You are on page 1of 14

Department of Mechanical Engineering, IUPUI

ME 414 Thermal-Fluid Systems Design

Project 2: Heat Exchanger Optimization

Instructor: John Toksoy


May 6, 2005

Group Members:
Luke Jones
Justin Gast
Mike Hughett

1
Problem Statement
‰ Design a heat exchanger given 80,000kg/hr of distilled water will
enter at 35°C and leave at 25°C and transfer heat to
140,000kg/hr raw water entering from a 20°C supply.
ƒ Desired heat transfer rate = m& C p ∆T = 928.5 kW
ƒ No baffles, neglect fouling, single pass.

‰ Optimize the weight, shell and tube pressure drops, and heat
transfer of the design using the DOE capabilities of both Matlab
and Minitab software.

2
Tools Utilized
‰ Matlab
ƒ Utilized the provided Matlab code to perform the heat exchanger
analysis
‰ Minitab
ƒ Used in the selection of critical design parameters
ƒ Provided tools needed to optimize Matlab heat exchanger design
calculations
ƒ Aided in optimization
‰ Iterative optimization process

3
Where to Start?
‰ Input given values from problem definition
‰ Obtained desired to calculated heat transfer ratio of 1 by trial and
error
‰ Ran DOE study using Minitab to find the main effects of the
variables and their interactions
‰ Eliminated insignificant variables

4
Funnel Effect
Shell ID, Tube OD, Length, Tube Material, Shell
Thickness, Fluid Allocation, Layout Angle,
Shell Thickness

Minitab

2-3 Critical
Variables
5
Main Effects Plots

6
Design Decisions
‰ Counter Flow
ƒ Parallel Flow Not an Option Æ
‰ 1.25 Pitch Ratio (rule of thumb)
‰ Square Pitch
ƒ Clean surfaces
ƒ 90 degree layout angle
‰ Tube Material
ƒ Aluminum: ▲Heat Transfer ▼Low Weight
‰ Shell Thickness set to 1 mm (determined from
hoop stress analysis)

7
Elimination from Evaluation
‰ After more Main effects plots were run, the 3 key variables
discovered were: length, tube OD, and shell ID

‰ Next, a multi-level DOE was run in Matlab to determine good


starting points for design optimization

8
Main Effects of 3 Critical Parameters

9
Heat Exchanger Optimization
‰ Analyzed Factorial Design to create Pareto charts of design parameters.
This shows the weight each variable has on the design specification
‰ Verified that the statistical p-values were below 0.1

10
Iterative Optimization
DOE 1 DOE 2 DOE 3

+/- 20% +/- 15% +/- 10%


(Matlab (Matlab (Matlab
Check) Check) Check)

DOE 4 DOE 5
Matlab Results:
Weight = 1051 kg
+/- 5% ∆P Tube = 978 Pa
(Matlab (Matlab
Check) Check)
∆P Shell = 914 Pa
Q = 928.6 kW
11
Cost Consideration
‰ While custom parts provide the most efficient heat exchanger design,
manufacturing costs must be considered in the Total Cost of Ownership

TCO = Initial Costs + Maintenance + Repairs


‰ Using standard tube sizes greatly reduces initial costs, thereby reducing the TCO

Standard Tube and Shell Size Optimization: Selected Material Sizes:


‰ Weight = 1005 kg ‰ Shell Diameter: 21.25 inches
‰ Heat transfer rate = 928.3 kW ‰ Tube Diameter: 20BWG ½ inch
ƒ Desired-to-calculated ratio of 1.00 ‰ Tube Length: 3.477 meters*
‰ Shell side pressure drop = 788 Pa
‰ Tube side pressure drop = 687 Pa * There is no defined standard length

Even Better than the Minitab Optimization!!


12
Conclusions
‰ Heat Exchanger optimization was a success
‰ The standard tube and shell diameters provides the optimal weight,
tube and shell pressure drops, and desired heat transfer
‰ One concern: the average tube velocity is 0.28 m/s for our optimal
design, which is lower than the recommended velocity to prevent
settling
ƒ Because distilled water is being used in the tubes, settling is
unlikely
‰ TCO of our design is minimized:
ƒ Low material weight Æ initial costs minimized
ƒ Low pressure drops Æ initial costs and operational costs
minimized
ƒ Square pitch Æ maintenance costs minimized (time=money!!)
13
Questions?

14

You might also like