as long as you maintain the graphical objects in
their current visual relationship to each other
 The most certain way to maintain the connectivity of your
model is to Lock the objects to the Workspace so they cannot
be moved
I4 - 7
Connecting Pipes and Junctions (2)
 To establish a connection between a junction and a pipe, the

following three steps are required:
1) Graphically connect the objects on the Workspace (the pipe
endpoint must terminate within the boundaries of a junction
icon)
2) Enter data for the pipes through the Pipe Property window or
globally
3) Enter data for the junctions through the Junction Property
window or globally
I4 - 8
Editing Objects
 The objects you place on the Workspace can be edited with

the editing commands from the Edit menu or the Toolbar
 Objects can be cut, copied, pasted, duplicated, and deleted
 These operations can be performed on individual objects or
on groups
 One level of undo is available for each editing operation
through the Edit menu
I4 - 9
Lay Out the Model
 Need to find the maximum pressure in system and at valve

 Model looks as below
I4 - 10
Using the Checklist
 The Checklist tracks the status of your model

– Communicates what items must be completed before you can
run the model
 First item is always checked off because AFT Impulse assigns
default Steady Solution Control parameters
– The default Steady Solution Control parameters work
satisfactorily in most cases
 You can open the Checklist box from the Toolbar, View menu,
or Quick Access Panel
I4 - 11
Checklist – Quick Access Panel
 Checklist status available from Status Light on the Quick

Access Panel
Status Light
I4 - 12
Using the Object Status Feature
 Each pipe or junction object requires some minimum input

data
 Until each object has the required input, it is "undefined"
 The Show Object Status feature checks the required data for
each object and reports to the user which objects are and are
not defined
– Undefined object numbers change color (to red by default)
– Holding the shift key down while right clicking on an object will
display a listing of undefined items for that object
I4 - 13
Using the Object Status Feature (2)
 Show Object Status is toggled on and off from the Workspace

Toolbar (flood light) or the View Menu
 Show Object Status should be used selectively because it
slows down the Workspace graphics if left in the ON state
– For large models, users should turn it ON only when needed
I4 - 14
Using Undefined Objects Window
 Opened from the View menu, undefined pipes and junctions

are displayed in lists
 Click on a pipe or junction to see undefined properties
I4 - 15
System Properties Window
 System Properties Window is opened from the Analysis Menu

 This window allows the user to input the fluid properties
– Density
– Viscosity
– Bulk Modulus of Elasticity
– Vapor Pressure (optional, used for transient cavitation)
 Fluid properties can be entered directly or selected from the
database of fluids
– Users only need to enter a temperature for fluids in the database
and the physical properties are calculated
– The fluid database is customizable
 User can also change the gravity level and atmospheric
pressure, and other System Data
I4 - 16
Entering Pipe and Junction Data
 Data for pipes and junctions are entered into Properties

Windows
 Properties Windows are opened either by double-clicking or
single-click then pressing enter for the pipe or junction of
interest
– Properties windows may also be opened by double clicking an
object within the Model Data and Output windows
 Data can also be entered through Global Edit Windows
I4 - 17
Input Data For Pipes
 For the demo problem we will modify the data for each pipe
individually
– The data can also be entered using Global Pipe Edit
 All pipes must have data for
– Length
– Diameter
– Roughness
– Wavespeed (for transient models)
 In addition, each pipe must have two connecting junctions
I4 - 18
Input Data For Junctions
 All junctions must have

– Elevation data
• Connecting pipes are assumed to travel linearly between junctions
– Sufficient number of connecting pipes
• Number of connecting pipes is different for each junction type
I4 - 19
Data For Reservoir Junctions
 All Reservoir junctions can have from 1-25 connecting pipes
I4 - 20
Data For Branch
 Branch junctions can have from 2-25 pipes

 Branches are general connectors and can also be used as
boundaries for diameter changes
I4 - 21
Data For Valve
 Valve junctions can have from 1-2 pipes
I4 - 22
Inspecting Objects
 The data in a pipe or junction can be reviewed quickly using

the inspection feature
 Inspecting is done by pressing down the right mouse button
on the graphical pipe or junction
 Inspecting is much quicker than opening the Properties
Window
– Using the inspection window also does not clear the output
results as opening a Properties window can
I4 - 23
Inspecting Objects – Quick Access Panel
 Pipe and Junction input/output data can be viewed in Quick

Access Panel
– Click the “Properties” tab on Quick Access Panel
– Select a pipe or a junction on Workspace
I4 - 24
Model Data Window
 The Model Data window is useful for reviewing the text input
for the model
– All data can be printed out for documentation
 Model Data can be accessed from the Window menu
 Use the Model Data window to do a quick sanity check of the
input
– Incorrect units or a typo become more obvious in Model Data
 Double-clicking the far left column of the tables opens the
appropriate Properties Window
I4 - 25
Pipe Sectioning Window
 Section Pipes Window is opened from the Analysis Menu

 Automates the selection of an optimal sectioning strategy
I4 - 26
Transient Control Window
 Transient Control Window is opened from the Analysis Menu

 This window gives user control of time length of simulation
and various calculation and reporting inputs
I4 - 27
Run the Model: Solution Progress
Window
 When a model is complete, the Run command is enabled
 The model can be run by choosing Run from the Analysis
Menu or clicking the appropriate toolbar icon
 When a model is running, the Solution Progress Window
displays
 The Solution Progress Window shows the status of the
Steady-State and Transient Solver's progress towards
convergence
 The Solution Progress Window allows you to Cancel or Pause
the run
 When the solution completes, you are notified
 When you select View Results, you are immediately
taken to the Output Window
I4 - 28
Output Window
 The Output Window displays text output for your model and is
accessed from the Window menu or toolbar
 The Output Control Window allows you to customize the
content of the output
 Each section may have multiple tabs to quickly view data by
type
 Print Content allows you to select the content of the printed
report
I4 - 29
Output Window (2)
 Transfer Results to Initial Guesses saves the current output

results as the initial conditions
– Transfer Results to Initial Guesses is accessed from the Edit
menu of the Toolbar (push pin)
 Warnings are placed into the General Results section
– When warnings exist the text color is changed to red
 Sort allows you to sort the Output according to one the
columns
 Double-clicking the column header allows you to change the
units for that column
I4 - 30
Output Window Transient Pipe Data
 The Transient Output tab shows data selected on the Pipe

Transient tab in Output Control
 The slider bar allows you to look at
numerical data for any time step
I4 - 31
Output Window Transient Max/Min Data
 The Transient Max/Min tab shows numerical maximum and

minimum data for each pipe or each pipe station
 Here the data is shown for each pipe because in Output
Control the Summary option
was chosen
– This is the default
I4 - 32
Output Window Transient Max/Min Data
(2)
 If the Detailed option is selected then max/min data is shown
for each pipe station
I4 - 33
Output Window Transient Junction Data
 Transient data is not available for junctions in the output as it

is of little benefit
– Much of it would be redundant to the available pipe transient
data
– Could result in long delays in displaying the Output window due
to the quantity of data
 A wide range of junction transient data is available from the
Graph Results with the numerical data for these graphs easily
exported
I4 - 34
Graph Results
 Graphs are created with the Graph Results Windows

– This window is one of the Primary Window tabs
– Graph Results can also be accessed from the Window menu
 Various parameters can be graphed by clicking on the Select
Graph Data button in the Graph Results window
 The graph can be printed, copied to the clipboard, or saved to
a file
 The graph x-y data can be exported to a file or copied to the
clipboard
I4 - 35
Graph Results (2)
 Graph the pressure up and

downstream of the valve
I4 - 36
Graph Results (3)
 Run an animation
– Use "Animate Using Solver"
– Must be a continuous path
of pipes
– To use "Animate Using Output
File" all pipe station output
needs to be saved in Transient Control
I4 - 37
Visual Report
 Visual Report allows you to see the results superimposed on

the Workspace graphic
– This is one of the Primary Windows
– Visual Report can be accessed from the Window menu
 The Visual Report Control allows you to select the type of
results you want to see
 You can print the image at full size or fit it to a single page
with Print Special
 Text locations are automatically saved with the model
I4 - 38
Visual Report (2)
I4 - 39
Pipe Supports and Wavespeed
 The type of pipe support selected on the Pipe Properties

window affects the calculated wavespeed
– The effect on wavespeed is not typically very significant
 The table on the next slide shows the effect of support type on
wavespeed in 10 inch Steel - ANSI, Schedule 40 pipes:
I4 - 40
Pipe Supports and Wavespeed (2)
Support Type Wavespeed (ft/s) Wavespeed (m/s)
Thin-Walled Anchored Upstream 4227 1288
Thin-Walled Anchored Throughout 4199 1280
Thin-Walled With Expansion Joints 4158 1267
Thick-Walled Anchored Upstream 4197 1279
Thick-Walled Anchored Throughout 4171 1271
Thick-Walled With Expansion Joints 4132 1259
Circular Tunnel 4637 1413
I4 - 41
US
Input Data for Demo 1
Water at 70F
Pipe depths in Reservoirs J1, J2 are 10 feet

Pipe depth in Reservoir J5 is 5 feet
All piping Steel-ANSI, sch40, with a calculated wave speed using
‘Thick-Walled Anchored Throughout’ support
I4 - 42
SI
Input Data for Demo 1
Water at 21C
Pipe depths in Reservoirs J1, J2 are 3 meters

Pipe depth in Reservoir J5 is 1.5 meters
All piping Steel-ANSI, sch40, with a calculated wave speed using
‘Thick-Walled Anchored Throughout’ support
I4 - 43
I5. AFT Impulse Hands-On Modeling
Overview of Seminar
INTRODUCTION
Introduction To Scenario Manager
 The Scenario Manager allows you to keep variants of a model

all with the same model
 The types of changes that can be made are very broad
– Junctions can be turned on and off to evaluate different
operating conditions
– Pipe and junction data can be varied to parametrically evaluate
competing designs
– You can build an existing system as your base model then add to
the system to evaluate expansion possibilities on the existing
system
– You can easily evaluate different working fluids by setting them
up as different children scenarios
I5 - 1
Introduction To Scenario Manager (2)
 Scenarios are created, manipulated, and loaded using the

Scenario Manager window
 The Scenario Manager can be opened from the Tools menu in
the Workspace window, the Scenario Manager button on the
toolbar, or Quick Access Panel
From Quick
Access Panel
I5 - 2
Notes
I5 - 3
Notes
I5 - 4
US
Single Line Valve Closure
 Test Problem #1 (TEST1.IMP) - Find pressure surge at valve

 Compare to instantaneous equation results P    aV
1 pipe 1
Liquid Height = 500 ft
Depth = 30 feet L = 2000 ft 2
D = 20 inch Elevation = 0 ft
(19.25 in ID) Exit Valve, P = 0 psig
a = 3876 ft/s t Cv
(calculated) 0 530
Water at 65F
10 pipe sections .001 0
0 to 4.5 seconds 1 0
Ignore Cavitation
Pipe is Steel‐ANSI, STD (Schedule 20)
I5 - 5
US
Single Line Valve Closure (2)
 Instantaneous waterhammer calculation

P    aV
V   Q/A
V  17.06 ft3/sec / (2.021 ft2)
V  8.44 ft/sec
2 2
Pmax  62.3 3 * 3876 * 8.44 * 2 *  223 psia
ft s s 144 in 32.2 lbm ‐ ft
Pmax  440  223
Pmax  663 psia
I5 - 6
US
 What is the communication time for this system?

 How far negative does the pressure spike go at the valve?
 What is the maximum pressure at the valve if the valve is
closed linearly over 0.99 seconds?
I5 - 7
US
 Change the TEST1.IMP model to handle cavitation

(TEST1.IMP, /With Cavitation scenario)
– Use the “Discrete Vapor Cavity Model”
 What is the maximum pressure?
 How does it compare to the instantaneous waterhammer
prediction?
– How far above or below the maximum valve pressure is the
instantaneous waterhammer prediction?
– How does this difference compare to the negative pressure in
the original model?
 Does the maximum pressure change when the number of
sections is increased to 20 or 40?
I5 - 8
US
Model Ammonia Transfer System
 Find maximum pressure in the system with different closure

rates (TEST3.IMP)
– All pipe is Steel – ANSI with standard wall thickness, thick-walled
anchored upstream
– Look at valve closing in 0.5, 1.0 and 2.0 seconds
I5 - 9
US
Model Ammonia Transfer System (2)
P1 P2 P3
L = 100 ft L = 300 ft L = 150 ft
8 inch 10 inch 10 inch
1 2 3 4
Surface Elev. = 20 ft
Surface Elev. = 40 ft Abrupt Expansion Surface Pressure = 250 psig
Surface Pressure = 250 psig Elevation = 0 ft
Pipe Depth = 5 ft
Pipe Depth = 20 ft
Valve
Elevation = 0 ft
Ammonia at 75F
t (sec) Cv
0 to 5 seconds
0 1000
Model Cavitation using the
? 0
“Discrete Vapor Cavity Model”
I5 - 10
US
Answers to Problems
 TEST1 pressures at valve

– Instantaneous pressure is:
• Instantaneous Waterhammer Calculation = 663 psia
• Impulse Results = 671.4 psia
» ~1% difference in results
– Communication time is 1.032 seconds (2L/a)
– Pressure goes to 202 psia below zero
– With 0.99 second closure the max pressure at the valve is 668
psia – about the same
– With 1.2 second closure the max pressure at the valve is 602
psia
I5 - 11
US
Answers to Problems (2)
I5 - 12
US
 TEST1 with cavitation (Discrete Vapor Cavity Model)

– Pressure is higher than “maximum theoretical pressure” (875 vs.
663; 212 psi higher)
– Negative pressure w/o cavitation was 202 psi below zero and
with cavitation it is 212 psi higher than maximum theoretical – a
similar magnitude
– With 20 sections max pressure is 855 psia - with 40 sections
max pressure is 926 psia
I5 - 13
US
I5 - 14
US
 TEST3 results
Closure Max Stag. Pressure*
Time (sec) (psig)
0.5 599
1 576
2 365
– (*) The first two cases yield different pressures when the
sectioning is varied
• This is a result of the cavitation model
– The 2 second closure case does not cavitate
I5 - 15
Notes
I5 - 16
SI
Single Line Valve Closure
 Test Problem #1 (TEST1.IMP) - Find pressure surge at valve

 Compare to instantaneous equation results P    aV
1 pipe 1
Liquid Height = 150 m
Depth = 10 m L = 600 m 2
D = 20 inch
Elevation = 0 m
(48.9 cm ID)
Exit Valve, P = 0 barG
a = 1181 m/s
t Cv
(calculated)
Water at 18C 0 530
10 pipe sections .001 0
0 to 4.5 seconds 1 0
Ignore Cavitation
Pipe is Steel‐ANSI STD (Schedule 20)
I5 - 17
SI
 Instantaneous waterhammer calculation

P    aV
V   Q/A
V  0.4794 m3/sec / (0.1878 m2)
V  2.553 m/sec
kg m m
Pmax  998.7 * 1181 * 2.553  1,515,000 Pa
m3 s s
Pmax  3,011,173  1,518,000 Pa
Pmax  4,529,173 Pa  4.53 MPa  45.3 bar
I5 - 18
SI
 What is the communication time for this system?

 How far negative does the pressure spike go at the valve?
I5 - 19
SI
 Change the TEST1 (SI).IMP model to handle cavitation

(TEST1 (SI).IMP, /With Cavitation scenario)
– Use the “Discrete Vapor Cavity Model”
 What is the maximum pressure?
 How does it compare to the instantaneous waterhammer
prediction?
– How far above or below the maximum valve pressure is the
instantaneous waterhammer prediction?
– How does this difference compare to the negative pressure in
the original model?
 Does the maximum pressure change when the number of
sections is increased to 20 or 40?
I5 - 20
SI
Model Ammonia Transfer System
 Find maximum pressure in the system with different closure

rates (TEST3 (SI).IMP)
– All pipe is Steel - ANSI with standard wall thickness, thin-walled
anchored upstream
– Look at valve closing in 0.5, 1.0 and 2.0 seconds
I5 - 21
SI
Model Ammonia Transfer System (2)
P1 P2 P3
L = 30 m L = 91 m L = 46 m
8 inch 10 inch 10 inch
(20.3 cm ID) (25.5 cm ID) (25.5 cm ID)
1 2 3 4
Surface Elev. = 12 m Surface Elev. = 6 m
Surface Pressure = 18.2 bar Surface Pressure = 18.2 bar
Pipe Depth = 6 m Pipe Depth = 1.5 m
Abrupt Expansion
Elevation = 0 m
Valve
Ammonia at 24C Elevation = 0 m
0 to 5 seconds t (sec) Cv
Model Cavitation using the 0 1000
“Discrete Vapor Cavity Model” ? 0
I5 - 22
SI
Answers to Problems
 TEST1 (SI) pressures at valve

– Instantaneous pressure is:
• Instantaneous Waterhammer Calculation = 45.3 bar
• Impulse Results = 45.8 bar
» ~1% difference in results
– Communication time is 1.016 seconds (2L/a)
– Pressure goes to 13.94 bar below zero, at the valve
– With 0.99 second closure the max pressure at the valve is 45.8
bar – about the same
– With 1.2 second closure the max pressure at the valve is 38.2
bar
I5 - 23
SI
I5 - 24
SI
 TEST1 (SI) with cavitation (Discrete Vapor Cavity Model)

– Pressure higher than “maximum theoretical pressure” 66.1 vs.
44.3; 20.8 bar higher)
– Negative pressure w/o cavitation was 13.94 bar below zero and
with cavitation it is 21.81 bar higher than maximum theoretical –
a similar magnitude
– With 20 sections max pressure is 58.6 bar - with 40 sections max
is 62.2 bar
I5 - 25
SI
I5 - 26
SI
 TEST3 (SI) results
Closure Max Stag. Pressure*
Time (sec) (bar)
0.5 41.26
1 38.22
2 25.91
– (*) The first two cases yield different pressures when the
sectioning is varied
• This is a result of the cavitation model
– The 2 second closure case does not cavitate
I5 - 27
What is Line Pack?
 Line pack is a pressure rise due to recovery of pressure

previously lost to friction
I5 - 28
US
What is Line Pack? (2)
 Consider the model “Line Pack Example.imp”, Base scenario

– Steady-state conditions shown below for a high viscosity oil
pipeline
• 28 miles (45 km) long
– Question: What will the pressure be at J2 after the valve closes
and the fluid has come to rest?
I5 - 29
US
 Answer: The pressure at J2 (and the entire pipeline) be 1499

psig after the fluid has come to rest
 How much did the pressure rise?
– Originally the J2 pressure was 301 psig and at the end it will be
1499 psig
• 1499 – 301 psig = 1198 psid
 This is due to line pack
I5 - 30
SI
 Consider the model “Line Pack Example.imp”, Base scenario

– Steady-state conditions shown below for a high viscosity oil
pipeline
• 28 miles (45 km) long
– Question: What will the pressure be at J2 after the valve closes
and the fluid has come to rest?
I5 - 31
SI
 Answer: The pressure at J2 (and the entire pipeline) be 103

barG after the fluid has come to rest
 How much did the pressure rise?
– Originally the J2 pressure was 20.8 psig and at the end it will be
103.4 barG
• 103.4 – 20.8 barG = 82.6 bar
 This is due to line pack
I5 - 32
Line Pack and Waterhammer
Instant Valve Closure
 What is the pressure rise from the instantaneous

waterhammer equation for this example?
P    aV
2 2
Pmax  53 3 * 3,572 * 9.92 * 2 *  301 psig
ft s s 144 in 32.2 lbm‐ ft
US Pmax  405  301
Pmax  706 psig
kg m m 1 bar
Pmax  849 *1,089 *3.02 * + 20.8 barG
m3 s s 100,000 Pa
SI Pmax  27.9 + 20.8
Pmax  48.7 barG
I5 - 33
Line Pack and Waterhammer (2)
Instant Valve Closure (First 10 seconds)
I5 - 34
Instant Valve Closure (All 6 minutes)
I5 - 35
Maximum Pressure * Peak potential pressure
= 1904 psig
= 131 barG
Line pack pressure rise
= 1198 psi
= 83 bar
Instantaneous water-
hammer pressure rise
= 405 psi
= 28 bar
* In the absence of cavitation

I5 - 36
Max/Min Pressure Peak potential pressure
= 1904 psig
= 131 barG
Line pack pressure rise
= 1198 psi
= 83 bar
Instantaneous water-
hammer pressure rise
= 405 psi
= 28 bar
I5 - 37
Line Pack and Waterhammer Video

