MONITORING AND CONTROL OF HYDRO POWER PLANT

Matei Vinatoru


University of Craiova, Faculty of Automation, Computers and Electronics
Str. A. I. Cuza Nr. 13, Craiova, RO 200396





Abstract: In hydro power plants from Romania, there is a major interest for the
implementation of digital systems for monitoring and control replacing the conventional
control systems for power, frequency and voltage. Therefore is necessary to develop
mathematical models capable to accurately describe both dynamic and stationary
behaviour of the hydro units, in order to be able to implement digital control algorithms.
Moreover, it is necessary to implement systems for monitoring and control of hydro
power plants in a cascade system along a river, in order to optimize the use of the river
resources. This paper presents the possibilities of modelling and simulation of the hydro
power plants and performs an analysis of different control structures and algorithms.


Keywords: Hydro power plant, Control system, Digital Control Algorithms.





1. INTRODUCTION

This paper discusses the aspects of modelling and
design of hydro power plants and control of hydro
power groups. There are presented computing
methods for pressure losses on the water intake pipes
from the reservoir to the turbine, aspects regarding
the energy transformations for different turbine types
(Pelton, Francis, and Kaplan) in order to determine
the net hydraulic power of the plant and to determine
the control options for the turbines.

Also, are presented different control structures for
the power groups and hydro power plants. In this
case, are analyzed different SCADA monitoring and
control systems, for an assemble of hydro power
plants on the same water stream as well as SCADA
system for high-energy power plants with Kaplan
turbines, and the case studies for the modelling,
design and study of the power groups, with and
without water towers.

This paper will present several possibilities for the
modelling of the hydraulic systems and the design of
the control system.

2. THE DESIGN OF CONTROL SYSTEMS FOR
KAPLAN HYDRAULIC TURBINES

2.1. Introduction

Different construction of hydropower systems and
different operating principles of hydraulic turbines
make difficult to develop mathematical models for
dynamic regime, in order to design the automatic
control systems. Also, there are major differences in
the structure of these models. Moreover, there are
major differences due to the storage capacity of the
reservoir and the water supply system from the
reservoir to the turbine (with or without surge
chamber). The dynamic model of the plants with
penstock and surge chamber is more complicated
than the run-of-the-river plants, since the water feed
system is a distributed parameters system.


2.2. Modelling of the hydraulic system for run-of-
the-river hydropower plants

These types of hydropower plants have a low water
storage capacity in the reservoir; therefore the plant
operation requires a permanent balance between the
water flow through turbines and the river flow in
IFAC Workshop ICPS'07
2007, July 09-11
Cluj-Napoca, Romania
-44-


order to maximize the water level in the reservoir for
a maximum efficiency of water use. Next, we will
determine the mathematical model for each
component of the hydropower system.

A. Hydraulic turbine. The hydraulic turbine can be
considered as an element without memory since the
time constants of the turbine are less smaller than the
time constants of the reservoir, penstock, and surge
chamber, if exists, which are series connected
elements in the system.
As parameters describing the mass transfer and
energy transfer in the turbine we will consider the
water flow through the turbine Q and the moment M
generated by the turbine and that is transmitted to the
electrical generator. These variables can be
expressed as non-linear functions of the turbine
rotational speed N, the turbine gate position Z, and
the net head H of the hydro system.
Q = Q(H, N, Z) (1)

M = M(H, N, Z) (2)
Through linearization of the equations (1) and (2)
around the steady state values, we obtain:

z a n a h a q
Z
Z
Q
N
N
Q
H
H
Q
Q
. . .
13 12 11
+ + =

=
(3)
z a n a h a m
Z
Z
M
N
N
M
H
H
M
M
. . .
23 22 21
+ + =

=
(4)
where the following notations were used:
0
Q
Q
= q ,
0
N
N
= n ,
0
M
M
= m ,
0
H
H
= h ,
0
Z
Z
= z which
represent the non-dimensional variations of the
parameters around the steady state values.

B. The hydraulic feed system. The hydraulic feed
system has a complex geometrical configuration,
consisting of pipes or canals with different shapes
and cross-sections. Therefore, the feed system will
be considered as a pipe with a constant cross-section
and the length equal with real length of the studied
system. In order to consider this, it is necessary that
the real system and the equivalent system to contain
the same water mass. Let consider m
1
, m
2
... m
n
the
water masses in the pipe zones having the lengths l
1
,
l
2
,...,l
n
and cross-sections A
1
, A
2
,...,A
n
of the real feed
system. The equivalent system will have the length
L=l
1
+l
2
+...+l
n
and cross-section A, conveniently
chosen. In this case, the mass conservation law in
both systems will lead to the equation:

= =
=
n
i
i i
n
i
i
A l l A
1 1
. (5)
Since the water can be considered incompressible,
the flow Q
i
through each pipe segment with cross-
section A
i
is identical and equal with the flow Q
through the equivalent pipe

Q=v.A=Q
i
=v
i
.A
i
for i=1, 2,...,n (6)
Where v is the water speed in the equivalent pipe,
and v
i
is the speed in each segment of the real pipe.
From the mass conservation law it results:
Q
A l
l
A l
l Q
A
Q
v
i i
i
i i
i i
.
. .