– Line Pack Example.mp4 (0:32)
• Impulse model file “Line Pack Example.imp”
I5 - 38
Overview of Seminar
INTRODUCTION
Pipe and Junction Details
 Pipes and Junctions are defined using Properties windows

 The Global Edit windows allow you to change data for several
pipes or junctions at the same time
 The required data in each Properties window can be
highlighted
– This allows you to easily see what data is needed or missing
 A note can be saved for each pipe and junction
 Right clicking on a junction will display an input summary
 The Quick Access Panel can also be used to review the input
and output of a pipe or junction
I6 - 1
The Pipe Properties Window
 Each pipe must have an ID number, length, diameter,
roughness model, and two connecting junctions
– If modeling transients, a wavespeed is also required
 AFT Impulse can model one pipe geometry
– Cylindrical pipe
 AFT Impulse provides default pipe material data for eight
types of pipe
 AFT Impulse offers eight methods for specifying Newtonian
frictional models
– Four methods rely on the Darcy-Weisbach friction factor method
of calculating pipe pressure drop
– One uses the Hazen-Williams empirical factor
– Two are for hydrocarbons
– One is frictionless
I6 - 2
Pipe Fittings and Losses
 Multiple additional (minor) losses can be added to any pipe

Losses are grouped by type
Select the quantity from the list
The losses of this type in the
pipe are listed here
User Favorites allow for much faster
navigation to frequently used fittings
The picture helps All of the losses are listed
describe the loss on the Summary tab I6 - 3
Optional Pipe Information
 An initial guess flow

rate can help the
Steady-State Solver
reach convergence
faster
 For each pipe you
can choose to display
the ID number, name,
size and schedule on
the Workspace
 Design factors add
safety margin to
calculations
I6 - 4
Optional Pipe Information (2)
 Specify that this pipe

is one of a number of
parallel pipes
 Change pipe line
thickness and color
 Pipes can be
specified as partially
full
 Enter intermediate
elevations for pipes
I6 - 5
Design Alerts
 Design Alerts allow you to specify maximum or minimum
parameters that, if exceeded, will generate special output
information
– Design Alerts can be specified in the Pipe and Junction
Properties windows or the Design Alert Manager located in the
Tools menu for pipes and junctions
I6 - 6
Design Alert Manager
 Design Alert Manager accessed from Tools menu allows quick

and global Design Alert allocation to pipes and junctions
List of Existing
Design Alerts
Specify
Pipes/Junctions
Define Design to apply Design
Alert maximum Alert to
or minimum
criteria Create/Edit
Design Alert
I6 - 7
The Junction Properties Window
 Each junction type uses a different Properties window

 Each junction must have an elevation
 You can choose junctions from a Custom Component
database list (if you create one) to specify the junction
 Junctions with two pipes connected may have outlet
elevations specified different than the inlet
 Junctions which can connect to more than two pipes will show
the connected pipes in a table in which you can enter loss
factors and individual elevations
I6 - 8
Optional Junction Information
An initial guess pressure can
help the Solver reach
convergence faster
For each junction you can

choose to display the ID
number, name, both or
neither on the Workspace
Design factors add

safety margin to
calculations
The size of each

individual junction icon Some junctions have Special Conditions which can be set to
be changed override the default behavior of the junction. Usually this will
turn off or close the junction to flow. Relief valve will be open
when the Special Conditions are set
I6 - 9
Global Pipe Editing
 Global Pipe Edit, on the Edit menu, assists you in making

large scale changes to your model
 To globally edit the pipe data
– Select the pipes to change from the list, or use the Select
Special tool
– Click on the Select Pipe Data button to show the Pipe Properties
window
– Enter the data you want to change and click OK
– Choose the parameters to be changed and click the Apply
Selections button
– Confirm that you want to change the selected pipes by clicking
OK or Cancel the change
I6 - 10
Global Pipe Editing (2)
4) Apply the data
1) Choose which
pipes to edit
3) Choose which
parameters to
2) Select the data change
which then displays
at the right
I6 - 11
Global Junction Editing
 You can change the general data or junction specific data

 To globally edit the common junction data
– Select All Junctions in the drop-down list box
– Click Select Common Junction Data
– Enter the new data in the Common Junction Data Edit window
and click OK
– Check the boxes on the right to select the parameters to be
changed
– Click on the Apply Selections button.
I6 - 12
Global Junction Editing (2)
Choose which
junctions to edit
Enter the
new data
Choose which
parameters to
change
I6 - 13
 To globally edit specific junction data

– Select option for Specific Data For… and then the junction type
in the drop-down list box and choose the junctions to change
– Click on the Select Specific Junction Data button to show the
Properties window for the junction type (the button will display
the junction type)
– Enter the data you want to change and click OK
– Choose the parameters to be changed and click the Apply
Selections button
I6 - 14
Choose the
junction type
Choose which
parameters to
change
Choose which
junctions to edit
I6 - 15
Design Factors
 Design Factors are available for pipe friction, junction friction,

fitting losses, and pump curves
– These can be used as safety factors in the design
– They can be set for each pipe/junction or globally
I6 - 16
Communication Times
 Be aware of the relevant communication times in your system

– Check the frequency content of your results vs. the
communication time
 Use Infinite Pipe junctions to eliminate non-interacting parts of
the pipe system
I6 - 17
Dead End Junctions
 Dead ends are more important in waterhammer modeling

than in steady flow - do not neglect them too quickly!
 Dead ends are areas that can absorb pressure waves and
change the transient behavior of the system
I6 - 18
Be Careful Assigning Flowrates
 What happens if you have an inflow assigned then close the

valve?
– The inflow will continue with nowhere to go, causing pressure

buildup in the system
 Your initial thinking about a problem should focus on pressure
type boundary conditions
– Use flow boundary conditions judiciously
– Usually they are inappropriate during transients because the flow
rates change during the transient
I6 - 19
Valid Example Using Assigned Flows
 Measured field data

for flows and pressures
was recorded
surrounding a pump trip
and valve slamming
event, which caused
excessive pressures in
the system.
 Assigned Flow Junctions were used to represent delivery
points at the extents of the model.
I6 - 20
Valid Example Using Assigned Flows (2)
 Transient flows and

triggers were adjusted
using an iterative
approach which led to
a very good match
between model and test
data.
 After developing an
accurately calibrated model of the system, multiple solutions
for the waterhammer issues were evaluated and the best
solution was selected and implemented.
 This study has been published on our website:
– http://www.aft.com/about-us/platinum-pipe-award
I6 - 21
US
Hands-on Example: Pressure Transient
 A piping system delivers water to two reservoirs. The supply

pressure cycles between 50 and 100 psia over a 10 second
period. Determine how the flowrates vary to each of the
discharge reservoirs over time. (Pressure Transient
Example.imp)
I6 - 22
US
Pressure Transient Example (2)
 Fluid is water at 70 deg. F

 Pipe data:
– Use the Global Edit feature to enter the piping data
– Steel – ANSI pipe, all STD (Schedule 40) with lengths of 1000 ft,
and diameters of 10 inches
– Pipe P1 has 3 smooth flanged, 1.5 r/D, 90 deg. bends
• Add them as Fittings & Losses
 Branch data:
– Elevation is 0 feet
I6 - 23
US
 Reservoir data:
– J3 surface elevation is 30 feet, pipe depth is 10 feet, surface
pressure is 0 psig
– J4 surface elevation is 10 feet, pipe depth is 10 feet, surface
pressure is 0 psig
 Assigned Pressure data:
– Elevation is 0 feet
– Steady-state stagnation pressure is 50 psia
I6 - 24
US
Pressure Transients Example (4)
 Assigned Pressure transient data:

– The pressure at J1 varies according to the following profile:
Time (s) Pressure (psia)

0 50
2.5 85
5 100
7.5 85
10 50
– The transient repeats

 Assume 4 pipe sections when sectioning the pipe
 Run the simulation for 30 seconds
I6 - 25
US
Pressure Transient Example Results
 Plot the flow rates at the reservoirs. Save the graph as a

Graph Set called “Flow Rates at Discharge Reservoirs”
 Plot the inlet stagnation pressure in pipe P1
 Load the “Flow Rates at Discharge Reservoirs” and show the
graph
I6 - 26
US
Pressure Transient Example Results (2)
I6 - 27
Notes
I6 - 28
Notes
I6 - 29
SI
Hands-on Example: Pressure Transient
 A piping system delivers water to two reservoirs. The supply

pressure cycles between 3.5 bar and 7 bar over a 10 second
period. Determine how the flowrates vary to each of the
discharge reservoirs over time. (Pressure Transient Example
(SI).imp)
I6 - 30
SI
 Fluid is water at 21 deg. C

 Pipe data:
– Use the Global Edit feature to enter the piping data
– Steel - ANSI pipe, all STD (Schedule 40) with lengths of 300
meters, and diameters of 10 inches (25.45 cm ID)
– Pipe P1 has 3 smooth flanged, 1.5 r/D, 90 deg. bends
• Add them as Fittings & Losses
 Branch data:
– Elevation is 0 meters
I6 - 31
SI
 Reservoir data:
– J3 surface elevation is 10 meters, pipe depth is 3 meters,
surface pressure is 0 barG
– J4 surface elevation is 3 meters, pipe depth is 3 meters, surface
pressure is 0 barG
 Assigned Pressure data:
– Elevation is 0 meters
– Steady-state stagnation pressure is 3.5 bar
I6 - 32
SI
Pressure Transients Example (4)
 Assigned Pressure transient data:

– The pressure at J1 varies according to the following profile:
Time (s) Pressure (bar)
0 3.5
2.5 6.0
5 7.0
7.5 6.0
10 3.5
– The transient repeats

 Assume 4 pipe sections when sectioning the pipe
 Run the simulation for 30 seconds
I6 - 33
SI
Pressure Transient Example Results
 Plot the flow rates at the reservoirs. Save the graph as a

Graph Set called “Flow Rates at Discharge Reservoirs”
 Plot the inlet stagnation pressure in pipe P1
 Load the “Flow Rates at Discharge Reservoirs” and show the
graph
I6 - 34
SI
Pressure Transient Example Results (2)
I6 - 35
Overview of Seminar
INTRODUCTION
The Five Primary Windows
 Workspace window
– Where you build the model visually and see the model layout
 Model Data window
– A text-based complement to the Workspace window
– Shows input data in text form
 Output window
– Displays the results of the analysis in text form
– Lets you produce attractive, effectively organized printed output
reports
I7 - 1
The Five Primary Windows (2)
 Visual Report window

– Merges the output data with the pipe system layout from the
Workspace
– Allows customized layout of the information for documentation
and presentation purposes
 Graph Results window
– A full-featured, customizable Windows plotting capability
– Results can be viewed in a variety of ways for evaluating system
performance or identifying important trends
I7 - 2
The Workspace Window - Editing
 You can zoom out to see a larger area

– By using the Zoom control or clicking on the + or - magnifying
glasses on the toolbar
 Objects can be selected as a group in several ways
– Using the mouse to highlight pipes and junctions
– Using the SHIFT key while clicking on the objects
– Using Select Flow Path on the Edit menu
– Using the Select Special tool on the Edit menu
– If a group or groups have been defined, by selecting Groups on
the Edit menu
I7 - 3
The Workspace Window – Editing (2)
 The Reference Flow Direction of a pipe can be changed

 The selected objects can be renumbered in three ways
– "Automatic Renumber“, "Renumber Increment" and "Renumber
Wizard"
 The Find tool will move the Workspace window to show a pipe
or junction
 Notes may be added with Annotation tool
 A background may be displayed
– Especially useful in conjunction with Snap
 A bitmap file may be displayed as a Workspace background
I7 - 4
The Workspace Window - Arranging
 The selected objects can be scaled or flipped

 The selected objects can be aligned
The assembly flipped vertically and horizontally
I7 - 5
The Workspace Window - Rotating
 A group of pipes and junctions can be rotated
I7 - 6
The Workspace Window - Selection Tool
Dragging left-to-right selects all objects Dragging right-to-left selects all objects
completely within the selection box completely or partially within the selection box
I7 - 7
The Workspace Window - Zoom to Fit
 The Workspace can be "zoomed to fit" so that the entire

model is visible
Zoom to Fit
I7 - 8
Quick Access Panel – Overview Map
 Drag the View Port over large models to zoom in on the

Overview Map
View Port
Access to Overview Map I7 - 9

Workspace - Annotation Manager
 Annotations can be hidden and displayed in scenarios of your

choice
I7 - 10
Using Select Special
 Select Special is a powerful tool which selects or deselects

objects based on certain criteria
 If accessed from the Workspace it will select objects on the
Workspace
 If accessed from other windows it will select items in a list box
 It can set the list box items based on what is selected on the
Workspace
I7 - 11
Using Select Special (2)
 Criteria options include

– What is selected on the Workspace
– Objects which are open
– Objects with the Special Conditions set
– Objects within a range of ID numbers
– Junctions of a certain type
– Pipe diameter or length
 Invert will select objects not selected and deselect objects
which are selected
 You can perform several selection operations at a time
I7 - 12
Select Special Example
 Select everything except the Assigned Pressure junctions

1) Select Junction Type, Assigned
Pressure, then click the Select
Junctions button
2) Choose Invert to get all of the

3) Select all pipes non-Assigned Pressure junctions
I7 - 13
The Model Data Window
 The three areas of the Model Data window show all the
information about the model
 Detailed information about the Pipe Fittings & Losses can be
displayed in the Pipe data area
 Model Data is a valuable tool for identifying modeling errors
 Use the Model Data Control to customize what is displayed
 You can optionally show just the pipes and junctions in which
you are interested
 You can select just the data in which you are interested
I7 - 14
The Model Data Window (2)
 Arrow buttons collapse input section

 Each section can be resized manually
I7 - 15
The Model Data Window (3)
 Intermediate pipe elevations can be shown

 Junction curve fit raw data can be shown
I7 - 16
The Output Window
 The Output window is the primary vehicle for communicating

the results of an analysis in text form
 The Output Control, on the Tools Menu, is used to customize
the Output window
– You choose the output parameters and units for the pipes and
junctions
– You set the order of the parameters in the output
– You choose what general information to display in the output
– You enter a title and reference information
– You select the number of digits and numerical display options
I7 - 17
The Output Window (2)
 You can optionally show only the pipes and junctions in which
you are interested
 Special report tables in the General Output include:
– Summary for Pump, Valve, and Reservoir data in the General
section
– Warnings regarding solution
– Differences in parameters from one junction to another (Junction
Deltas)
 Double clicking on an item will open its Properties window in
read-only format
I7 - 18
The Output Window (3)
 Double-clicking the column header allows you to change the

units for that column
 Dragging the column can change the order displayed
 Arrow buttons collapse input section
 Each section can be resized manually
I7 - 19
Using The Output Control
 Example: Open “Pump Trip with Backflow.imp”

 Set the following Output Control Parameters:
– Volumetric Flow Rate (gal/min)
– Static inlet Pressure (psia)
– Static outlet Pressure (psia)
– Pressure Loss Static Total (psid)
– Show only junctions 21, 22, 31, and 32
– Show only pipes 1, 2, 11, and 12
I7 - 20
Using the Output Control (2)
I7 - 21
The Visual Report Window
 Integrates the pipe system layout from the Workspace window

with information in text form
 The values to be shown may come from the Model Data
values or the Output results
 Text can be moved to give the best visual presentation
 Notes may be added using the annotation button on the
toolbar
 The content is controlled using the Visual Report Control
window
– You choose the parameters to show
– You select to show the title and additional information
– You choose font, symbols and legend options
I7 - 22
Using Visual Report Control
 Choose the parameters you want to be displayed
These parameters will be

shown on the Visual Report
I7 - 23
Using Show Selected Options
 Each object can be set to show any combination of object,

name, number and data
Don’t show name or data for the junctions
Only show the data for

pipes (no name or number) I7 - 24
Using The Color Map
 A Color Map can be used to visualize values of various

parameters by coloring the pipes, for example,
– You can show graphically where maximum or minimum
pressures are located
– You can better visualize the pipe diameters or lengths in a model
I7 - 25
Using the Color Map (2)
The Color Map in the legend
Color map can be

automatically generated
Pipes are colored according

to max stagnation pressure
I7 - 26
The Graph Results Window
 You can generate high quality printed graphs

 Four types of graphs can be selected using the Select Data
window
– Transient parameter at a pipe station (time on the x-axis)
– Transient parameter at a junction (time on the x-axis)
– Profile along a flow path (distance on the x-axis)
– Imbalanced forces (visible only when force sets are specified)
 Graph x-y data can be copied directly to the clipboard
 You can graph multiple parameters on Stacked Graphs with a
common X-axis
 You can create dual Y-axis graphs
I7 - 27
Graph Control on the Quick Access Panel
 Graphs may be saved as Graph List Items in folders in the

Graph List Manager and quickly regenerated after rerunning
the model or running other scenarios
 Animation feature allows user to replay transient data along a
flow path with dynamic graphics
 Multiple profile plots can be cross-plotted using Workspace
Groups
 Each type of graph is defined on color-differentiated tabs
– Select the tab to view the applicable parameters for that graph
type
 Can perform extensive graph formatting from the Formatting
flyover panel
I7 - 28
Graph Control on Quick Access Panel (2)
 Graph Control is a tab on the

Quick Access Panel (QAP)
– Allows for easy graph generation
and configuring
I7 - 29
Graph List Manager
Add Graph to List Update Current
Graph in List
 Graph folders can be created to organize your graph list items
Graph Folders
Graph List Items
I7 - 30
Graph Tabs
 Graph Tabs allow for multiple graphs

– Graphs Tab can be tiled
I7 - 31
Graph Guide
 The Graph Guide helps you build graphs using Direction

Bubbles
Graph Guide
I7 - 32
Stacked Graphs
 Stacked graphs can show different parameters using the

same X-axis for easy comparison
I7 - 33
Dual Y-Axis Graphs- Different Parameters
 Two different parameters can be shown on the same graph

using a second Y-axis
I7 - 34
Dual Y-Axis Graphs- Same Parameter,
Different Units
 Dual Y-axes can be used to show the same parameter with
different units
I7 - 35
Design Alerts on Graphs
 The Design Alerts can be cross-plotted to show, for example,

maximum operating pressure
– Design alerts are set in the Graph Control next to the parameter
I7 - 36
Cross-Plot Graphs Using Groups
 Multiple pipe sequences can be plotted as profile graphs
I7 - 37
Zoom and Pan on Graphs
 To pan, hold the left mouse key and drag the mouse
 To zoom, click on graph and scroll the mouse wheel
 To zoom on an axis, place mouse over axis and use scroll
wheel
 Can use range finder to zoom in on desired portions of the
graph
I7 - 38
Graph Formatting
 The formatting control can be accessed on a flyover panel

on the QAP
 Title and axis texts can be
changed
 The Size All - and +
buttons are convenient to
change all at once
 The axis scales and tick
marks can be set
I7 - 39
Graph Formatting (2)
 Right-clicking on the Title or an axis allows quick editing
 Right-clicking on the graph

brings up a menu of additional
functions and formatting options
- Formatting is listed at the
bottom
I7 - 40
Other Graph Functionality
 Crosshairs can be shown to callout a data point

 A range finder (shown at the bottom) can be used to zoom
into a portion of the graph
I7 - 41
Other Graph Functionality (2)
 Right-clicking on a graph list folder allows:

– Printing all graphs in the folder
– Generating all graphs in the same tab or separate tabs
– Renaming, moving and deleting
 Data points for each series are grouped together and can be
displayed in a list
– Clicking on a point on the graph with the crosshair selected will
highlight the data in the list
– Clicking on data in the list will select the point on the graph
– Can convert points of interest to annotations on the graph
I7 - 42
Create Graphs from the Workspace
 Right-click on objects on Workspace to directly generate a

graph
I7 - 43
Animating the Results
 Animation feature allows user to replay transient data along a

flow path with dynamic graphics
 Must be along a continuous path of pipes
– Multiple paths can be animated using Workspace Groups
 Can animate Stacked and Dual Y-Axis graphs simultaneously
 Use "Animate Using Solver" to run the animation using the
solver
– Do not have to save data points for all pipe stations, so output
file size can be reduced
I7 - 44
Animating the Results (2)
 Use "Animate Using Output File" to run the animation using

the data stored in the output
– To use this feature, all pipe station output needs to be saved in
Transient Control, which can dramatically increase the output file
size for some models
– Gives more flexibility to control animation
• Time steps can be skipped
• The animation can be started at any point in the simulation
• Can improve animation performance
I7 - 45
Can be recorded and Select the pipes
saved to a *.mp4 or in the flow path
*.wmv movie file type
Control the animation
speed
Choose an
The current animation option
simulation time is
shown in the
callout
I7 - 46
 Multiple animations can be shown at the same time
I7 - 47
Overview of Seminar
INTRODUCTION
Transient Initiation
 Junction transients can be initiated in one of two ways:

– Time
– Events
 Time transients are specified relative to the simulation time
– Time transients are always initiated
 Event transients are specified relative to some criteria being
satisfied
– Event transients will not occur unless the criteria is satisfied
I8 - 1
Time Transients
 To specify, select Time in the Initiation of Transient area

 Time transients are relative to the Start and Stop time in
Transient Control
 Time transients always act based on the time specified
 Time transients are the default
I8 - 2
Time Transients (2)
 Here it is specified that:

– The valve state remains
unchanged for 0.5
seconds
– The valve partially closes
over 1.5 seconds
– The valve remains at the
new state (Cv = 20) from
2 seconds and on
I8 - 3
Event Transients
 Event transients can be one of three types:

– Single event
– Cyclic dual events
– Sequential dual event
 Event transients do not act unless and until some criteria is
met
I8 - 4
Single Event Transient
 To specify, select Single Event in the Initiation of Transient

area
 Specify the event criteria
– Specify the Event Type
• This is some parameter such as pressure, velocity, pump speed or
valve position
• There are over fifteen choices
– Specify the event value and location within the system
• This will be at a particular pipe or junction
 Time Zero in the Transient Data is relative to the time when
the criteria is first met
I8 - 5
Single Event Transient (2)
I8 - 6
Cyclic Dual Event Transients
 To specify, select
Dual Event Cyclic in
the Initiation of
Transient area
 There are two criteria
to specify (First Event
and Second Event)
I8 - 7
Cyclic Dual Event Transients (2)
 The first transient will initiate when the first event criteria is
satisfied
 Once the first transient is initiated, the second event criteria
becomes active
– The Second Transient will initiate when the Second Event criteria
is met
 The junction will cycle between the two transients
I8 - 8
Sequential Dual Event Transients
 To specify, select Dual

Event Sequential in the
Initiation of Transient
area
 This type of transient
functions the same as
Cyclic Dual Event,
except it never cycles
back to the first transient
I8 - 9
Junctions With Inherent Event Logic
 There are three junction types in AFT Impulse which have

built-in, or inherent, event logic
– Check valve
– Relief valve
– Vacuum breaker valve
 The user does not need to specify the nature of the events,
and in fact is not allowed to
 The inherent event logic is very similar to the Dual Event
Cyclic logic
I8 - 10
Event Messages
 When events occur, they are logged to the Event Message

reports in the Output window
– There are two reports: one sorted by junction number, and the
other sorted by time
I8 - 11
Thought Experiment: A Pseudo-Check
Valve
 One can make a regular valve with a cyclic dual event behave
just like a check valve
Valve closes on negative flow Valve opens with sufficient pressure difference
I8 - 12
I9. Steady and Transient Special
Conditions
Overview of Seminar
INTRODUCTION
Why Use Special Conditions?
 There are both steady-state and transient special conditions

– Steady-state special conditions are just called "Special
Conditions"
 Transient Special Conditions are usually just for convenience
– Avoids having to delete transient data already entered
 In two cases, there is a special use of Transient Special
Conditions which simulate an "infinite pipe"
– This is discussed later in this section
I9 - 1
Why Use Special Conditions? (2)
 There are two main reasons for using steady-state special

conditions:
– First is for modeling convenience
• When you want to run a simulation with a valve closed, rather than
delete the valve junction and add a Dead End junction, you can just
turn the valve off and leave it in the model
• It will behave just like a Dead End
– Second is for numerical reasons
• Steady-state solution algorithms do not work very well with zero
flows
• Using a Special Condition allows the Impulse Steady-State Solver
to skip the flow solution for certain pipes and junctions and just set it
to zero
• This is very helpful for convergence
I9 - 2
Special Conditions
 Special Conditions relate to the initial steady operation

 Common uses are to close valves and turn off pumps
 Can be specified on the junction Optional tab, Edit menu or
toolbar
 Vacuum breaker valves,
surge tanks, gas accumulators,
relief valves and liquid
accumulators can be ignored
using the Special Condition to
simplify the location of surge
suppression equipment
I9 - 3
Special Conditions with Transients
 When modeling transients, Special Conditions relate to the

initial state of the system
 Special Conditions must be matched properly with the initial
data of the transient
 If a Special Condition is set and there is no transient data,
then the Special Condition remains in effect for the entire
transient
I9 - 4
Special Conditions with Valve Transients
 Here, the valve is initially closed

 At time zero the Cv value must be zero to be consistent with
the initial closed state
– Impulse checks for
consistency such as this
I9 - 5
Special Conditions with Pump Transients
 For pumps, Special Conditions will be for pump startup

transients
I9 - 6
(2)
 The "Pump Off With Through Flow" is usually for centrifugal
pumps with a pump curve entered
 The transient has an initial speed of zero
– Here the "Without Inertia"
model is used
I9 - 7
(3)
 The "Pump Off With Through Flow" is also used for inertia
modeling using four quadrant pump startups
– Here the "Startup - Four Quadrant
Known Speed" model is used
I9 - 8
(4)
 The "Pump Off With No Flow" is usually for positive
displacement pumps
 The transient has an initial flow of zero
I9 - 9
Transient Special Conditions
 Transient Special Conditions are specified on the junction

Transient Data tab
 The default Transient Special Condition is None
 The most frequent use for a Transient Special Condition is to
"Ignore the Transient Data"
 The Assigned Flow and Assigned Pressure junction support a
third type called "No Reflection - Infinite Pipe"
I9 - 10
No Reflection - Infinite Pipe
 Normally Assigned Flow junctions fix the flow during steady-

state and transient calculations
– The transient flow rate can be varied with time
– The Assigned Pressure works identically but with
pressure
 When "No Reflection - Infinite Pipe" is chosen,
then the steady-state flowrate or pressure is not
fixed during the transient
– Both the flowrate and pressure vary as the
junction absorbs all transient waves while reflecting none
 Proper use of "No Reflection - Infinite Pipe" can significantly
reduce model runtimes that have long pipelines with no
compromise on accuracy
I9 - 11
No Reflection - Infinite Pipe (2)
 The conditions for using "No Reflection - Infinite Pipe" are as

follows:
– The waves that travel into the pipe will not reflect during the time
frame of the simulation
– The pipe will have no nearby reflecting points such as:
• Valves
• Pumps
• Tees
• Diameter changes
• Material or wall thickness changes, or anything which changes the
wavespeed
 A minimum of two pipe sections should be used in the pipe
that is connected to the infinite pipe junction
(i.e., two times the controlling pipe length)
I9 - 12
Other Special Features at Junctions
 Three junction types can behave in a unique way for steady-

state
– Surge Tank
– Gas Accumulator
– Assigned Flow
 During steady-state, surge tanks and gas accumulators by
default behave exactly like branch junctions
– Mass is balanced at these junctions and thus no net
inflow/outflow exists
I9 - 13
Other Special Features at Junctions (2)
 There may be occasions where one wants these junctions to

act like a pressure-type junction during steady-state
– A good example is when the user wants to track the liquid level
in a topping tank
 One can do this with features on the Optional tab
– For the Surge Tank this has been moved to the main Tank Model
tab
 Users can input the surge tank steady-state liquid height
– Mass is no longer balanced during steady-state
 For gas accumulators users can input the steady-state
pressure
– This only works for cases where there is no net inflow
I9 - 14
Other Special Features at Junctions (3)
 Normally for assigned flow junctions, if the steady-state flow is

zero (using a Special Condition) then the junction acts like a
dead end
– Impulse solves for the local pressure
 The assigned flow junction can also be made to act like an
assigned pressure junction
– This allows a pressure to be input at the location, and then the
flow rate changed during the steady-state
I9 - 15
Notes
I9 - 16
US
Hands-On Model: Pump Startup With
Events
 Find the maximum pressure in a water transfer system for the
following three cases:
– Both pumps start
– One pump starts while the other remains off (assume J2 starts)
– One pump starts while the other is already operating at full
speed and valve is open (assume J5 starts)
– The complete solution is saved in the model file "Pump Startup
With Event Transient (complete).imp”
I9 - 17
US
Events (2)
 Build the water transfer system or use the pre-built example
and add the transient data
– If building from scratch:
• Use the same pipe and jct numbering scheme
• Use all the pipe and junction data on following pages
– If using pre-built model:
• Open the model file "Pump Startup With Event Transient.imp“
• Skip ahead to the “Transient Data” pages that follows
I9 - 18
US
Events (3)
Pipe and Junction Data:
I9 - 19
US
Events (4)
Pipe and Junction Data (cont.):
 Water temperature is 60 F
 All pipes are Steel - ANSI with standard wall thickness
 J2 and J5 pump curve data are:
Q (gpm) H (ft)
0 150
1000 140
2000 120
 J3 and J6 have Cv = 1000 when fully open

 Pipe depth at J1 is 20 feet
 Pipe depth at J10 and J11 are 10 feet
I9 - 20
US
Events (5)
Transient Data:
 Pumps J2 and J5 startup speed transient is linear over two
seconds
– Use the "Without Inertia" model where speed is an input
Time (sec) Speed
0 0
2 100
10 100
I9 - 21
US
Events (6)
Transient Data (cont.):
 When pump J2 is off then valve J3 is closed to prevent
backflow
backflow
 When the J2 pump starts, the J3 valve opens over two
seconds with the following transient (the J5 pump and J6
valve behave the same)
Time (sec) Cv
0 0
1 800
2 1000
10 1000
I9 - 22
US
Events (7)
 The J9 valve is closed until the stagnation pressure at the J7
branch reaches 45 psig or greater, at which point it starts to
open
– Hint: There is no "Junction Pressure" event type, but there is one
for pipes. Use one of the pipes connected to the branch as a
basis for the event
– Full open Cv = 500
– The J9 valve has the following opening profile:
Time (sec) Cv
0 0
2 400
5 500
10 500
I9 - 23
US
Events (8)
 Use 1 section in the controlling pipe, and run the simulation
for 10 seconds
 Ignore cavitation in Transient Control
I9 - 24
US
Results: Pump Startup With Events
Case Max Stag. Pressure (psig)
Two Pumps Start 72.8 (at pump discharge)
One Pump Start 71.6 (at pump discharge)
One Pump Start With One On 73.3 (in pipe P11)
Note: Results differ slightly if you use variable pipe resistance as was done here
I9 - 25
US
 Note that the case with one pump starting while the other is
running causes the maximum transient in pipe P11
 Also note that there are sub-atmospheric pressures in this
case
I9 - 26
US
Note the Event Messages
I9 - 27
Notes
I9 - 28
US
Hands-On Model: Gasoline Tank Farm
 Gasoline flows into a

tank farm
– The gasoline is flowing
into Desert Petroleum's
tank (J102)
– The transient occurs
when the flow is
redirected to West
Coast Energy's tank
(J202)
– To achieve this, valve
J201 is opened as J101
is closed Model file on disk is: "Tank Farm - Using Infinite Pipes.imp"
I9 - 29
US
Hands-On Model: Gasoline Tank Farm (2)
 Plot the flowrates and pressures at the valve inlets

 How much gasoline flows into Desert Petroleum's tank during
the transient?
– Hint: There are two ways to do this:
• You can use Integration Query
• You can plot integrated values and then save the x-y data to file
I9 - 30
US
 All pipes are Steel - ANSI, schedule 40

 The restraint for each pipe should be "Thick-Walled Anchored
Throughout"
 Gasoline properties are:
– Specific Gravity = 0.759
– Dynamic Viscosity = 0.000321 lbm/s-ft
– Bulk Modulus = 150,000 psia
– Vapor Pressure = 9 psia
 J1 steady-state flowrate is 13,500 barrels/hr
 Tank J102 is initially full, with the pipe at a depth of 40 feet
 Tank J202 is initially empty, with the pipe at a depth of 0 feet
I9 - 31
US
 The J101 valve has a full open Cv = 15,700
 J101 closes over 34 seconds with the following profile
time (sec) Cv
0 15700
19 15700
19.2 12400
22 12400
22.2 9400
26 9400
28 200
34 0
 The J201 valve opens over 12 seconds with the following

profile time (sec) Cv
0 0
10 150
11 800
12 4000
30 4000
 Run the simulation for 35 seconds I9 - 32

US
Results: Gasoline Tank Farm
I9 - 33
US
Results: Gasoline Tank Farm (2)
 Total flow through pipe 102, inlet (station 0) yields total flow
into tank
 Using Integration Query on Tools menu yields: 86.14 barrels
I9 - 34
US
Hands-On Model: Gasoline Tank Farm -
Part II
 Use an infinite pipe junction at J1 and rerun model and plot
results
 The length of pipe P1 should be shortened
– Use two pipe sections so the length is twice the controlling pipe
 What was the difference in run time?
I9 - 35
US
Results: Gasoline Tank Farm - Part II
 Results should be almost identical to previous model

– Our experience for this model showed a difference of a factor of
over 60
– This could be the difference between a one-minute and one-hour
run
I9 - 36
SI
Events
 Find the maximum pressure in a water transfer system for the
following three cases:
– Both pumps start
– One pump starts while the other remains off (assume J2 starts)
– One pump starts while the other is already operating at full
speed and valve is open (assume J5 starts)
– The complete solution is saved in the model file “Pump Startup
With Event Transient (SI) (complete).imp”
I9 - 37
SI
Events (2)
 Build the water transfer system or use the pre-built example
and add the transient data
– If building from scratch:
• Use the same pipe and jct numbering scheme
• Use all the pipe and junction data on following pages
– If using pre-built model:
• Open the model file "Pump Startup With Event Transient (SI).imp“
• Skip ahead to the “Transient Data” pages that follows
I9 - 38
SI
Events (3)
Pipe and Junction Data:
I9 - 39
SI
Events (4)
Pipe and Junction Data (cont.):
 Water temperature is 16 C
 All pipes are Steel - ANSI with standard wall thickness
 J2 and J5 pump curve data are:
Q (m3/hr) H (m)
0 46
230 43
450 37
 J3 and J6 have Cv = 1000 when fully open

 Pipe depth at J1 is 6 meters
 Pipe depth at J10 and J11 are 3 meters
I9 - 40
SI
Events (5)
Transient Data:
 Pumps J2 and J5 startup speed transient is linear over two
seconds
– Use the "Without Inertia" model where speed is an input
Time (sec) Speed
0 0
2 100
10 100
I9 - 41
SI
Events (6)
backflow
backflow
 When the J2 pump starts, the J3 valve opens over two
seconds with the following transient (the J5 pump and J6
valve behave the same)
Time (sec) Cv
0 0
1 800
2 1000
10 1000
I9 - 42
SI
Events (7)
 The J9 valve is closed until the stagnation pressure at the J7
branch reaches 310 kPa(g) or greater, at which point it starts
to open
– Hint: There is no "Junction Pressure" event type, but there is one
for pipes. Use one of the pipes connected to the branch as a
basis for the event
– Full open Cv = 500
– The J9 valve has the following opening profile:
Time (sec) Cv
0 0
2 400
5 500
10 500
I9 - 43
SI
Events (8)
 Use 1 section in the controlling pipe, and run the simulation
for 10 seconds
 Ignore cavitation in Transient Control
I9 - 44
SI
Case Max Stag. Pressure (kPa(g))
Two Pumps Start 501.0 (at pump discharge)
One Pump Start 492.5 (at pump discharge)
One Pump Start With One On 504.7 (in pipe P11)
Note: Results differ slightly if you use variable pipe resistance as was done here
I9 - 45
SI
Results: Pump Startup With Events (2)
 Note that the case with one pump starting while the other is
running causes the maximum transient in pipe P11
 Also note that there are sub-atmospheric pressures in this
case
I9 - 46
SI
Note the Event Messages
I9 - 47
Notes
I9 - 48
SI
Hands-On Model: Gasoline Tank Farm
 Gasoline flows into a

tank farm
– The gasoline is flowing
into Desert Petroleum's
tank (J102)
– The transient occurs
when the flow is
redirected to West
Coast Energy's tank
(J202)
– To achieve this, valve
J201 is opened as J101
is closed Model file on disk is: "Tank Farm - Using Infinite Pipes (SI).imp"
I9 - 49
SI
 Plot the flowrates and pressures at the valve inlets

 How much gasoline flows into Desert Petroleum's tank during
the transient?
– Hint: There are two ways to do this:
• You can use Integration Query
• You can plot integrated values and then save the x-y data to file
I9 - 50
SI
 All pipes are Steel - ANSI, schedule 40

 The restraint for each pipe should be "Thick-Walled Anchored
Throughout"
 Gasoline properties are:
– Specific Gravity = 0.759
– Dynamic Viscosity = 0.47770 cp
– Bulk Modulus = 10342.1 bar
– Vapor Pressure = 0.6205 bar
 J1 steady-state flowrate is 2146.33 m3/hr
 Tank J102 is initially full, with the pipe at a depth of 12 meters
 Tank J202 is initially empty, with the pipe at a depth of 0
meters
I9 - 51
SI
 The J101 valve has a full open Cv = 15,700
 J101 closes over 34 seconds with the following profile
time (sec) Cv
0 15700
19 15700
19.2 12400
22 12400
22.2 9400
26 9400
28 200
34 0
 The J201 valve opens over 12 seconds with the following

profile time (sec) Cv
0 0
10 150
11 800
12 4000
30 4000
 Run the simulation for 35 seconds I9 - 52

SI
Results: Gasoline Tank Farm
I9 - 53
SI
Results: Gasoline Tank Farm (2)
 Total flow through pipe 102, inlet (station 0) yields total flow
into tank
 Using Integration Query on Tools menu yields: 13.66 meters3
I9 - 54
SI
Hands-On Model: Gasoline Tank Farm -
Part II
 Use an infinite pipe junction at J1 and rerun model and plot
results
 The length of pipe P1 should be shortened
– Use two pipe sections so the length is twice the controlling pipe
I9 - 55
SI
Results: Gasoline Tank Farm - Part II
 Results should be almost identical to previous model

– Our experience for this model showed a difference of a factor of
over 60
– This could be the difference between a one-minute and one-hour
run
I9 - 56
Overview of Seminar
INTRODUCTION
Steady-State Pump Modeling
 Steady-state pump operation can be modeled in three basic

ways:
– Pump curve
• Useful for centrifugal pumps
• Possibly with a user specified operating speed
– Assigned flow
• Useful for positive displacement pumps
– Pump with variable speed controller
• Used in conjunction with a pump curve
• Useful for centrifugal pumps with VFD's
– If no pump transient is being modeled, the above three steady-
state pump models are still applicable during system transients
– If a pump transient is being modeled, each of
the three steady-state models have
different aspects
I10 - 1
Pump Curves and Speed Changes
 When using a pump curve,

the effect of pump speed
changes must be
accounted for
 Pump curves can be
adjusted for operation at
different speeds using
affinity laws
– Also called homologous
pump laws
2
Q1 n1 ∆H1  n1 
= =  
Q2 n2 ∆H 2  n2 
I10 - 2
Pump Curves and Speed Changes (2)
 The simplest model is for

the user to input the speed
transient
– No other information about
the pump is needed
– This model is called
"Without Inertia"
 The pump speed can
frequently be estimated for
startup cases
 For pump trips, it is difficult 2
Q1 n1 ∆H1  n1 
to determine the speed = = 
Q2 n2 ∆H 2  n2 
I10 - 3
Pump Curves and Speed Changes (3)
 Pumps with user defined speed transients can be time or

event based
 For pump startups, the initial transient speed will be zero
– This is used in conjunction with the "Pump Off With Through
Flow" Special Condition
– Can be time or event based
 For pump trips, the initial transient speed will be the same as
the steady-state
– Can be time or event based
I10 - 4
US
Model Pump Trip
 Determine the maximum and minimum transient pressures

during a pump trip for the following system (TEST2.IMP)
– Water system at 70 degrees F
– Reservoir at 10 feet elevation needs to be pumped up a hill to a
reservoir at 200 feet elevation (assume pipes are at a 10 feet
depth)
– The total pipe length is 1000 feet (assume pump is 10 feet from
supply)
– The pipe is 4 inch Schedule 40 Steel-ANSI
– Pump elevation = 0 feet
– Run the model for 30 seconds
I10 - 5
US
Model Pump Trip (2)
 Pump data:
– Use the transient pump model “ Without Inertia” and enter the
pump speed profile given below
– Add a check valve in the Pump Properties that closes when the
forward velocity reaches zero
Q (gpm) Head (ft) Time (s) Speed (%)
0 400 0.0 100
250 390 0.5 40
500 380 2 20
1000 340 5 15
30 5
 Save this model to disk as we will revisit it later
I10 - 6
US
Answers to Pump Trip Model
 TEST2 results
– Max pressure is 163.3 psig at pump discharge
– Min pressure is -2.2 psig somewhere in pump discharge line
(Station 89)
I10 - 7
Notes
I10 - 8
SI
Model Pump Trip

during a pump trip for the following system (TEST2 (SI).IMP)
– Water system at 21 degrees C
– Reservoir at 3 meters elevation needs to be pumped up a hill to
a reservoir at 60 meters elevation (assume pipes are at a 3
meter depth)
– The total pipe length is 305 meters (assume pump is 3 meters
from supply)
– The pipe is 4 inch (10.2 cm ID) Schedule 40 Steel-ANSI
– Pump elevation = 0 meters
– Run the model for 30 seconds
I10 - 9
SI
Model Pump Trip (2)
 Pump data:
– Use the transient pump model “ Without Inertia “ and enter the
pump speed profile given below
Q (m3/hr) Head (m) Time (s) Speed (%)
0 122 0.0 100
57 119 0.5 40
114 116 2 20
227 104 5 15
30 5
 Save this model to disk as we will revisit it later
I10 - 10
SI
Answers to Pump Trip Model
 TEST2 (SI) results