= = = (7)
The dynamic pressure loss can be computed
considering the inertia force of the water exerted on
the cross-section of the pipe:

dt
dv
g
L
A a A L a m F
i
.
.
. . . .

= = =

(8)
Where L is the length of the penstock or the feed
canal, A is the cross-section of the penstock, is the
specific gravity of water (1000Kgf/m
3
), a is the water
acceleration in the equivalent pipe, and g=9,81 m/s
2

is the gravitational acceleration.
The dynamic pressure loss can be expressed as:

dt
dQ
A l
l
g
L
dt
dv
g
L
A
F
H
i i
i i
d

= = =
.
. .
(9)
Using non-dimensional variations, from (9) it results:

dt
Q
Q
d
A l
l L
H
Q
H
H
i i
i
d d
d
|
|
.
|

\
|

0
0
0
0

(10)
Or in non-dimensional form:

dt
dq
T h
w d
= (11)
Where T
W
is the integration constant of the
hydropower system and the variables have the
following meaning:

=
i i
i
d
w
d
d
d
A l
l L
H
Q
T
Q
Q
q
H
H
h

0
0
0 0
, , (12)
It must be noted that this is a simplified method to
compute the hydraulic pressure loss, which can be
used for run-of-the-river hydropower plants, with
small water head. If an exact value of the dynamic
pressure is required, then the formulas presented in
[8], sub-chapter 8.4 The calculation of hydro energy
potential shall be used.
Using the Laplace transform in relation (11), it
results:
h
d
(s) = - sT
w
.q(s), and, ) (
1
) ( s h
sT
s q
d
w
= (13)
Replacing (13) in (3) and (4) and doing some simple
calculations, we obtain:

) (
1
) (
1
) (
11
13
11
12
s z
s T a
a
s n
s T a
a
s q
w w
+
+
+
= (14)
) (
1
) (
1
) (
11
13
11
12
s z
s T a
s T a
s n
s T a
s T a
s h
w
w
w
w
d
+

+
=

(15)
) (
1
) (
1
) (
11
13
23
11
12
21
s z
s T a
s T a
a s n
s T a
s T a
a s m
w
w
w
w
|
|
.
|

\
|
+
+
|
|
.
|

\
|
+
=

(16)
The mechanical power generated by the turbine can
be calculated with the relation P=..Q.H (see [8]
sub-chapter 8.5 Hydraulic turbines), which can be
used to obtain the linearized relations for variations
of these values around the steady state values:
p = .g.Q
o
.h + .g.H
o
.q (17)
-45-


Where is the turbine efficiency, and , Q, and H
were defined previously.
On the other hand, the mechanical power can be
determined using the relation P=M.=2.M.N,
which can be used to obtain the linearized relations
for variations of these values around the steady state
values:

( )
m
P
N
p
P
N
n
o
o
o
o
2
2 2
= (18)
Where P
0
=M
0
.
0
is the steady state power generated
by the turbine for a given steady state flow Q
0
and a
steady state head H
0
, and N
0
is the steady state
rotational speed.
Using these relations, the block diagram of the
hydraulic turbine, for small variation operation
around the steady state point, can be determined and
is presented in figure 1, where the transfer functions
for different modules are given by the following
relation:

|
|
.
|

\
|
+
=
|
|
.
|

\
|
+
=
+
=
+
=
+
=
+
=
s T a
s T a
a s H
s T a
s T a
a s H
s T a
s T a
H
s T a
s T a
s H
s T a
a
s H
s T a
a
s H
w
w
mz
w
w
mn
w
w
hz
w
w
hn
w
qz
w
qn
11
13
23
11
12
21
11
13
11
12
11
13
11
12
1
) ( ,
1
) ( ,
1
,
1
) ( ,
1
) ( ,
1
) (

(19)

For an ideal turbine, without losses, the coefficients
a
ij
resulted from the partial derivatives in equations
(12 - 16) have the following values: a
11
=0,5;
a
12
=a
13
=1; a
21
=1,5; a
23
=1. In this case, the transfer
functions in the block diagram are given by the
following relation:

s T . ,
s T
H ,
s T . ,
H H
w
w
hn
w
qz qn
5 0 1 5 0 1
1
+
=
+
= = (20)


|
|
.
|

\
|
+
=
+
=
s T . ,
s T
, H ,
s T . ,
s T
H
w
w
mn
w
w
hz
5 0 1
5 1
5 0 1
(21)
s T . ,
s T . .
s T . ,
s T
) s ( H
w
w
w
w
mz
5 0 1
5 0 1
5 0 1
1
+