– Max pressure is 1126 kPa(g) at pump discharge
– Min pressure is -14.4 kPa(g) somewhere in pump discharge line
(Station 90)
I10 - 11
Calculated Speed Changes Using Inertia
 For more accurate pump trip modeling, the effects of pump

inertia must be included
– Transient pump speed becomes an output parameter, rather
than an input
– Additional pump data is required to use these models
 AFT Impulse offers two transient models for pump trips:
– Pump Trip With Inertia and No Backflow or Reverse Speed
– Pump Trip With Inertia - Four Quadrant
 Pumps can be tripped or started based on events
– Pumps can trip and then restart, and vice versa
 Pumps have an anti-reverse rotation option
I10 - 12
Calculated Speed Changes Using Inertia
(2)
 If there is a check valve to prevent backflow, the best choice
is the first one: "Pump Trip With Inertia and No Backflow"
– This model requires the following:
• Pump curve and power vs. flow data
• Rated pump speed in rpm
• Inertia of pump, entrained water and motor
I10 - 13
Pump Trip With Inertia and No Backflow
 Calculation methodology
dω
T = −I T = unbalanced torque
dt
I = inertia
P = Tω ω = rotational speed
1 dω t = time
∫ dt = − ∫
I T P = power
∆t 1
= − ∆ω
I T
∆
ωnew = ωold − Told t
I
 Once we know the new speed, we plug this into the same
affinity law routines as we do when the speed is specified by
the user
I10 - 14
Pump Trip With Inertia and No Backflow
(2)
 The user enters the power vs. flow, the rated pump speed in
rpm, and the estimated rotating inertia
I10 - 15
Estimating Rotating Inertia
 There are two parts to the rotating inertia

– Pump and entrained water
– Motor
 Wylie provides equations to estimate the inertia*
0.9556 1.48
7
𝑃𝑃 𝑃𝑃
𝐼𝐼𝑝𝑝 = 1.5 10 𝐼𝐼𝑚𝑚 = 118
𝑁𝑁 3 𝑁𝑁
 These equations have been incorporated into a window that
may be accessed from the Transient tab on the Pump
Property window
 Note Europeans often base inertia on diameter (rather than
radius) and values differ by a factor of four from above value
* Wylie, E.B., V.L. Streeter & L. Suo, Fluid Transients in Systems,

Prentice Hall, Englewood Hills, New Jersey, 1993. See page 148-149.
I10 - 16
Pump Inertia and Specific Speed
Estimates
 If an inertial Pump Transient model is selected, the “Estimate
Rotating Inertia and Calculate Specific Speed” window may
be used to estimate the pump inertia and specific speed
– Click the “Estimate Inertia and Specific Speed…” button located
on the Inertia Data tab
I10 - 17
Estimates (2)
 Enter pump rated
parameters
 Pump rated speed
value is imported
automatically from the
Pump Property
window
 The value may also
be entered manually
I10 - 18
Estimates (3)
 The rated head, flow,
and the BEP can be
calculated, if the pump
head curve, and the
efficiency or power
curves have been
entered on the Pump
Properties window
 If the pump data is not
available, the values
are entered manually
I10 - 19
Estimates (4)
 By default, the density
is set to the same
density value used in
System Properties
 By unchecking the
box, a different
density value may be
entered
I10 - 20
Estimates (5)
 After the rated values
are entered, click the
“Calculate” button
– Calculates the
estimated rotating
inertia values
– Calculates the pump
specific speed
I10 - 21
Estimates (6)
 The Calculated Values
section displays the
results calculated from
Wylie’s equations
– Power
– Pump Rotating
Inertia
– Motor Rotating
Inertia
I10 - 22
Estimates (7)
 The Results section
displays additional
calculated results
– Total Rotating Inertia
• Sum of Pump and
Motor inertia
• Required for inertial
pump transient
calculations
– Specific Speed
• Displayed in three
different units
• Used to select four
quadrant data set
I10 - 23
Estimates (8)
 The results can be
transferred back to the
Pump Properties
window
– Make appropriate
transfer selections
– Click “Transfer and
Close” button
I10 - 24
Estimates (9)
 The pump total
rotating inertia value
is transferred directly
to the Pump
Properties window
 The pump specific
speed value is used
to select the most
appropriate four
quadrant data set
I10 - 25
Four Quadrant Modeling
 When pumps can spin and flow backwards, then four

quadrant methods are required
 The four quadrants are:
– Positive flow, positive speed
– Positive flow, negative speed
– Negative flow, positive speed
– Negative flow, negative speed
 Conventional methods use the "Suter Method"
– For a detailed description see Wylie, et al. (*)
* Wylie, E.B., V.L. Streeter & L. Suo, Fluid Transients in

Systems, Prentice Hall, Englewood Hills, New Jersey, 1993.
See page 144-151.
I10 - 26
Four Quadrant Modeling (2)
 Four quadrant data is usually not available from pump

manufacturers
– They only take in the first quadrant
 Impulse provides twenty-one four quadrant data sets
 Users of these methods rely on data for pumps with similar
specific speed
 Four quadrant methods make use of four parameters
– Pump head rise
– Pump flow rate
– Pump torque
– Pump speed
I10 - 27
 The head rise and flow rate across the pump are available
from the compatibility equations (we saw these in Section 3)
Pi , new = C P − B P m
 i , new
Pi , new = C M + B M m  i , new
 This provides a way to solve for two of the four variables

 We need two more pieces of data
 Four quadrant data provides correlated/measured data for
pump torque and speed in all four quadrants, and how these
relate to flow rate and head
I10 - 28
 The Suter method changes the four variables into

dimensionless form by dividing through by the rated value
α= N
NR
β= T
TR
H
h=
HR
Q
v=
QR
 Two combined parameters are defined (called WH and WB in
Wylie) h
FH =
α 2 + v2
β
FB =
α 2 + v2
I10 - 29
 FH and FB are specified, based on correlated data, and

increments of θ where:
ϑ = π + tan−1 v
α
 Some Suter method

users define θ as:
α
ϑ = tan−1
v
 AFT Impulse allows the

user to choose which
definition to use
I10 - 30
 If the four quadrant data (FH and FB) are available, they can
be entered into the table
– The user would set the
"Four Quadrant Data
Source" as "User Specified"
 More typically, one of the
predefined data sources
can be used using the
nearest specific speed (Ns)
I10 - 31
 With data for FH and FB and the two compatibility equations,

there are four equations and four unknowns
 These are solved to obtain the new head, flow rate, torque
and speed at each time step
I10 - 32
Partial Four Quadrant Data
100% Speed:
Standard
Operating
Zone
Rated
Operating
Point
I10 - 33
Four Quadrant Pump Head Curves
 AFT Impulse allows users to enter manufacturer head and

power (or efficiency) curves on the Pump Model tab
– Used for steady state calculations
– These curves are no longer used once the transient calculation
begins
I10 - 34
Four Quadrant Pump Head Curves (2)
 After the transient begins, AFT Impulse uses the head curve
derived from the four quadrant pump data
I10 - 35
Four Quadrant Pump Head Curves (3)
 These two different curves are rarely, if ever, identical

– This means the pump operates on a different curve during the
transient than it does during the steady state
– If an appropriate four quadrant data set was selected, generally
the curves are very similar
 The four quadrant head curves are typically developed based
on the pump’s rated operating point (usually BEP)
 Using different curves in steady state and transient
calculations creates some potential difficulties
I10 - 36
Issues When Using Different Curves
 Pump curve and four quadrant curve based on pump BEP

– In this case, the curves are very similar
– The curves will always meet exactly at the BEP
I10 - 37
Issues When Using Different Curves (2)
 The pump may operate away from BEP during Steady State
– The head predicted by the different curves is different
– When the solution is handed off to the transient solver, an
artificial transient will likely occur
•
I10 - 38
 Impulse 4 handled this by developing the four quadrant pump

curve based on the steady state solution rather than the pump
BEP
– This eliminated the artificial transient by ensuring the steady
state pump curve and the transient pump curve always produced
the same head at the steady state condition
– This tends to work fairly well if the steady state pump operation
is near BEP
– However, the more the steady state solution varies from the BEP,
this method tends to force bigger deviations between the steady
state pump curve and the four quadrant pump curve
I10 - 39
 As the steady state operation moves away from the BEP, the
four quadrant curves are forced to meet the pump curve at the
operating point
– Steady state and transient curves tend to deviate more
I10 - 40
Four Quadrant Pump Curves
 Impulse allows the use of the four quadrant pump curve

during both the steady state and transient solutions
– Eliminates the artificial transients associated with using different
pump curves
– Eliminates the “warping” of the four quadrant curve to meet
steady state operating conditions away from the BEP
 The four quadrant pump curve used in the steady state is
developed using the BEP of the user entered pump data
– User must still enter a pump head and power (or efficiency)
curves so the BEP can be determined
– A user defined BEP point can be entered on the Pump
Configuration window
I10 - 41
Four Quadrant Pump Curves (2)
 The user may select to use the four quadrant pump curve in
the steady state on the Pump Properties window Transient tab
I10 - 42
Four Quadrant Pump Curves (3)
 The four quadrant pump curves can also be compared to the

manufacturer’s curves by clicking the “Compare Curves”
button
I10 - 43
Pump Startups With Four Quadrant Data
 The four quadrant data can also be used for pump startups
 Impulse offers two four quadrant models for pump startup
1) Four quadrant with known speed
• Here the speed is an input
• The four quadrant data is used to calculate the pump response
2) Four quadrant with inertia and known motor torque/speed
• Motor torque/speed is used to calculate the pump speed
I10 - 44
Pump Startups Using Motor Torque and
Speed
 Impulse offers two inertial startup models
– Startup with inertia and no back flow or reverse speed
– Startup with inertia - four quadrant with inertia and known
torque/speed
 For these cases the motor torque is entered by the user as a
function of speed
 When four quadrant data is not available, the user provided
pump data for head vs. flow and power vs. flow are used in
the calculations, and thus calculation can only be made for
positive or zero flows (i.e. a check valve is present)
 The addition of the four quadrant data allows calculations for
back flow or reverse speed
I10 - 45
Pump Startups Using Motor Torque and
Speed (2)
Typical motor
speed/torque data
For models using pump startup

with inertia models, the user
supplies the speed/torque data
I10 - 46
No Transient – Four Quadrant
 In some cases a pump with no transient will experience

backflow due to conditions in the system
– In such cases the second quadrant data will be assumed to have
a head equal to the shutoff head
– This can cause convergence issues and is also not very
accurate
 The “No Transient – Four
Quadrant” option can be used
to model this situation
I10 - 47
Modeling Pumps With Control Transients
 When using a pump curve and flow (or pressure) controller,

the control point can be varied with time
– The resulting speed is calculated over time
I10 - 48
Modeling Pumps With Flow Transients
 Pumps with flow transients can be used to model positive

displacement pumps
– The "Repeat Transient" feature is useful in modeling the periodic
flow rate
I10 - 49
Modeling Multiple Pumps
 AFT Impulse allows you to specify that a pump represents

one of a set of parallel or series pumps
 This simplifies entering extra piping and dealing with the
impact on run time
 This only works when all pumps start, trip or in general
operate together
I10 - 50
Notes
I10 - 51
Notes
I10 - 52
US
Hands-On Model: Pump Trip
 Build models shown below and compare results from two

pump trip models ("Pump Trip With Multiple Pumps.imp")
– The two models will be identical except for the pump transient
model
I10 - 53
US
Hands-On Model: Pump Trip (2)
 Fluid is water at 60 degrees F

 The pipes at the inlet reservoirs are 200 feet deep
 The pipes at the outlet reservoirs are 140 feet deep
 All pipes are 30 inch STD Steel - ANSI
 There are four parallel pumps in each system
 The data for an individual pump is:
– Rated Speed = 1760 rpm
– Rotational Inertia = 475 lbm-ft2
– Performance curve:
Q (gpm) H (ft) Power (hp)
0 408 165
1000 384 183
2000 351 225
3000 288 261
4000 189 279
5000 72 285 I10 - 54
US
 For J2, choose "Trip With Inertia and No Backflow or Reverse

Speed" for the transient model
 For J12, choose "Trip With Inertia - Four Quadrant" transient
model
– Choose the Four Quadrant Data Source as Ns = 1270 in gpm
units
– On the Pump Model tab, specify that a Check Valve is at the
discharge which closes when the velocity is zero and reopens at
zero delta pressure/head
 Section the pipes with 4 sections in the controlling pipe
 Run the models and compare the pump speed decay, pump
discharge pressure transients, pump flow rates transients,
and overall pressure profile
– Run the models for 60 seconds I10 - 55
Notes
I10 - 56
US
Results: Pump Trip
 Results are very similar
I10 - 57
US
Results: Pump Trip (2)
I10 - 58
Notes
I10 - 59
Notes
I10 - 60
SI
Hands-On Model: Pump Trip
 Build models shown below and compare results from two

pump trip models ("Pump Trip With Multiple Pumps (SI).imp")
– The two models will be identical except for the pump transient
model
I10 - 61
SI
 Fluid is water at 16 degrees C

 The pipes at the inlet reservoirs are 61 meters deep
 The pipes at the outlet reservoirs are 43 meters deep
 All are 30 inch (74.3 cm ID) STD Steel - ANSI
 There are four parallel pumps in each system
 The data for an individual pump is:
– Rated Speed = 1760 rpm
– Rotational Inertia = 20.016 kg-m2
– Performance curve:
Q (m3/hr) H (m) Power (kW)
0 124 123
227 117 136
454 107 168
681 88 195
908 58 208
1136 22 212 I10 - 62
SI
 For J2, choose "Trip With Inertia and No Backflow or Reverse

Speed" for the transient model
 For J12, choose "Trip With Inertia - Four Quadrant" transient
model
– Choose the Four Quadrant Data Source as Ns = 0.46
– On the Pump Model tab, specify that a Check Valve is at the
discharge which closes when the velocity is zero and reopens at
zero delta pressure/head
 Section the pipes with 4 sections in the controlling pipe
 Run the models and compare the pump speed decay, pump
discharge pressure transients, pump flow rates transients,
and overall pressure profile
– Run the models for 60 seconds
I10 - 63
Notes
I10 - 64
SI
Results: Pump Trip
 Results are very similar
I10 - 65
SI
Results: Pump Trip (2)
I10 - 66
I11. Pipe Sectioning and Transient
Control
Overview of Seminar
INTRODUCTION
Pipe Sectioning Can Impact Runtime
 The run time for a simulation is related directly to the total

number of computations required
– The number of computations depend on the total number of time
steps (shown on the Transient Control window) and,
– The total number of pipe sections (which can be determined
from the display table in the Section Pipes window)
 Run times also depend to a lesser degree on how frequently
data is written to disk and which stations are written (Transient
Control window)
I11 - 1
Pipe Sectioning Can Impact Runtime (2)
 Using double the number of sections results in four times the

number of computations
– This occurs because the time step must also be halved
 If the controlling pipe is very short, this requires more sections
in all other pipes in the model
– It may be prudent to either neglect the shortest pipe or identify a
way to combine it into a longer pipe
I11 - 2
Section Pipes Window
 The Section Pipes

window automates
the process of finding
the most suitable pipe
sectioning
 A range of values for
the number of
sections in the
controlling pipe can
be evaluated
– By default Impulse
searches up to 5
sections
I11 - 3
Section Pipes Window (2)
 The search is conducted by varying the wavespeed over a

range of choices and searching for the option that produces
the least amount of error
– The range is controlled by specifying the “bandwidth” of the
variation percentage. The default bandwidth is 25%.
– The search is repeated for each number of sections in the
controlling pipe specified in the search criteria
 The option with the least amount of error for each number of
sections in the controlling pipe is displayed in the Search
Results table
I11 - 4
 If the resulting error is

larger than the
maximum allowable
error, the selection is
colored red, and the
next higher option is
selected by default
– Maximum allowable
error is 10% by
default
– Results with error
above the maximum
allowable error can
still be selected
I11 - 5
 Detailed information
for the selected pipe
sectioning is
displayed in the
“Selected Sectioning
for Model” table
– Specific information
for each pipe in the
model is displayed
– The Sectioning
Details can be
hidden
I11 - 6
 By clicking “Show
Advanced Settings”
the search criteria
can be modified
– Maximum Allowable
Error
– Search Bandwidth
– Search increment
I11 - 7
Combining Pipes
 Pipes that are good candidates to be combined are ones that:

– Are connected by static elements such as an elbow or a valve
that is always open
– Have the same pipe size
 The Combine Pipes feature on the Section Pipes window will
automatically combine pipes that satisfy the criteria
I11 - 8
Example of Ignoring Shortest Pipe
 Here pipe P4, with a 36 inch diameter, is the controlling pipe

– Run time is over 1 hour
I11 - 9
Example of Ignoring Shortest Pipe (2)
 Here pipe P4 has been deleted and run time is less than a
second
– How do you think this will affect the results?
I11 - 10
Transient Output Files
 Transient models generate much more output data than

steady-state
– Calculations are performed at numerous time steps
 Ongoing access to this data is needed for several reasons
– Creating graphs of results
– Reviewing results of models run previously without having to
rerun them
– Comparing results of different runs
 To facilitate these activities, output data is written to a file
external to the *.imp file
I11 - 11
Transient Output Files (2)
 Such files have the following naming convention:

– "Input File Name" + # + ".out" where # is a number
• The output file is located in the same folder as the input file
– The number is needed to distinguish among different output files
generated from different scenarios
– Unless you want to move the models files to a different PC or
archive the results, the name of the output file is unimportant
– If you move both input and all output files, you can open the
models on other PC's and have access to all modeling results
I11 - 12
Transient Output File Size Issues
 Transient output files can get extremely large if all data is

saved
 Fortunately, it is usually not necessary to save all of the data
 AFT Impulse offers several features for reducing output file
size while retaining important data
I11 - 13
Transient Output File Size Issues (2)
 The file keeps the following data:

– At all or selected time steps:
• Pressure
• Mass flowrate
• Vapor volume
• Slurry information if the SSL module is used
– The maximum and minimum values for all time steps for all pipe
stations:
• Static pressure
• Stagnation pressure
• Mass flowrate
• Vapor volume
• Slurry information if the SSL module is used
I11 - 14
 The approximate output file size and model run time are
shown at the bottom of Transient Control
I11 - 15
 Output data for each time step exists

– All of this data can be written to file,
or only at
certain intervals
– Saving data at intervals is usually
acceptable
• Here data is written every 10th time
step
• This reduces the output file size by
nearly 90%
• It also can have a noticeable impact on model runtime
– It is recommended that data be saved at least every 25th time step
• In more complicated models where the output file
size is extremely large one may consider
skipping more data points
I11 - 16
 One can limit the number of pipe

stations written to the output file
– For systems with very long pipes, this
can reduce output file sizes by a huge
amount
 A good choice is to save only the Inlet
and Outlet stations
– If results indicate you need to save
more stations, you can change Transient Control and rerun
– The choice can be applied on a pipe by pipe basis
 When a pipe station is not saved, then it cannot be plotted
 Even if a station is not saved, the max/min values are saved
and can be included in profile plots
I11 - 17
Transient Control and Graph Results
If one chooses to save All Stations
to the transient output file...
…then all stations are shown in Select Graph Data
…and a transient graph can

be created for all stations
I11 - 18
Transient Control and Graph Results (2)
If one chooses to save Inlet and
Outlet to the transient output file...
…then only those two stations are
shown in Select Graph Data
…and a transient graph can be

created for only those two stations
I11 - 19
Transient Control and Graph Results (3)
However, if one chooses to save Inlet
and Outlet to the transient output file...
…then Max/Min profile graphs can be

specified in Select Graph Data that
include unsaved stations
…and a profile graph can be created
I11 - 20
Artificial Transients and Transient
Control
Impulse can stop the
run if artificial
transients are present
in the model by
selecting the check
box on the Transient
Control window
If necessary, the
criteria for detecting
artificial transients
can be adjusted
I11 - 21
I12. Special Topics Including
Troubleshooting
Overview of Seminar
INTRODUCTION
Getting the Right Results
 There are a number of modeling problems AFT sees

frequently
 This section offers numerous strategies and suggestions for
approaching modeling problems
I12 - 1
First Run the Model in Steady-State
 It is good practice to first run a model in Steady Only mode

before running the transient
I12 - 2
Plot Input Data in Graph Results
 The Graph Results window allows you to plot transient data

that should match input data
– Transient pressure conditions
– Valve Cv
– Pump speed
– Etc.
 This can assist with input data verification
I12 - 3
General Input Errors
 AFT Impulse offers significant help for identifying connection

errors, but data entry errors can be hard to find
 Finding input errors in pipe and junctions can be tedious
– Each object needs to be checked individually
 The Model Data Window lets you look at all pipes and junction
data in column format
– This is much easier to check than workspace objects
 Each column should be scanned down to see if there are any
unit errors or numbers that are wildly different than the others
– Example: all pipe lengths in miles but one is in feet
I12 - 4
Use the Sort Feature
 If a big model runs but the results look incorrect, use the Sort
feature in the Output Window to look for extremes in velocity,
pressure drop or other parameters
 An input error (like incorrect diameter) may be easier to see
by its effect on the results
I12 - 5
Tee/Wye Junctions Complications
 As mentioned earlier, tee/wye junctions are frequently

misused
– If tee losses have a small effect on this system, Branch junctions
or simple tee loss models may be preferred
 On some models with multiple Tee/Wye junctions the flow split
dependence causes convergence problems with the Relative
Tolerance Method
 In such cases the Absolute Tolerance method is better
 In some transient calculations, detailed Tee/Wye junctions
have been seen to produce results that appear to be
unrealistic
– In such cases using simple tees may be a
better alternative
I12 - 6
Check Valves
 Check Valves in AFT Impulse have inherent logic which

controls opening and closing
 Three types of check valve:
– Traditional (Without Inertia)
– Single Closure Based on Fluid Deceleration (Without Inertia)
– Swing Check (With Inertia)
I12 - 7
Without Inertia Check Valve Models
 Traditional check valve

– By default, AFT Impulse assumes the valve closure is
instantaneous
– This can lead to very large pressure surges
– This assumption may not be a realistic representation of the
actual valve closure
– User has the option of entering valve opening and closing
transients
• The actual opening and closing of the valve is controlled by the logic
in the junction itself
• The user can specify the amount of time it
takes for the valve to open and close
I12 - 8
Without Inertia Check Valve Models (2)
 Single Closure Based on Fluid Deceleration

– Attempts to predict the maximum reverse velocity at which the
valve will close
– This velocity is based on experimental predictions involving the
fluid deceleration and various check valve types from A R
Thorley’s Fluid Transients in Pipeline Systems (1981)
– The check valve does not reopen once it closes
– Only one check valve at a time in the model can use this option
I12 - 9
With Inertia Check Valve Model
 Swing Check
– Uses a torque balance to estimate valve position over time for a
swing check valve
– Requires more detailed input to describe the valve
• Physical properties of the valve, such as disk size and submerged
weight
• Inertia properties of the valve
• Spring data, if applicable
I12 - 10
Check Valve Special Condition
 Check valves can be set closed in steady state through

Special Conditions
– They can either stay closed or allowed to open as the transient
progresses
I12 - 11
Check Valve Transients (4)
 In this example, the downstream tank level starts higher than

the upstream, then lowers to below and raises back higher
– The check valve is set closed in steady state
– Then opens when the level drops low enough for forward flow
– Finally recloses as the level rises
I12 - 12
The Weir Junction
 A weir is a dam-like device that is used to measure flow and

to increase upstream water level (HGL)
 A weir can act like a seal pit which raises the outlet pressure
of the in-flowing pipe to avoid back flow and air intake
 Impulse can model three types of sharp-crested weirs
– Rectangular
– V-notch
– Cipolletti
I12 - 13
The Weir Junction (2)
 In Impulse a weir is an open channel device with an inlet and

outlet pipe
 There are three basic modeling aspects to define a weir
– An upstream area (bottom elevation and cross sectional area)
– A downstream area (bottom elevation and cross sectional area)
– The weir barrier (type, elevation and width)
 For steady-state the flow rates are equal in all three areas
– The flow entering the upstream area equals the flow over the
weir and leaving the downstream area
 During the transient the inlet and outlet flow rate can be
different
– The upstream and downstream levels can vary
I12 - 14
The Weir Junction Example
 The discharge from a holding tank to a reclaim pond is

controlled by a valve and a rectangular weir
 The valve is closed rapidly over 2 seconds and fully closed
within 3 seconds
 Determine the liquid levels in the upstream and downstream
areas over time
*Reference “Weir Example.imp” on seminar disc

I12 - 15
The Weir Junction Example (2)
 The liquid levels over time
I12 - 16
The Turbine Junction
 The AFT Impulse turbine junction is intended to model a

hydroelectric turbine, in which hydraulic power is converted to
shaft power, which drives an electric generator
– Models only the reaction-type turbine, specifically, the Francis
turbine
– The turbine four quadrant data relies on having a complete set of
rated turbine performance parameters
– If no transient is defined for the turbine, it will continuously adjust
to changes in supply pressure and/or discharge pressure
I12 - 17
Partial Four Quadrant Pump Data
 Sometimes four quadrant pump data does not include data

points for all four quadrants
 This can cause convergence issues during the transient pump
calculations
 The AFT Impulse help file lists which of the four quadrant data
sets do not include full data sets
 If the four quadrant data you have selected is not a complete
data set, and it is causing calculation issues, try using the
next closest data set which contains full four quadrant data
I12 - 18
Specifying a Stagnant Region Pressure
 A stagnant region is where no pressure is defined

– Between two closed valves, for example
 During the steady state calculation this will result in a “No
Solution” shown in the output
 A transient run cannot be started with undefined pressures in
the model
 To avoid this, a stagnant region pressure can be entered in
one of the pipes
– This will define the pressure in the entire region
– Only one pressure can be entered for the region
I12 - 19
Specifying Stagnant Region Pressure (2)
 Consider the model where there are initially three closed

valves J7, J8 and J9
– There is no defined pressure between these valves and the
region is stagnant
 Valves J7 and J9 will be opened during the transient run
I12 - 20
 Running the model in steady state there is no solution for

pipes P5, P6 and P8
 But there is a pressure just upstream of valve J7 at the outlet

of pipe P4 that we can use
I12 - 21
 This pressure can be specified at the inlet of pipe P5

– That is just downstream of the valve
 This provides a pressure for the entire stagnant region
I12 - 22
 With this pressure matching upstream of the valve the

transient simulation can be run
 Valves J7 and J9 are instantly opened
– Pressures oscillate on either side of valve J7
I12 - 23
 A warning will be given if multiple pressures are defined
I12 - 24
Artificial Transients
 Before beginning a transient run, AFT Impulse performs a

pseudo time step for diagnostic purposes
 This calculation is performed with all of the transients ignored
 An artificial transient occurs when the steady-state results do
not match the results from this pseudo time step
 This situation can introduce false transients into the system
 If undetected, this artificial transient will interact with the
physical transient being modeled and corrupt the results
I12 - 25
Artificial Transients (2)
 Before AFT Impulse initiates the transient simulation it

performs a “zero time step computation” [pseudo time step]
and compares the predicted results with the results from the
steady-state solution
– If the difference exceeds the
artificial transient criteria specified
in the Transient Control window
(0.5% by default) a warning
message is displayed
– If the Stop Run if Artificial Transient
Detected option is selected the
simulation will be terminated
immediately
I12 - 26
Artificial Transients (3)
 Artificial transients can be detected manually

 Run the AFT Impulse model
– Do not use any of the transient input (e.g., by setting the
Transient Special Conditions to ‘Ignore Transient Data’)
• This will cause the system to remain at the original steady-state
condition
– Run the model for 10 to 100 time steps
 If no artificial transients exist, all pressure and flow solutions
will remain steady with time
 If pressures and/or flows change and transients are evident,
then an artificial transient exists
I12 - 27
Artificial Transients Example
 Open the “Artificial Transient.imp” example file to the “With

Artificial Transients” scenario, and run the simulation
 The run is terminated with Critical errors due to Artificial
Transients
– The results displayed in the output are not valid, because the
simulation did not run to a normal convergence
I12 - 28
Artificial Transients Example (2)
 Pipes P1, P2, and P11 show warnings that the initial pressure
conditions are not steady
– Checking Solution Control shows that the pressure and flow
tolerances have been set to 1.0E-4, rather than the AFT Impulse
default of 1.0E-5
– The user may have changed this to achieve a faster solution to
the steady-state
– Change the solution control values to the default AFT Impulse
values
I12 - 29
Artificial Transients Example (3)
 Pipe P103 shows a ‘zero flow’ warning

– This pipe ends at a dead end junction, which causes the zero
flow
– With zero flow, even extremely small tolerances in the solution
are relatively large
– In the Artificial Transient section in the Transient Control window,
set 'Ignore Flows Less Than' to 1 gal/min
– This will cause these minor variations around zero flow to be
ignored
 Rerun the model with the above changes
– The artificial transient issues have been resolved
I12 - 30
Vapor Cavitation - Beware!
 Vapor cavitation changes the wavespeed of the fluid

 For the DVCM and DGCM, AFT Impulse assumes:
– the wavespeed is constant
– all vaporization is localized at the computing sections
– the vapor bubbles do not move
 These assumptions make cavitation modeling less certain
than all liquid modeling
I12 - 31
Vapor Cavitation - Beware! (2)
 Be careful interpreting transient cavitation results!

 When using the “Discrete Vapor Cavity Model”:
– The first one or two pressure spikes after cavity collapse tend to
be the most reliable
– After the first one or two spikes, the non-physical pressure
spikes frequently begin
– Pay attention to the maximum vapor volumes in relation to the
computing volume
– Attention should be focused on the major trends of the pressure
results
• These tend to be where more pressure solutions are grouped
together, rather than isolated spikes
I12 - 32
Cavitation – DVCM and DGCM
 The existence of persistent cavitation in a simulation can

result in unrealistic results
– This is a result of limitations in the mathematical model used to
calculate transient cavitation in AFT Impulse
– Cavitation is actually a two-phase flow phenomenon, which can
be approximated by sophisticated mathematical models
• These models do not model true two-phase flow conditions
• In the presence of persistent cavitation, these mathematical models
can break down
I12 - 33
Cavitation – DVCM and DGCM (2)
 There are two fairly common situations where cavitation

issues arise in AFT Impulse models
– Cavitation exists in the initial model steady state
• Solution of the initial model steady state results in cavitation in the
system
– Cavitation exists in the final model steady state
• The final state of the system after the transient phenomenon have
subsided results in persistent cavitation in the system
I12 - 34
 Cavitation exists in the initial model steady state

– In this case, when AFT Impulse detects the presence of
cavitation in the initial steady state solution, an error message
will be generated, and the transient analysis is aborted
 Open the “Cavitation Issues Example.imp” example model file
to the “Cavitation at Initial Steady-State” scenario and run it
– The initial conditions in this simulation result in cavitation during
the initial steady state solution
– Note the resulting error message
I12 - 35
 Cavitation exists in the final model steady state

– In this circumstance, it is sometimes harder to determine the
validity of the cavitation results
– In some cases, the effect on the results is dramatic.
• Continuous creation and collapse of vapor cavities can result in
unrealistic pressure surges due to the creation of excessively large
vapor volumes
– In other cases, the results may be less obvious
I12 - 36
Cavitation – DVCM (5)
 Open the “Cavitation Issues Example.imp” example model file

to the “Cavitation at Final Steady-State” scenario and run it
– In this scenario, the fixed flow pump flow rate is reduced from
400 gal/min to 200 gal/min
– The final state of the system results in persistent cavitation at the
high point in the system
– Notice the “Critical Warning” message shown below:
I12 - 37
 The pressure plot shows that cavitation starts to occur at

about 1.5 seconds, which corresponds to the beginning of the
vapor volume increase
I12 - 38
 The vapor volume plot shows the growth of the vapor volume
– The total vapor cavitation volume preferably should not exceed
10% of the volume of the computing section (you will get a
warning message if this occurs)
– In this case, the pipe section volume is about 0.5 ft3 , while the
vapor volume is approaching 5 ft3
 Note that the vapor volume at the outlet to pipe P3 (at the
high point) continues to grow over time
 Vapor cavitation is assumed to occur at computing stations in
the pipes
I12 - 39
 Open the “LNG-2.imp” example file

 This model has persistent cavitation in pipes P71-P78
 Plot the static pressure and vapor volume vs. time for the
outlet of pipe P73
– P73 begins to cavitate after about 20 seconds, resulting in a
number of pressure spikes due to vapor cavity collapse
– Looking at the vapor volume of the cavities, however, reveals
very large volume in relation to the pipe section volume
• The volume of the sections in P73 is about 0.6 m3
• The volume of the vapor cavities is notably larger than 10% of the
section volume
– The pressure spikes in this system may not be realistic
I12 - 40
I12 - 41
DGCM User Input Parameters
 The gas cavity model requires only three special input

parameters: Ψ, α, and Reference Pressure. So what are
they?
 First lets deal with Ψ. In simple terms, Ψ is a weighting factor
between previous flow results (known) and current flow
calculations (not known)
 Now for the details: the growth or contraction of a cavity
volume depends on the flow deficit (outflow – inflow)
integrated over time. However, the method of characteristics
(MOC) is a finite difference method, so the evolution of flow
and cavity volume between t and ∆t is unknown and the
integral must be approximated using the trapezoid rule
I12 - 42
DGCM User Input Parameters: Ψ
 The trapezoid rule assumes that the change in flow with

respect to time is essentially linear. To the extent that linear
assumption is invalid, a weighting factor (i.e. Ψ) should be
used to bias the result toward t + ∆t (the new conditions).
Think of it as the trapezoid rule with a partially flat top. Using a
Ψ value of 0.95 implies that the change in volume, from a
previous flow deficit to a new flow deficit (or surplus) is
dominated by the new flow
 For the DGCM, a Ψ between 0.5 and 1.0 should be used
 Generally, DGCM results are more stable when Ψ is close to
1.0. The default is 0.95 and works well for the vast majority of
cases
I12 - 43
DGCM User Input Parameters: α and Ref
Pressure
 The DGCM treats distributed gas bubbles as one aggregate
volume. The mass of air within the volume is assumed
constant, as well as the temperature within the volume
 Assuming constant mass and temperature, the ideal gas law
can be used to relate volume to pressure alone
 α is the void fraction, or the potion of the overall section
volume that is vacant
 With the section volume known ahead of time, a reference
condition can be specified using a reference α and reference
pressure
I12 - 44
DGCM User Input Parameters: α and Ref
Pressure (2)
 With known reference parameters, the right hand side of the
ideal gas law can be calculated and will remain constant for
the entire transient simulation
 The default values of α and reference pressure are 10-7 and 1
atm, respectively
 The default values work well for the majority of cases and
should not be modified without careful analysis
 A large initial void fraction would imply slug flow and should
be avoided
 A small initial void fraction would tend toward DVCM behavior
because the initial volume would be vanishingly small and
ineffective
I12 - 45
Use Relaxation
 Relaxation acts as a damping factor on the flow rate and

pressure updates and is helpful for highly non-linear systems
such as pumped systems
Selecting Fixed relaxation

allows flow rate and
pressure relaxation to be
individually specified
I12 - 46
Use Relaxation (2)
 Multiple parallel pumps frequently require a flow rate

relaxation of 0.1 or lower
– We have used 0.001 is some cases
 For multiple parallel pump systems, using flow rate relaxation
in addition to a broad brush initial flow rate guess can be
helpful
 Keep the flow rate relaxation at least one or two orders of
magnitude larger than the relative tolerances
I12 - 47
Use Relaxation (3)
 Pressure relaxation adjustment would be used if the pressure

tolerance is oscillating around a value
– Only pressure relaxation values of 0.5 or 1.0 should be specified
 Automatic relaxation is 0.5 for pressure, begins at 1.0 for flow
and decreases based on the solution progress, is sufficient for
most models
I12 - 48
Try Absolute Tolerance
 Some models converge much better with absolute tolerance

than relative tolerance
 What this usually means is that there is at least one element
of the model that is close to the real answer in absolute terms
but does not lock in on a relative (percentage) basis
 Absolute Tolerance is the only change required for some
problem models
I12 - 49
More on Absolute Tolerance
 Some models converge very slowly

– You get the feedback on the Solver activity in the Solution
Progress Window
 It is often easier to see where the Solver is going in absolute
terms than it is in relative terms
 Some models that are converging slowly do not appear to be
converging on a relative tolerance basis even though they are
I12 - 50
More on Absolute Tolerance (2)
 If you choose only relative tolerance or only absolute, the

other out of tolerance value is shown on the Solution Progress
window in a gray box
– You can monitor the progress of unused tolerance values even
though the Solver does not use them for convergence (unless
you change the setting in Solution Control)
I12 - 51
Turn Off Parts of the Model
 AFT Impulse's Workspace makes it easy to turn off large

sections of a model so a portion of it can be run
 Turning off parts of the model can help isolate the problem
portion of the model
I12 - 52
Turn Off Transient Data
 Using Transient Special Conditions, different transient inputs

can be disabled
 Turning on one transient at a time and running the model may
help you understand what is happening
I12 - 53
Check Event Messages
 The Output window has two reports on "event messages”

– Events sorted by junction
– Events sorted by time
 Check this to see if system behavior is in accordance with
expectations
– Check valve closures
– Relief valve opening
– Event based transient initiation
I12 - 54
Use Animation
 If the model runs but gives strange results, use animation

features to understand what is happening
I12 - 55
Send the Model to AFT
 If all else fails you can send the model to AFT

 We need to run some models in source code to help identify
the problem
 We place a great deal of emphasis on quality technical
support
 Refer to the AFT User Forum, http://forum.aft.com/
 Send model to AFT by INTERNET at support@aft.com
I12 - 56
P
A
Orifice
Connector
Q Pipe
A
L
C
Overview of Seminar
INTRODUCTION
Surge Suppression
 Surge pressures that are excessively high can burst pipes

and damage equipment
 Surge pressures that reduce the local pressure can result in:
– Crushed pipes
• Due to atmospheric pressure exceeding the internal liquid pressure
– Cavitation and liquid column separation
• Can then cause large pressure spikes when the cavity collapses
– Sub-atmospheric pressures that are unacceptable for drinking
water pipelines
 Surge pressures can create imbalanced forces that move
pipes, dislodge insulation and ultimately break supports
I13 - 1
Surge Suppression (2)
 Options for surge reduction include:

– Reducing steady-state operating velocity (since surge is directly
related to velocity changes)
– Slowing the change in system operating conditions
– Changing the profile of how equipment operates in transient
mode
– Installing surge suppression equipment
I13 - 2
Velocity Reduction
 Surge pressures can be reduced by using larger diameter

pipes which thereby reduces velocity
I13 - 3
Slowing System Operation Changes
 Slowing system operation changes usually means slowing the

rate of valve position changes
 One way to do this is to use parallel valves with staggered
closure times
 Swaffield* discusses how closing the valve more quickly at
first and then more slowly at the end can result in significant
surge reduction
– Examples are given where the first 80% of closure occurs over
the first 20% of closure time, with the final 20% closure occurring
over the remaining 80% of the time
 It should be noted that obtaining accurate transient valve data
can be difficult
* Swaffield and Boldy, "Pressure Surges in Pipe and Duct Systems"

I13 - 4
Surge Suppression Equipment
 Options for surge suppression equipment depend on steady-

state pressure levels and whether one is protecting for high
pressure or low
 Common choices include:
– Gas Accumulators
– Surge Tanks
– Relief Valves
– Vacuum Breaker Valves
I13 - 5
Waterhammer Videos - Accumulators

– How a Bladder Surge Tank can alleviate column
separation2.wmv (1:16)
• Gas bladder accumulator upstream of valve which closes
• online version - http://www.youtube.com/watch?v=E6NIA4LxPPw
– accumulator after pump trip.wmv (1:23)
• Gas bladder accumulator close to pump - after pump trip
• online version - http://www.youtube.com/watch?v=kiTzez0x6aQ
I13 - 6
Gas Accumulators
PA
 Gas accumulators are enclosed vessels

pre-charged with a common gas such as Orifice
air or nitrogen Connector
QA
 When a surge pressure reaches the LC
Pipe
accumulator, the gas in the vessel

compresses or expands to allow liquid in
or out of the vessel
– This has the tendency to reduce the surge pressure
– It also changes the frequency response of the pipe system
I13 - 7
Gas Accumulators (2)
PA
 Accumulators are not a cure-all for surge

problems in a pipe system Orifice
– They can help if engineered properly Connector
Q Pipe
– Most accumulators in industrial installations L
A
C
are not specified properly, and do not offer

the protection the designer intended
– If you have to specify one, the best approach is to use
waterhammer software and try different options to determine the
best size and location
I13 - 8
Modeling Gas Accumulators
 Important parameters are the polytropic constant and initial

gas volume
 You also can model a flow
restrictor (such as an orifice)
or a short connecting pipe
– These change how the
accumulator interacts with
the system
 Max and Min Volumes
prevent volumes from
exceeding certain sizes
I13 - 9
Modeling Gas Accumulators (2)
 During steady-state conditions accumulators behave like

branch junctions
 Accumulators can also act as
a reference pressure
– This is entered in the "Initial
Pressure for Zero…" field on
the Optional tab
 This allows the accumulator
to act as an Assigned Pressure
junction during the steady-
state only when the net flow
is zero
I13 - 10
Surge Tanks
 Surge tanks are similar to gas accumulators,

but are open to the atmosphere
 When a surge pressure reaches the surge
tank, the liquid in the tank rises or falls
– This has the tendency to reduce the surge
pressure
– It also changes the frequency response of
the pipe system
I13 - 11
Surge Tanks (2)
 Surge tanks are more commonly used for

low pressure water systems
– If the steady-state pressure is high, the
surge tank will need to be very tall
– Surge Tanks are more commonly in civil
engineering systems such as hydroelectric
facilities
 Surge Tanks are similar to Reservoirs in
Impulse
– Reservoirs do not change liquid level
automatically, and are thus more massive
liquid sources
I13 - 12
Modeling Surge Tanks
 Surge Tanks can drain in

reality, but Impulse does not
model this
– This allows air into adjacent
pipes
 Surge Tanks can spill over the
top when the Tank Height is
entered
 Geometries other than pure cylinders can be modeled by
entering the tank accumulated volume as a function of height
 Short connector pipes and orifices can be modeled
 Surface pressure transients can be modeled to represent
known pressures in an enclosed tank
– These can be time or event based I13 - 13
Modeling Surge Tanks (2)
 A hybrid
accumulator/surge tank
can be modeled called
a "Dipping
Tube Vessel" using
features on
Tank Model tab
 During steady-state
conditions
accumulators behave
like branch junctions
– Impulse solves for the
liquid height
I13 - 14
Modeling Surge Tanks (3)
 The surge tank can be used as a reference pressure during

steady-state
– It then acts just like a reservoir junction
– The liquid height is input by the user (in the Initial Liquid Height
field)
– The surge tank no longer represents a point
of mass balance
I13 - 15
Relief Valve Modeling
 The relief valve model

used in Impulse has
numerous changes from
previous versions
– Layout of information
– Opening/closing
profiles
– Constant backpressure
– Set pressures
– Opening/closing rate
limits
I13 - 16
Relief Valve Modeling (2)
 Impulse offers three relief valve configurations
– Internal
• Relieves into another pipe which was previously isolated.
• Two connecting pipes
• Acts like a regular valve which is closed until set pressure is
reached
– Exit
• Relieves to an external ambient pressure such as the atmosphere
• One connecting pipe
• Associated back pressure must be specified
I13 - 17
Relief Valve Modeling (3)
 Impulse offers three relief valve configurations
– Inline Exit
• Relieves to an external ambient pressure, but is located between
two pipes
• Valve is a lossless connection when closed, and allows flow
between the pipes
I13 - 18
Opening/Closing Profiles
 The relief valve offers four types of profiles that define the
behavior of the valve as it is opening and closing
– Instant
• The Instant profile causes the relief valve to open or close fully
instantaneously as the opening or closing conditions are met
I13 - 19
Opening/Closing Profiles (2)
– Time
• The Time profile causes the valve to open or close according to a
Cv vs. Time profile, as the opening or closing conditions are met
• The opening and closing profiles are entered on the Transient tab,
as needed.
I13 - 20
– Pressure
• Causes the valve to passively open or close due to pressure/dP
variations according to a Cv vs. Pressure/dP profile
• The profile is entered on the Loss Model tab
• Maximum rate limits can be applied to Pressure profiles.
I13 - 21
– Never
• Only available as a closing profile
• Easy modeling of a relief valve that opens, but never closes again
(e.g., rupture disk)
I13 - 22
Remote Sensing
 The set pressure is evaluated at a specified remote location

(other than the upstream side of the relief valve)
 This allows the relief valves to be more responsive to surge
events by opening the valve before a pressure surge arrives
directly at the valve
I13 - 23
Opening/Closing Profile Combinations
 Opening and closing profiles can be used in different

combinations
– A relief valve could open instantly, and then close according to a
passive pressure profile
 There are a few special considerations regarding some
combinations of profiles
I13 - 24
(2)
 Pressure as the opening profile
– If Pressure is selected as the opening profile, then Pressure is
the only option for a closing profile
– Logically, a valve that opens passively according to pressure
must also close passively according to pressure.
 Never
– This profile can only be applied as a closing profile, allowing the
valve to open, but never close
– This is useful for modeling rupture disks
I13 - 25
(3)
 Time – Pressure
– Valve will begin to open according to the opening time transient
profile, as defined on the Transient tab
– If the valve Cv from the time profile exceeds the valve Cv based
on the passive pressure profile for the given pressure conditions,
the valve will switch to passive pressure control, and proceed to
open or close based on the pressure profile for the duration of
the transient
– This is to prevent the time profile from causing the valve Cv to
“overshoot” the valve Cv predicted by the pressure profile for the
pressure conditions during the transient and introducing
unrealistic transients
I13 - 26
Valve Setpoints
 After the opening and closing profile types have been defined,
the valve setpoints must be entered.
 The valve setpoints determine when the valve opens and
when it closes.
 The setpoints can be entered in terms of either head, or
pressure/dP.
I13 - 27
Valve Setpoints – Exit Pressure
 If the relief valve is an Exit valve, or an Inline Exit valve, the

external exit pressure must be entered.
I13 - 28
Valve Setpoints – Constant Backpressure
 Relief valves can be hydraulically balanced (constant

backpressure), or non-hydraulically balanced
 Hydraulically balanced
– Downstream valve mechanism is isolated from the downstream
pipe pressure
– Often by attaching a bellows device to the downstream side of
the valve stem
– Backpressure is constant, the valve setpoints can be specified in
terms of the actual upstream pressure
I13 - 29
Valve Setpoints – Constant Backpressure
(2)
 Non-hydraulically balanced
– Downstream mechanism is not isolated from the downstream
pressure
– Pressure acting on downstream valve stem is not necessarily
constant as transients are acting on the system
– The pressure required for the valve to open or close will be
relative to the downstream pressure
– Valve setpoints must be specified as differential pressures
I13 - 30
Valve Setpoints – Pressure Setpoints
 Set Pressure
– The pressure at which the valve will open
– All valve profile types must specify the Set Pressure
 Blowdown Pressure
– Some relief valves close at a pressure that is lower than the set
pressure
– Blowdown Pressure must be specified for all opening and
closing profiles, unless the Never closing profile is selected
– If the relief valve closes at the same pressure where it opens,
then the Set Pressure and the Blowdown Pressure should be set
to the same value.
I13 - 31
Valve Setpoints – Pressure Setpoints (2)
 Overpressure
– When Pressure type profiles (passive relief) are used, then the
Overpressure must be specified.
– Overpressure is the pressure required to cause the relief valve to
open to its fully open position.
I13 - 32
Loss Model
 The Relief valve loss model information is entered on the

Loss Model tab.
 If Instant, or Time profiles are selected, then the user has the
option of entering a constant Cv, Kv, or K factor for the valve.
– This value represents the fully open loss value for the valve.
 If a Pressure profile is selected, then the variable Cv or Kv
profile must be entered.
I13 - 33
Loss Model – Variable Data
 Linear based on setpoints

– Variable loss profile will be based on the valve pressure
setpoints entered on the Valve Model tab
– The valve loss will vary linearly between the Blowdown pressure
and the Overpressure
– Requires that the Full Open loss value be entered.
I13 - 34
Loss Model – Variable Data (2)
 Non-linear
– User can enter a more detailed loss profile
– First data point must have a Cv (or Kv) of 0, at a pressure equal
to the Blowdown pressure entered on the Valve Model tab
• This represents the valve in its closed state.
– Final data point must have a loss value equal to the valve’s fully
open loss value, at a pressure equal to the Overpressure
entered on the Valve Model tab
• This represents the valve in its fully open position.
– The data points between the Blowdown pressure and the
Overpressure will vary based on the values entered by the user.
I13 - 35
Opening/Closing Rate Limits
 If a Pressure-Pressure type profile is specified, the maximum

valve opening and closing rates can be specified
– This allows modeling limiting devices which prevent the valve
from opening or closing too quickly
– If, during the transient, the pressures would cause the valve to
open or close faster than the rates specified in the rate limits
table, the valve loss value will be set to the value dictated by the
rate limits.
– Defined on the Loss Model tab
I13 - 36
Opening/Closing Rate Limits (2)
I13 - 37
Vacuum Breaker Valves
 Vacuum breaker valves (also known as air valves) are

primarily used to protect against low pressure conditions
 When the liquid pressure drops below some set pressure
(usually atmospheric), air (or any gas you choose) is allowed
in to maintain near atmospheric pressure at that location
 Vacuum breakers are typically located at high points in the
system which are most vulnerable to low pressure transients
I13 - 38
Vacuum Breaker Valves (2)
 Typical designs allow air to flow in and out

 Five conditions can exist:
– Valve closed
• Behaves like a branch junction
– Valve open with subsonic flow in
– Valve open with sonic flow in
– Valve open with subsonic flow out
– Valve open with sonic flow out
I13 - 39
Modeling Vacuum Breaker Valves
 Important input are the inflow and

outflow CdA values
– These are the air flow areas
 The inflow CdA is usually as large
as possible to allow rapid air flow in
 The outflow CdA is usually much
smaller than the inflow to let the air
out slowly and prevent rapid collapse of the liquid columns
which can cause pressure surge
 Transients can start with air already in the system
– These valves can be modeled as "air eliminators"
 Air flow is assumed to be adiabatic
I13 - 40
Three Stage Air Release Valves
 AFT Impulse allows users to model three stage “Anti-shock”

air release and vacuum break valves
– This is a particular type of air release valve sold by Vent-O-Mat
valves
 Three stage valves work the same as regular valves when
gas is flowing in, but they have two orifice sizes when gas
flows out
 Pressure difference or volumetric flowrate criteria specify
which orifice to use
– When the actual value is less than the specified value, the
normal Outflow CdA is used
– When the actual value is greater than the specified value, the
intermediate orifice CdA is used
I13 - 41
Three Stage Air Release Valves (2)
 Intermediate orifice CdA values are typically much smaller

than the normal outflow CdA
– This slows the rate of release of the last amounts of gas in the
system, and results in slower fluid velocities as the valve closes
– This translates into greatly reduced surge pressures than those
caused by rapidly closing air release valves
I13 - 42
Three Stage Air Release Valves (3)
 Important input are the

inflow, outflow, and
intermediate CdA values
– These are the air flow areas
 The intermediate outflow
CdA is typically very small in
relation to the outflow CdA
– This slows the rate of
discharge as the last of the
air is expelled
I13 - 43
Liquid Accumulators
 A liquid accumulator is a liquid volume inside a pipe system

 Some examples are:
– Expansion joints in a pipe which have higher flexibility than the
pipe
– An enclosed, liquid-solid vessel in the pipe system
 Liquid accumulators are not, by design, used as surge
suppression devices
– They are just parts of a pipe system that can interact in a certain
way
I13 - 44
Modeling Liquid Accumulators in Impulse
 Important parameters are the

elasticity and initial volume
 The elasticity is how much the
vessel changes volume for a
given change in pressure
– This is frequently difficult to
determine and can only be
approximated
 The initial volume is the volume of the vessel at the start of
the simulation
– As the pressure changes, Impulse tracks the volume changes
over time
I13 - 45
Modeling Partially Full Pipes
 Pipes which connect to vacuum breaker valves, exit valves,

spray discharge or assigned pressure junctions can drain
during the transient, or start the transient partially full
 Pipe drainage is limited to the specified pipe - pipes will not
drain from pipe to pipe
The partially full
pipe option is set
on the Pipe
Properties window
Options tab
I13 - 46
Modeling Partially Full Pipes (2)
When the partially
full pipe option has
been selected, data
related to the
partially full pipes
becomes available
when selecting graph
data
I13 - 47
Overview of Seminar
INTRODUCTION
Calculating Unbalanced Forces
 AFT Impulse can calculate unbalanced forces in a system

based on the transient system pressures
– In AFT Impulse “forces” is typically used to imply “unbalanced
hydraulic forces”
 Both differential and point forces can be calculated
 Calculated forces can be plotted in the Graph Results window
 Calculated forces can also be exported to Force vs. Time data
files which can then be imported into stress analysis
applications such as CAESAR II and TRIFLEX
NOTE: Pipe force calculation is not reliable when

cavitation occurs.
I14 - 1
Defining Force Sets
 In order to process force data, the locations where the forces

are to be calculated must be defined
– The forces may be defined as differential or point forces
 Forces are defined on the Force Output tab in the Transient
Control window
– A force report showing the input details is shown in the Model
Data (ensure that Show Force Sets is enabled in the Other tab of
the Model Data Control window)
– Design Alerts can be applied to Force Sets and shown in the
Force Summary
I14 - 2
Defining Force Sets (2)
 Force Sets are defined by a Start Node and an End Node

 For each node the user must specify:
– The pipe where the force is to be calculated
– The distance from the start of the pipe (Station 0) to the location
of the node
– The closest pipe station to the specified distance along the pipe
will be selected automatically
I14 - 3
Defining Force Sets (3)
 Once a force set is defined, AFT Impulse can calculate the

differential force between the two defined points
– The resulting forces are displayed in the Output and can be
displayed in the Graph Results window or exported to a
Force/Time file
– A typical application of force set data would be to select the pipe
endpoints between a pair of pipe elbows
 A Point force can be defined by selecting the Point option in
the Force Set column
– This disables the cells for the end node, as they have no
meaning in this application
– For point forces, the ambient pressure that these forces will act
against can be specified
I14 - 4
Unbalanced Forces Across 3+ Pipes
 Unbalanced forces may be

calculated across three or
more pipes
– Pipes must be connected
sequentially end to end
I14 - 5
Unbalanced Forces Across 3+ Pipes (2)
 All of the pipes in the path must

be added to a named group
I14 - 6
Unbalanced Forces Across 3+ Pipes (3)
 Define the force set in the Transient Control window

– Select “Difference (3+ Pipes)” as the Force Type
– Select the appropriate Group Name under Pipe Group
I14 - 7
Calculating Unbalanced Forces
 There are multiple forces that come into play when calculating
unbalanced forces
– Pressure difference between points
– Friction forces
– Fluid momentum changes (including area changes)
 Traditionally the force imbalance is based solely on the
difference in pressure between points, with adjustments to
account for the hydrostatic pressure due to elevation
differences
I14 - 8
Calculating Unbalanced Forces (2)
 Friction and momentum are not typically taken into account,

though each will bias the pressure difference and hence the
force calculation
 In cases where force sets span across two pipes and the
pipes have a different diameter, the area change itself cannot
create an unbalanced force even though the static pressure
changes
– AFT Impulse automatically accounts for this in the force balance
calculation
I14 - 9
Unbalanced Forces Traditional Method
 Consider forces acting in the system shown

– There is a pressure difference between nodes 1 and 2 that is
due solely to frictional pressure drop in the pipe and across the
valve – this is not included in the force balance below
– All forces at Node 1 are balanced by forces at Node 2 in steady-
state
– If frictional force is not included in the force balance calculation,
then the steady-state situation will show a force unbalance which
is not correct
I14 - 10
Unbalanced Forces Traditional Method
(2)
I14 - 11
Friction and Unbalanced Forces
 The “Include Friction” option on the Force Data tab in the

Select Graph Data window and in the CAESAR II Export or
TRIFLEX Export window will account for this
– To include friction effects, the option to save the required data
must be selected in the Transient Control window
– Note that selecting this option can have a significant effect on the
output file size
• To include friction effects, AFT Impulse must save data for every
pipe station in pipes that are included in any force sets
I14 - 12
Friction, Momentum and Unbalanced
Forces
 Now consider unbalanced friction forces and momentum
changes acting in the same system
– In steady-state these will be in balance - during a transient they
usually will not
– Typically (but not always) the forces caused by momentum
changes are relatively small compared to the other transient
forces in a system
– The “Include Friction” and “Include Momentum” options in the
Force Data graphing and the CAESAR II/TRIFLEX Export
window will account for this
I14 - 13
Friction, Momentum and Unbalanced
Forces (2)
Friction & pressure loss forces Other forces + PAA + PBA = 0
Force = PA x A Force = PB x A
Momentum = mAΔVX,A Momentum = mBΔVX,B
A = πD2/4
ΣFfriction + PAA + PBA =

mAΔVX,A - mBΔVX,B
I14 - 14
Viewing Force Data
 Force data can be graphed from the Transient Force Data tab
in the Select Graph Data window
– Select the force pair data to be graphed
– Select whether to include the momentum and friction effects
I14 - 15
Viewing Force Data (2)
 The forces calculated by Impulse represent the force of the

fluid acting on the pipe and supports.
I14 - 16
Exporting Force Data
 Transient force/time data can be exported by copying the data

to the clipboard from the Graph Results window
 Alternatively, the “Save Graph Data As…” option on the File
menu in the Graph Results window allows the plotted
force/time data to be saved as a text file
 After the force/time data has been exported, the user has the
option of bringing the data into other applications
– Spreadsheets
– Stress analysis
I14 - 17
Exporting Force Data - CAESAR
II/TRIFLEX
 Transient force/time data can be exported in a format that can
be loaded directly into the CAESAR II or TRIFLEX pipe stress
application
– Select “Export CAESAR II Force File” (or TRIFLEX) from the File
menu
– This will open the Export CAESAR II Force File window
– This window allows the user to define the content of a force/time
data file and export it
I14 - 18
US
Transient Force Hands-on Example
 A gravity drain system transfers water from one tank to

another through a valve in the connecting pipeline. It is
necessary to determine the transient forces that occur when
the valve is closed in order to analyze pipe stresses and loads
on the pipe supports. Compare results both with and without
friction effects (Model “Valve closure with pipe forces.imp”)
I14 - 19
US
Transient Force Hands-on Example (2)
 Open the “Valve closure with pipe forces.imp” model file

– The valve in the system is set to close 90% of the way over 2
seconds
 Define two force sets:
– Between 480 ft and 520 ft
(Vertical leg)
– Between 520 ft and 1040 ft
(across the valve)
 Note: The lengths indicated
in the figure are measured
from the J1 reservoir
I14 - 20
US
Answers to Hands-on Example
 The force sets should be defined as shown

– Note that the “Length to End Node” for the force set across the
valve is defined as 40 feet and not 1040 feet
• This is because the End Node is actually located in Pipe 5 and the
length to the nodes is defined from Station 0 of the pipe on which
the node is located
– Also note that the option to save the required data to include
friction effects has been checked
I14 - 21
US
Answers to Hands-on Example (2)
 The unbalanced force data for the 480 ft to 520 ft force set
without friction and momentum effects included should appear
as shown
– Note that at time 0, which represents the initial, steady state
results, the force imbalance is approximately -2700 lbf. This
imbalance is due to the presence of friction.
I14 - 22
US
 The unbalanced force data for the 480 ft to 520 ft force set
with friction and momentum effects included should appear as
shown
I14 - 23
US
 The unbalanced force set for the 520 ft to 1040 ft section

without friction or momentum included should appear as
shown
– There is a steady-state force imbalance of more than -1300 lbf
where in reality there would be no force imbalance.
I14 - 24
US
 The largest magnitude force imbalance is larger than in the

case of the ‘480 ft to 520 ft’ force
– This is due to the greater frictional pressure loss from the longer
run of piping and the valve between the two locations in this
force set
I14 - 25
US
 The unbalanced force set for the 520 ft to 1040 ft section with
friction included should appear as shown
– With friction included there is no force imbalance under steady
state conditions
– The force imbalance is about 30% less than that calculated
without friction, and in the opposite direction
I14 - 26
SI
Transient Force Hands-on Example
 A gravity drain system transfers water from one tank to

another through a valve in the connecting pipeline. It is
necessary to determine the transient forces that occur when
the valve is closed in order to analyze pipe stresses and loads
on the pipe supports. Compare results both with and without
friction effects (Model “Valve closure with pipe forces
(SI).imp”)
I14 - 27
SI
Transient Force Hands-on Example (2)
 Open the “Valve closure with pipe forces (SI).imp” model file
– The valve in the system is set to close 90% of the way over 2
seconds
 Define two force sets:
– Between 150 m and 160 m
(Vertical leg)
– Between 160 m and 300 m
(across the valve)
 Note: The lengths indicated
in the figure are measured
from the J1 reservoir
I14 - 28
SI
Answers to Hands-on Example
 The force sets should be defined as shown

– Note that the “Length to End Node” for the force set across the
valve is defined as 15 m and not 315 m
• This is because the End Node is actually located in Pipe 5 and the
length to the nodes is defined from Station 0 of the pipe on which
the node is located
– Also note that the option to save the required data to include
friction effects has been checked
I14 - 29
SI
 The unbalanced force data for the 150 m to 160 m force set
without friction effects included should appear as shown
– Note that at time 0, which represents the initial, steady state
results, the force imbalance is about -12 kN. This imbalance is
due to the presence of friction.
I14 - 30
SI
 The unbalanced force data for the 150 m to 160 m force set
with friction effects included should appear as shown
I14 - 31
SI
 The unbalanced force set for the 160 m to 300 m section

without friction or momentum included should appear as
shown
– There is a steady-state force imbalance of about -5000 N where
in reality there would be no force imbalance.
I14 - 32
SI
 The largest magnitude force imbalance is larger than in the

case of the ‘150 m to 160 m’ force
– This is due to the greater frictional pressure loss from the longer
run of piping and the valve between the two locations in this
force set
I14 - 33
SI
 The unbalanced force set for the 160 m to 300 m section with
friction included should appear as shown
– With friction included there is no force imbalance under steady
state conditions
– The force imbalance is about 35% less than that calculated
without friction and in the opposite direction
I14 - 34
I15. Using Scenario Manager
Overview of Seminar
INTRODUCTION
Scenario Manager
 The Scenario Manager allows you to keep variants of a model

all with the same model
– When changes are made to the base model, they are
automatically passed downward
– Changes at lower levels do not pass upwards
I15 - 1
Scenario Manager (2)
Current
Workspace
scenario
Scenario tree Notes can be

added for
each Scenario
Rename, delete,
clone, promote
& save
scenarios by
clicking here
Create a new
scenario by
clicking here
I15 - 2
Quick Access Panel
 The Quick Access Panel provides convenient utilization of all

Scenario Manager features.
I15 - 3
Types Of Changes
 The types of changes that can be made are very broad

– Junctions can be turned on and off to evaluate different
operating conditions
– Pipe and junction data can be varied to parametrically evaluate
competing designs
– You can build an existing system as your base model then add to
the system to evaluate expansion possibilities on the existing
system
– You can easily evaluate different working fluids by setting them
up as different children scenarios
– You can compare a newly-built clean system to one that has
been in service for a period of time with worn/corroded pipes,
etc.
I15 - 4
Data Propagation
 Changes to ancestors propagate to all descendants if the

descendant data has not been modified
 Changes to descendants never propagates to ancestors
I15 - 5
Data Propagation (2)
 For many users, it is easiest Diameter Length

to grasp Scenario Manager Base 3 25
when it is explained how the Child #1 __ __
coding logic is actually Gr. Child #1 __ __
implemented
 Blank fields for children,
grandchildren, etc., mean to
look to the parent for the
data
– If the parent is blank, then
look to the grandparent
 The Base Scenario never
has blank fields
I15 - 6
 Here Child #1 does not Diameter Length

have a blank field, so it’s Base 3 25
Diameter would be 2, not 3 Child #1 2 __
 Gr. Child #1 would have a Gr. Child #1 __ __
Diameter of 2
– Both still have Lengths of
25
I15 - 7
 Changing the Base Diameter Length

Scenario Diameter from 3 to Base 3 25
6 would not impact Child #1 Child #1 2 __
or any descendants in that Gr. Child #1 __ __
line
 Changing the Length from
Diameter Length
25 to 40 would also change
Base 6 40
the length in Child #1, Gr.
Child #1 2 __
Child #1, and any __ __
Gr. Child #1
descendants of Gr. Child #1
I15 - 8
Data Propagation (5) Diameter Length
Base 3 25
Child #1 2 __
 Even if the Gr. Child #1 has Gr. Child #1 3 __
the same Diameter as the

Base, it is not linked to the Diameter Length
Base because it and its Base 3 25
parent are not blank Child #1 3 __
Gr. Child #1 3 __
– Any change to the Base
Diameter would not affect
any descendent because
Child #1 is not blank Diameter Length
 If the Diameter in Child #1 Base 3 25
is changed to be the same Child #1 __ __
as the Base, it will be Gr. Child #1 __ __
“blanked out” the next time
the scenario is loaded
– So will Gr. Child #1 I15 - 9
 Here, Child #1 would have the Diameter Length

following: Base 3 25
– Diameter = 2 Child #1 2 __
Gr. Child #1 __ 15
– Length = 25
– Changes to Base Diameter
will not affect Diameter
– Changes to Base Length will
affect Length
I15 - 10
 Here, Gr. Child #1 would have Diameter Length

the following: Base 3 25
– Diameter = 2 Child #1 2 __
Gr. Child #1 __ 15
– Length = 15
– Changes to Base Diameter
will not affect Diameter
– Changes to Child #1 Diameter
will affect Diameter
– Changes to Base Length or
Child #1 Length will not affect
Length
I15 - 11
Ancestral Data
 Ancestral source of data can be viewed for all pipes and

junctions in Model Data
Scenario data can be

colored for easier viewing
Scenario names shown at left
Parameters which change

are highlighted
I15 - 12
Links to Parent
 A link to a parent may be re-established by returning the

attribute to the same value as that of its parent
– This can be done manually be entering the value or selecting
Same As Parent from within a pipe or junction properties
window, Solution Control or System Properties.
 Links are identified by comparing attribute values on a pipe or
junction number by number basis.
– This means that renumbering a scenario will break all links with
its parent (since numbers must unique)
I15 - 13
Links to Parent (2)
 You can make a pipe have the same attribute as its parent by
choosing Same As Pipe: “Parent Pipe Data”
– Junctions function similarly
I15 - 14
Using Batch Runs
 AFT Impulse has the ability to run multiple simulations

sequentially in batch mode
– The Batch Run window is opened by selecting ‘Start Batch
Run…’ from the File menu
 There are two types of batch runs
– Multiple scenarios
– Multiple model files
 Generate a list of
simulations to run in
batch mode by Once the list of scenarios or
model files to run is created,
selecting either the clicking ‘Start Run’ will load
and run the selections in
‘Add Scenarios’ or order
‘Add Model Files’ button
I15 - 15
Notes
I15 - 16
US
Hands-on Example: LN2 Transfer
 Open the “LN2 Transfer.imp” example file

 A liquid nitrogen transfer system is used to fill two storage
tanks. Only one tank is filled at a time. Model the change-over
from filling Tank A to filling Tank B. Assume the tank levels
stay constant during the time frame of the simulation. The
valves cycle open and closed simultaneously. Determine the
maximum system stagnation pressures for three change-over
scenarios:
– Instantaneous valve cycle
– 5 second valve cycle
I15 - 17
US
Hands-on Example: LN2 Transfer (2)
 Use 2 sections in the controlling pipe

 Use the Discrete Vapor Cavity Model for Transient Cavitation
 Run the model for 20 seconds
 Does cavitation occur for any of the cases?
 It is decided that the piping in the system should be 8 inches
instead of 6 inches. What are the new maximum system
pressures?
I15 - 18
US
LN2 Transfer Results
 6-inch pipe:
Case Max Stagnation Pressure (psig)
Instantaneous valve cycle 175.6 (cavitation)
5 second valve cycle 109.0 (cavitation)
10 second valve cycle 105.3 (no cavitation)
 8-inch pipe:
Case Max Stagnation Pressure (psig)
Instantaneous valve cycle 145.6 (cavitation)
I15 - 19
Notes
I15 - 20
SI
Hands-on Example: LN2 Transfer
 Open the LN2 Transfer (SI).imp example file

 A liquid nitrogen transfer system is used to fill two storage
tanks. Only one tank is filled at a time. Model the change-over
from filling Tank A to filling Tank B. Assume the tank levels
stay constant during the time frame of the simulation. The
valves cycle open and closed simultaneously. Determine the
maximum system stagnation pressures for three change-over
scenarios:
– Instantaneous valve cycle
I15 - 21
SI
Hands-on Example: LN2 Transfer (2)
 Use 2 sections in the controlling pipe

 Use the Discrete Vapor Cavity Model for Transient Cavitation
 Run the model for 20 seconds
 Does cavitation occur for any of the cases?
 It is decided that the piping in the system should be 8 inches
instead of 6 inches. What are the new maximum system
pressures?
I15 - 22
SI
LN2 Transfer Results
 6-inch pipe:
Case Max Stagnation Pressure (kPa(g))
Instantaneous valve cycle 1233 (cavitation)
 8-inch pipe:
Case Max Stagnation Pressure (kPa(g))
Instantaneous valve cycle 1017 (cavitation)
I15 - 23
I16. Customizing Impulse & Using
Databases
Overview of Seminar
INTRODUCTION
Customizing Impulse
 Your User Options can be customized

 The Toolbox tools can be customized
 These preferences can be saved to a file
– The file can be shared between colleagues
– The file can be incorporated in to a network database
– You can develop files for different clients, reports and project
phases
I16 - 1
User Options – Unit Preferences
 Here you specify the units you generally work with to be the
defaults
Select the unit

system
You can set a

preferred unit
by selecting
one of the units
and clicking
here
I16 - 2
User Options – Pipes
 The pipe thickness on the Workspace can be set

 The pipe thickness on the Workspace can be varied
according to the pipe diameter
 The line style used when a pipe is closed can be chosen
I16 - 3
User Options – Pipes (2)
This path is closed
The pipes thickness will vary

from 2 to 4 pixels according to This is a 10 in. pipe This is an 8 in. pipe
diameter
I16 - 4
User Options – Junctions
 Junction Icon Styles, Sizes, and Label Locations can be

specified
I16 - 5
User Options – Labels
 Auto Increment Labels, Display Defaults, Allow Label

Movements
I16 - 6
User Options – Colors and Fonts
 The font and colors used by AFT Impulse can be changed
Click on
the
category
Then
choose a
color
Click here to
pick a font
I16 - 7
User Options – Other
 Specify the Workspace Symbols and Background Picture

Scaling
I16 - 8
User Options – Pipes and Junctions
 Here you specify the parameters you generally work with to

be the defaults
 You can still use the other parameters as needed
You can
specify a
default
pipe
material & You can specify
loss model default Design
Factors
You can
specify a
default
elevation
I16 - 9
User Options – Parameter Types
 Specify default flow rate type and head/pressure type
I16 - 10
User Options – Waterhammer Options
 Set Waterhammer specific Options
I16 - 11
Customizing Graphs
 Graph colors, fonts, axes, line styles, etc., can be customized

from the Formatting flyout area accessed on the Quick Access
Panel or Graph Results Toolbar, or by right-clicking on the
graph or axis itself
I16 - 12
Database Concept
 Internal database files -

– Internal Default database - fluid properties, fittings & losses, etc.
• This database is read only
• Impulse5.dat - located in the application folder
– Local User database - customizable file that may contain piping,
components, fluids, formatting, report configurations, etc.
• Imp_user5.dat - located in the \Documents and
Settings\UserName\Application Data folder
– Pipe Material Standards database - pipe materials from ANSI,
etc.
• This database is read only
• Steel - ANSI.dat - located in the Pipe Material Databases
folder below the application folder
I16 - 13
Database Concept (2)
 External database files

– Customizable files created in Database Manager - located on
any connected drive
– Read only as listed in database.lib - database.lib located in
application folder
 Custom databases are data used in the model building
process that can be saved and reused in the future
 By reusing previously entered data users can
– Save time by not re-entering data
– Control quality by using previously qualified data
I16 - 14
Custom Data Features
 AFT Impulse offers custom databases for these types of data:

– Components (Junctions)
– Fluid properties (Fluid Database)
– Pipe sizes (Material Database)
– Fittings & losses (Fittings Database)
– Solids properties (Solids Database used in SSL module)
 Databases can be both private and shared
– Private and shared data is merged in memory together when
AFT Impulse loads
 Multiple shared databases can be created and optionally
connected into a model
I16 - 15
Component Databases
 A junction can be added to the Component Database

 To add to the Component Database
– Enter data for the junction in its Properties Window
– Click on the junction to ensure it is highlighted
– From the Database Menu, select Add Component to Database
– Enter a name for the custom component
I16 - 16
Component Databases (2)
 To use the custom component

– Add a junction to the model of the same type as the custom
component
– Open the Properties Window for that junction
– Select the custom component name from the Database List
 To stop accessing the custom component choose None from
the Database List
 To edit the previous list select Component Database from the
Database Menu
I16 - 17
Fluid Databases
 A fluid and its associated properties (as they depend on

temperature) can be added to the Fluid Database
 To add to the Fluid Database
– From the Database Menu select Fluid Database
– Choose the Add New Fluid selection
– Enter a name and data for the fluid
• You can use raw data entered by hand or imported from file
• AFT Impulse will be perform a least-squares curve fit to the data
 To use the custom fluid
– Open the System Properties Window
– Select the custom fluid name from the Specified Fluid List
– Enter a temperature for the fluid
I16 - 18
Fluid Databases (2)
 To stop accessing the custom fluid choose Unspecified Fluid

 To edit the previous fluid data use the Fluid Database Window
and select Change Fluid Data
– You can also delete a fluid or change the fluid name
I16 - 19
Pipe Material Databases
 A pipe material and its associated sizes and schedules can be

added to the Material Database
 To add new material to the Material Database
– From the Database Menu select Pipe Material Database
– Click the New Material button
– Enter a name and data for the material entry, including the
friction data set
• You can use raw data entered by hand or imported from file
 To use the custom pipe material
– Open the Pipe Properties Window (or Global Pipe Edit)
– Select the custom material name from the Pipe Material List
– Select type and size
I16 - 20
Pipe Material Databases (2)
 To stop accessing the custom material choose (Unspecified)

 To edit the previous material data use the Pipe Material
Database Window and select the appropriate buttons
I16 - 21
Friction Data Sets
 Multiple Friction Data Sets can be associated with a pipe

material
 Friction Data Sets can also be associated with a particular
pipe size, or pipe type (e.g., schedule)
– Data Sets associated with Sizes apply to all Type sets within that
size
– Data Sets associated with a Type apply only to that type
 You can add Friction Data Sets to the default AFT Impulse
pipe materials
 Friction Data Sets are accessed in the Pipe Properties
window
I16 - 22
Local and Shared Network Databases
 The custom data you enter is stored in a single text file locally
on the hard drive
– This is called a local user database
– It is possible to copy this file to another user's hard drive so they
can use it too
– If you spend a lot of time entering custom data it is advisable to
back up the file
I16 - 23
Local and Shared Network Databases (2)
 You can create a network database if the AFT Impulse

software is being run from a network
– All users who run AFT Impulse from the network automatically
access the network database
– The network database is read-only, thus users cannot edit it
– Network databases are of great help in quality control because
data only need to be verified once
– Consult the Help System for detailed information on network
database management
 You can use the Database Manager capabilities to build new
databases and connect to selected databases across a local
or wide area network
I16 - 24
Using the Database Manager
 Databases can contain components, fluids, pipe materials,

insulation materials and program settings (Output Control,
User Options, etc.)
 The Database Manager allows you to find, connect to, and
use the information in these databases
I16 - 25
Using the Database Manager (2)
 The AFT Default Internal & AFT Impulse Local User

databases are ALWAYS connected!
The Click here to

available add another
databases engineering
are listed database to
here the Available
Databases
list
Currently
connected
databases
are listed
here
I16 - 26
 Contents of a connected database may be reviewed

Choose the database
Select the sections you
want to review and
Click on Review Content
Click here to display the

items in the selected
database section
I16 - 27
 In the Edit Database tab you can:

– Create a new database
– Select an existing database to edit
– Delete a database
 Data may be copied or moved from one database to another
Select which database will
be used for the source
data
Choose the section and
item(s) within that section
Click here to copy or move the selected

item(s) to the database being edited
I16 - 28
Chempak Fluid Property Database
 Chempak provides thermo-physical properties for ~700 fluids

– Both pure fluids and mixtures of these fluids may be used in the
model
I16 - 29
Chempak Fluid Property Database (2)
The Create New Mixture button
becomes visible when Chempak
Select Chempak Fluid or Mixture is selected
mixture
Select the fluid... Input

then select Add to temperature &
Model to use the pressure to
fluid calculate fluid
properties
To remove a
Select fluid
fluid, select it
property
here… then
variation here
select Remove
Fluid
Set accuracy
level here I16 - 30
Chempak Fluid Property Database (3)
 Fluids are selected from within the Create Mixture window to

assemble a mixture based on mass or mole fraction
Select a fluid
here Portion of the
mixture
Click here to add represented by
the fluid to the a fluid is set
mixture here
Any name may be Any or all

specified for a components
mixture may be
removed
Current mixture
components and
their fraction are
listed I16 - 31
I17. Introduction to AFT Impulse SSL
Module
Overview of Seminar
INTRODUCTION
Settling Slurry Module
 Incorporates the functionality of the SSL module for Fathom 8

– Predicts settling velocity to maintain acceptable operation
– Pump de-rating available
– Basic and Advanced slurry definition
– Solids data can be entered directly or from a database
 Extends slurry calculations by modifying the wavespeed and
bulk modulus to account for the solid particles
I17 - 1
Adjusting Wavespeed due to Solids
 The presence of solids changes the wavespeed

1
𝜌𝜌 1 − 𝐶𝐶𝑉𝑉 + 𝜌𝜌𝑠𝑠 𝐶𝐶𝑉𝑉
𝑐𝑐 2 = 𝑤𝑤
1 − 𝐶𝐶𝑉𝑉 𝐶𝐶𝑉𝑉 𝐷𝐷
+ + 𝜑𝜑
𝐾𝐾𝑤𝑤 𝐾𝐾𝑠𝑠 𝐸𝐸𝐸𝐸
Where:
𝜌𝜌𝑤𝑤 = liquid density
𝐾𝐾𝑤𝑤 = liquid bulk modulus
𝐷𝐷 = pipe internal diameter
𝐸𝐸 = Young’s modules for pipe wall
𝑡𝑡 = pipe wall thickness
𝜑𝜑 = pipe restraint factor
𝐶𝐶𝑉𝑉 = volumetric concentration of solid particles
𝐾𝐾𝑠𝑠 = solid particle bulk modulus (sand ~ 2.32E+6 psi, 16.0 GPa)
𝜌𝜌𝑠𝑠 = solid density
I17 - 2
SSL Module Example
 Description: Sand is being transferred from a quarry sand pit

to receiving deposits at two different locations
 Requirement: The ratio of the mixture velocity (Vm) to the
velocity at which the sand will settle out of the liquid and form
a stationary bed on the bottom of the pipe (Vsm) must be kept
greater than 1.2 for continual operation
 Determine: If the selected pipe sizes and velocity ratios are
acceptable during the transient where one valve starts to
close while the other opens further
– Change the pipe sizes to meet the Vm/Vsm requirement if
needed
I17 - 3
SSL Module Example (2)
 Open “Sand Transfer to Two Locations - SSL.imp” and

change P1 and P2 to 10 inch pipe
I17 - 4
 The Design Alert for Minimum Velocity Ratio was hit for pipe
P4, station 17 at 6.28 seconds
I17 - 5
 The profile graph of the velocity ratio from the sand pit to
deposit #1 shows the ratio falling below the minimum allowed
I17 - 6
 There are several factors that contribute to this problem

– Pipe P4 is slopped upward which increases the settling velocity
• Thus decreasing the velocity ratio
– Pipes P1 and P2 are the same size as the other pipes
• However, these pipes have a much larger flow rate than the other
pipes and therefore have a larger pressure gradient
 Increase the diameter of pipes P1 and P2 to 14 inch
– The size cannot be increased too much or the velocity ratio in
these pipes will fall below the 1.2 minimum
 Rerun the model
I17 - 7
 With 14 inch pipes the velocity ratio is satisfactory for the

entire length of the pipeline
I17 - 8
 The pressure profile from the sand pit to deposit #1
I17 - 9
 Time comparison of the velocity ratio in the critical pipes P4

and P14
I17 - 10
I18. More AFT Impulse Hands-On
Modeling
Overview of Seminar
INTRODUCTION
Platinum Pipe Award
Recognizing Excellence in Piping and Ducting System Modeling

Using AFT Software
 Win AFT software for your company and iPad for you
 Enter in one of the following categories

– Use of software features and model creativity
– Correlation to test/field data
– Operational benefits and sustainability
 Review the rules and enter on our website

– http://www.aft.com/platinum-pipe-award
PPA - 1
Platinum Pipe Award
AFT FathomTM Instrumental in Chrysler Energy Savings Project:
Plant Saves $194,000 Per Year
2015 Platinum Pipe Award Winner – Operational Benefits and Sustainability
Chrysler used AFT Fathom to model four
large central coolant filter systems in the
Dundee Engine plant located in Dundee,
Michigan, USA.
The existing cooling system was energy
intensive. Chrysler wanted to develop a
model that matched existing conditions,
and then use that model to predict
energy savings for different capital
expenditure scenarios. Thirty operating
scenarios were evaluated over four
systems for the project.
PPA - 2
Platinum Pipe Award
Analysis Using AFT FathomTM Reveals Root Cause for Fire
Pump System Over-Pressurization
2015 Platinum Pipe Award Winner – Correlation to Test/Field Data
In an undisclosed facility’s fire

pump system, Charles Williamson,
P.E., Senior Engineering Analyst at
CB&I, used AFT Fathom to
determine the root cause of an
over pressurization problem during
pump starts and to make
recommendations for corrective
action.
PPA - 3
Platinum Pipe Award
AFT FathomTM Finds Optimization Opportunities in Hydrocarbon
Separation Process
2015 Platinum Pipe Award Winner – Use of Software Features and Model
Creativity
Mr. Nelius Joubert, Senior
Operations Specialist at Sasol, used
AFT Fathom to model a complicated
hydrocarbon separation plant in
South Africa. The plant utilizes liquid
extractants to separate the
components of a liquid stream into
separate pure components.
The model will be used on a
continuous basis for troubleshooting
and evaluation of the plant’s
hydraulic health.
Liquid Extraction and Distillation Portion of the AFT Fathom Model
PPA - 4
AFT Notifications & Seminar Evaluation
 Please complete the AFT Notifications form (if you would like)
and the Technical Seminar Evaluation Form, and hand to
instructor when complete.
US
Hands-on Models
 Work these models in whatever order you prefer and at your

own pace
– Do not open these from disk!
– They are intended to be built from scratch unless otherwise
noted
– All models are included on your seminar disk so you can review
them at a later date
I18 - 1
US
Hands-on Models (2)
1. Spray water system, TEST8

Determine time for flow to come up to full flow
2. Pump with flow control transient, "Pump With Flow Controller.imp"
A pump with a flow control device has the control flow rate reduced over 5
seconds. Find the maximum and minimum pressures during the transient.
3. Pump trip with accumulator, TEST2
A pump trips and the pressure in some parts of the discharge pipe falls
below atmospheric. Size and locate a gas accumulator to maintain all
positive pressures.
4. Pump trip with backflow, TEST4
A pump trip where backflow occurs is modeled using four quadrant data.
Determine the maximum and minimum pressures in the system.
5. Cooling system pump trip with backflow, "Pump Trip with Backflow.imp"
A cooling system with two pumps experiences a pump trip. By the time the
valve closes, backflow has already occurred in the pumps. Find the
maximum and minimum pressures and look at the pump speed decay.
I18 - 2
US
Model a Spray System
 Find how long it takes for the flowrate to come up to the full
flow of 100 gpm at each spray from the closure state (Model
TEST8.IMP)
 Pipe data:
– Steel - ANSI pipe, all schedule 40, standard roughness of
0.00015 feet
 Fluid is water at 70 deg. F
 Inlet stagnation pressure is 171 psig
 Spray nozzle data:
– Sprays discharge to atmosphere and open in 0.1 second
– Flow Area = 0.5 square inches, Discharge coefficient = 0.6
I18 - 3
US
Model a Spray System (2)
Time (sec) CdA (in2)
0 0
0.1 0.3
10 0.3
El=10.0 feet
Typical
I18 - 4
US
Model Pump Flow Controller Transient
 Find maximum and minimum static pressure in system when

pump flow set point is changed ("Pump With Flow
Controller.imp")
 System uses water at 100 degrees F
 Pipe layout, sizes and elevations shown below
I18 - 5
US
Model Pump Flow Controller Transient (2)
 Pump data is:

Q (gpm) H (ft)
0 60
2000 55
4000 40
– Steady state flow control is 2000 gpm

– Flow transient is as follows:
time (sec) Q (gal/min)
0 2000
5 1000
10 1000
 Heat exchanger head loss is:

Q (gpm) H (ft)
0 0
1000 20
2000 80 I18 - 6
US
 Valve at J11 and J21 steady-state Cv is 500

 Valve at J21 closes as follows:
time (sec) Cv
0 500
0.5 100
1 0
10 0
 Reservoir J1 has pipes at 20 feet depth
I18 - 7
US
Add Accumulator to Pump Model
 Size and locate an accumulator to prevent sub-atmospheric

pressure
 Start from the model you built in Section 10-5 (TEST2.IMP)
 The accumulator will be connected to the system through an
integral short connector pipe with data as shown below
– Assume a polytropic coefficient of 1.2
I18 - 8
US
Model Pump Trip With Backflow

1 2 3
P1 P2
L = 1500 ft L = 1800 ft
Steel - ANSI (STD) Steel - ANSI (Sch 30)
30 inch 30 inch
Surface Elev. = 190 ft
 Fluid is water at 40 degrees F Surface Pressure = 1 atm
 Use four sections in the controlling pipe Pipe Depth = 30 ft
 Run simulation for 15 seconds
I18 - 9
US
Model Pump Trip With Backflow (2)
 Pump data:
– Pump is the inlet junction at J1, modeled as submerged with
HGL = 0 feet at suction
– No check valve - pump can flow and spin backwards
– Pump elevation = 0 feet
– Pump curve data is in a comma separated data file "test4 pump
curve data.txt"
• Units of imported data: Q in gal/min, H in feet, efficiency in decimal
• Use an “Interpolated X-Y Data” Curve Fit Type
– Transient model is "Trip With Inertia - Four Quadrant"
– Pump Speed = 1100 rpm
– Rotating Inertia = 400 lbm-ft2
I18 - 10
US
 Pump data (cont.):

– This pump represents 2 parallel pumps - specify on Optional tab
– Number of Data Points for Four Quadrant Data = 72
– Use four quadrant data is in a comma separated data file "test4
four quad data.txt"
α
– Theta is based on
υ
 J2 is a branch with elevation = 0 feet
– This represents a change in pipe properties (wavespeed and
diameter)
I18 - 11
US
Model Cooling System With Backflow
 A cooling water system

has two pumps which
trip ("Pump Trip with
Backflow.imp")
 The valves downstream
of the pump close, but
not before backflow
occurs through the
pumps
 What are the maximum
and minimum pressures
in the system?
 How does the pump
speed decay? I18 - 12
US
Model Cooling System With Backflow (2)
 Assume water at 60 degrees F

 Pipes in cooling basins are four feet below surface
 Each cooling tower has an overall flow area of 10 square feet
with a discharge coefficient of 0.6
 Pump data is: Q (gpm) H (ft) Power (hp)
0 112 105
2000 108 130
4000 104 160
6000 98 190
8000 90 210
10000 78 220
12000 60 210
– Initial speed is 1790 rpm , Pump Inertia is 140 lbm-ft2
– Transient model is “Trip With Inertia – Four Quadrant”
– Four Quadrant Data Source Ns = 2.78
I18 - 13
US
 Valves have the following data:

Time (sec) Cv
0 10000
5 1000
12 0
30 0
 The J11 condenser has the following loss data:

Q (gpm) H (ft)
0 0
20000 22
40000 88
I18 - 14
US
Answers to Hands-On Models
 TEST8.IMP
– It takes a little over 0.7 second for the final spray to reach 100
gpm
– After 1 second the flow drops below 100 gpm
Nearest Supply
Farthest From Supply I18 - 15

US
Answers to Hands-On Models (3)
 "Pump With Flow Controller.imp"

– Maximum pressure in pipe P21, station 1 = 55.2 psig
– Minimum pressure in pipe P23, station 0 = -5.3 psig
I18 - 16
US
 "Pump With Flow Controller.imp"

– Other output of interest (initial pump speed is 94%)
I18 - 17
US
 TEST2.IMP with accumulator

– There are many solutions to this problem
– We located the accumulator 495 feet downstream from the pump
and made the initial volume 6.0 cubic feet
– With this location and size, the following profile results
I18 - 18
US
 TEST4 results
– Max/Min static pressure profile and pump speed decay (pump
spins backwards)
– Max/Min static pressure
120 psig / 4.2 psig
I18 - 19
US
 TEST4 additional results

– Flowrate and HGL at pump discharge
I18 - 20
US
 "Pump Trip with Backflow.imp"
I18 - 21
SI
Hands-on Models
 Work these models in whatever order you prefer and at your

own pace
– Do not open these from disk!
– They are intended to be built from scratch unless otherwise
noted
– All models are included on your seminar disk so you can review
them at a later date
I18 - 22
SI
Hands-on Models (2)
1. Spray water system, TEST8 (SI)

Determine time for flow to come up to full flow
2. Pump with flow control transient, "Pump With Flow Controller (SI).imp"
A pump with a flow control device has the control flow rate reduced over 5
seconds. Find the maximum and minimum pressures during the transient.
3. Pump trip with accumulator, TEST2 (SI)
A pump trips and the pressure in some parts of the discharge pipe falls
below atmospheric. Size and locate a gas accumulator to maintain all
positive pressures.
4. Pump trip with backflow, TEST4
A pump trip where backflow occurs is modeled using 4-quadrant data.
Determine the maximum and minimum pressures in the system.
5. Cooling system pump trip with backflow, "Pump Trip with Backflow (SI).imp"
A cooling system with two pumps experiences a pump trip. By the time the
valve close, backflow has already occurred in the pumps. Find the
maximum and minimum pressures and look at the pump speed decay.
I18 - 23
SI
Model a Spray System
 Find how long it takes for the flowrate to come up to the full
flow of 22.7 m3/hr at each spray from the closure state (Model
TEST8 (SI).IMP)
 Pipe data:
– Steel – ANSI pipe, all schedule 40, standard roughness of
0.004572 cm
 Fluid is water at 21 deg. C
 Inlet stagnation pressure is 1200 kPa
 Spray nozzle data:
– Sprays discharge to atmosphere and open in 0.1 second
– Flow Area = 3.23 square cm, Discharge coefficient = 0.6
I18 - 24
SI
Model a Spray System (2)
Time (sec) CdA (cm2)
0 0
0.1 1.94
10 1.94
El=3 meters
Typical
I18 - 25
SI
Model Pump Flow Controller Transient
 Find maximum and minimum static pressure in system when

pump flow set point is changed ("Pump With Flow Controller
(SI).imp")
 System uses water at 38 degrees C
 Pipe layout, sizes and elevations shown below
I18 - 26
SI
 Pump data is:

Q (m3/hr) H (m)
0 18
450 17
900 12
– Steady state flow control is 450 m3/hr

– Flow transient is as follows:
time (sec) Q (m3/hr)
0 450
5 230
10 230
 Heat exchanger head loss is:

Q (m3/hr) H (m)
0 0
225 6
450 24 I18 - 27
SI
 Valve at J11 and J21 steady-state Cv is 500

 Valve at J21 closes as follows:
time (sec) Cv
0 500
0.5 100
1 0
10 0
 Reservoir J1 has pipes at 6 meters depth
I18 - 28
SI
Add Accumulator to Pump Model
 Size and locate an accumulator to prevent sub-atmospheric

pressure
 Start from the model you built in Section 10-9 (TEST2 (SI).IMP)
 The accumulator will be connected to the system through an
integral short connector pipe with data as shown below
– Assume a polytropic coefficient of 1.2
I18 - 29
SI
Model Pump Trip With Backflow

1 2 3
P1 P2
L = 450 m L = 550 m
Steel-ANSI (STD) Steel-ANSI (Sch 30)
30 inch 30 inch
(74.3 cm ID) (73.0 cm ID)
Surface Elev. = 59 m
 Fluid is water at 4 degrees C Surface Pressure = 1 atm
 Use four sections in the controlling pipe Pipe Depth = 10 m
 Run simulation for 15 seconds
I18 - 30
SI
 Pump data:
– Pump is the inlet junction at J1, modeled as submerged with
HGL = 0 meters at suction
– No check valve - pump can flow and spin backwards
– Pump elevation = 0 meters
– Pump curve data is in a comma separated data file "test4 pump
curve data (SI).txt"
• Units of imported data: Q in m3/sec, H in meters, efficiency in
decimal
• Use an “Interpolated X-Y Data” Curve Fit Type
– Transient model is "Trip With Inertia - Four Quadrant"
– Pump Speed = 1100 rpm
– Rotating Inertia = 16.85 kg-m2
I18 - 31
SI
 Pump data (cont.):

– This pump represents 2 parallel pumps - specify on Optional tab
– Number of Data Points for Four Quadrant Data = 72
– Use four quadrant data is in a comma separated data file "test4
four quad data.txt"
α
– Theta is based on
υ
– J2 is a branch with elevation = 0 meters
– This represents a change in pipe properties (wavespeed and
diameter)
I18 - 32
SI
Model Cooling System With Backflow
 A cooling water system

has two pumps which trip
("Pump Trip with
Backflow (SI).imp")
 The valves downstream
of the pump close, but
not before backflow
occurs through the
pumps
 What are the maximum
and minimum pressures
in the system?
 How does the pump
speed decay? I18 - 33
SI
 Assume water at 16 degrees C

 Pipes in cooling basins are 1.2 meters below surface
 Each cooling tower has an overall flow area of 0.9 square
meters with a discharge coefficient of 0.6
3
 Pump data is: 0Q (m /hr) H34
(m) Power (kW)
78
450 33 97
900 32 119
1400 30 142
1800 27 157
2300 24 164
2700 18 157
– Initial speed is 1790 rpm, Pump Inertia is 5.9 kg-m2
– Transient model is “Trip With Inertia – Four Quadrant”
– Four Quadrant Data Source Ns = 2.78
I18 - 34
SI
 Valves have the following data:

Time (sec) Cv
0 10000
5 1000
12 0
30 0
 The J11 condenser has the following loss data:

Q (m3/hr) H (m)
0 0
4500 7
9000 27
I18 - 35
SI
Answers to Hands-On Models
 TEST8 (SI).IMP
– It takes a little over 0.85 seconds for the final spray to reach 22.7
m3/hr
– After slightly less than 1 second the flow drops below 22.7 m3/hr
Nearest Supply
Farthest From Supply I18 - 36

SI
 "Pump With Flow Controller (SI).imp"

– Maximum pressure in pipe P2, station 76 = 375.9 kPa(g)
– Minimum pressure in pipe P23, station 0 = -34.5 kPa(g)
I18 - 37
SI
 "Pump With Flow Controller (SI).imp"

– Other output of interest (initial pump speed is 93%)
I18 - 38
SI
 TEST2 (SI).IMP with accumulator

– There are many solutions to this problem
– We located the accumulator 151 meters downstream from the
pump and made the initial volume 0.17 cubic meters
– With this location and size, the following profile results
I18 - 39
SI
 TEST4 (SI).imp results

– Max/min static pressure profile and pump speed decay (pump
spins backwards)
– Max/Min static pressure
9.43 / 1.323 bar
I18 - 40
SI
 TEST4 (SI).imp additional results

– Flowrate and HGL at pump discharge
I18 - 41
SI
 "Pump Trip with Backflow (SI).imp"
I18 - 42

AFT Impulse Instructions

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AFT Impulse Instructions

Uploaded by

Copyright:

Available Formats

AFT Impulse Seminar

Week of 22 May, 2017

 Applied Flow Technology (AFT), founded in 1993, is a world

 Models incompressible network pipe systems

 XTS – eXtended Time Simulation

 Models compressible network pipe systems

 GSC – Goal Seek & Control

 Models waterhammer/surge flow in pipe networks

 SSL – Settling Slurry simulation

 Chempak™ Property Database

 AFT products are being successfully applied to a broad range

 Provides consulting services beyond typical technical support

 Additional ways to use your hours:

 General purpose pipe network waterhammer and surge

 Can model systems in any generalized configuration

 Branching section (up to 25 pipes)

 No practical software limit to model size

 The AFT Impulse modeling process flows through five Primary

 Ability to activate SSL Module

 Cut, copy, paste, delete, duplicate and undo features

 The Reference Flow Direction of a pipe can be changed

 The Workspace image can be printed on printers and plotters

 Model Data is broken into three sections

 With a Workspace printout and the complete Model Data

 The Output window is the primary vehicle for communicating

 User can select all or portions of the Output Window content

 Visual Report allows user to display input and output results

 The Graph Results Window allows creation of full-featured

 AFT Impulse uses the Newton-Raphson Method to solve the

 Conflicting definitions exist in the literature for total and

 Momentum Equation (Bernoulli)

 The dynamic pressure and static pressure can be combined

Po ,1 + ρgh1 = Po , 2 + ρgh2 + ∆Ploss

 Traditional method of friction loss calculation uses the Darcy-

 Bernoulli’s equation then becomes

 Applying law of mass conservation to a branching section

 Substituting yields the following equation to be solved for

 where sgn = ±1 depending on flow direction

 We need to solve as many equations as there are flow splits

 The procedure for applying Newton-Raphson to a single

3. Substitute the improved guess back into the above

 When applied to a system of equations with P as the

where P is the vector of pressures and JF is the Jacobian matrix of error

 The diagonal derivative terms in the Jacobian can be

 The off-diagonal terms can also be calculated analytically

 Rather than inverting the Jacobian matrix, it is usually faster

 We need to solve for the values in vector, z, that satisfy the

pipe f L (ft) D (in) Fluid

− 0.5 0.5 − 0.5 0.5 − 0.5 0.5

P (psia), M (lbm/s) Solution shown in Test1.xls

 When the change in pressures and flow rates decreases to

 The previous example focused on the pressure solution

 Waterhammer is a transient phenomenon that occurs in a

 Waterhammer can occur in any liquid piping system

 Waterhammer can be caused by different physical

 The magnitude of a waterhammer transient is dependent on

 It can be shown that the maximum theoretical pressure surge

– Example: Density is 62 lbm/ft3, Wavespeed is 2500 feet per

Pmax  384 psia

 It can be shown that the maximum theoretical pressure surge

– Example: Density is 1000 kg/m3, Wavespeed is 1000 meters per

 The communication time is a helpful concept in understanding

 Any event that occurs in a time frame less than the

 Instructor – show video files