=
|
|
.
|

\
|
+
= (46)
2.3. Simulation results
Example. Let consider a hydroelectric power system
with the following parameters:
-Water flow (turbines): Q(5001000) m
3
/s, Q
N
=725
m
3
/s;
-Water level in the reservoir: H(1738) m, H
N
=30
m;
-The equivalent cross-section of the penstock
A=60m
2
;
-Nominal power of the turbine
P
N
=178MW=178.000kW;
-Turbine efficiency =0,94;
-Nominal rotational speed of the turbine N=71,43
rot/min;
-The length of the penstock l=l
i
=20m;

It shall be determined the variation of the time
constant T
W
for the hydro power system.
For the nominal regime, using relation (12), where
l
i
=20m, the time constant of the system is:
s T
w
82 , 0
60 81 , 9
20
30
725
=

= (47)
Next we will study the variation of the time constant
due to the variation of the water flow through the
turbine for a constant water level in the reservoir,
H=30m, as well as the variation due to the variable
water level in the reservoir for a constant flow
Q=725 m
3
/s.
In table I, column 3 and figure 2 a) are presented the
values and the graphical variation of the time
constant T
W
for the variation of the water flow
between 500 m
3
/s and 110 m
3
/s, for a constant water
level in the reservoir, H=30m. In table 8.12 column 4
and figure 2 b) are presented the values and the
graphical variation of the time constant T
W
for the
variation of the water level in the reservoir, for a
constant water flow, Q=725m
3
/s.

It can be seen from the table or from the graphs that
the time constant changes more than 50% for the
entire operational range of the water flow through the
turbine or if the water level in the reservoir varies.
These variations will create huge problems during
the design of the control system for the turbine, and
robust control algorithms are recommended.
gQ
o
gH
o
H
qn
(s)
H
qz
(s)
H
hn
(s)
H
hz
(s)
H
mn
(s)
H
mz
(s)
2N
o
/P
o
(2N
o
)
2
/P
o
p
n p
n
n
n
z
z
z
z
m
h
q
+
Fig. 1. The block diagram of the hydraulic turbine.
-46-


Table I. Variations of the time constant of the hydro
system

H Q Tw(H=30m) Tw(Q=725mc/s
17 1135,46 1,286065 1,449101
20 965,14 1,093155 1,231736
23 839,25 0,950569 1,071075
26 742,42 0,8408882 0,947489
29 665,62 0,7538998 0,849473
32 603,21 0,6832217 0,769835
35 551,51 0,6246598 0,703849
38 507,97 0,5753446 0,648282
In figure 3 is presented the block diagram of the
turbines power control system, using a secondary
feedback from the rotational speed of the turbine. It
can be seen from this figure that a dead-zone element
was inserted in series with the rotational speed sensor
in order to eliminate the feedback for 0,5%
variation of the rotational speed around the
synchronous value.

The constants of the transfer functions had been
computed for a nominal regime T
W
=0,8s. The
optimal parameters for a PI controller are: K
R
=10,
T
I
=0,02s. The results of the turbine simulation for
different operational regimes are presented in figure
4, for a control system using feedbacks from the
turbine power and rotational speed, with a dead-zone
on the rotational speed channel for 0,5% variation
of the rotational speed around the synchronous value
(a) Power variation with 10% around nominal value,
b) Rotational speed variation for power control).

Figure 3. Block Diagram of the control system for hydraulic turbines
n p
0.0389

0.0016

s 4 , 0 1
1
+

0.0025

1.1305

p
n
n
n
z
z
z
z
m
h
q
+
s 4 , 0 1
1
+
s
s
4 , 0 1
8 , 0
+

s
s
4 , 0 1
2 , 0 5 , 1
+

s
s
4 , 0 1
8 , 0 1
+

s
s
4 , 0 1
8 , 0
+

|
|
.
|

\
|
+
s T
K
I
R
1
1

-
-
+
P
*
0,6
0,7
0,8
0,9
1
1,1
1,2
1,3
1,4
1,5
15 20 25 30 35 40
[s]
[m]
H
T
w
b)
0,5
0,6
0,7
0,8
0,9
1
1,1
1,2
1,3
1,4
500 600 700 800 900 1000 1100 1200
T
w

Q
[s]
[mc/s] a)
Figure 2 Variation of the integral time constant T
W
: a) by the flow Q, b) by the water level H
-47-





























Fig. 4. Turbine Control simulations with feedbacks
from turbine power and rotational speed































Fig. 5 Control structure with only power feedback,
a) Power variation with 10% around nominal value,
b) Rotational speed variation for power control

In figure 5 are presented the variations of the turbine
power (graph a) and rotational speed (graph b) for
the control system using a feedback from the turbine
power but no feedback from the rotational speed .

This oscillation has no significant influence on the
performance of the system but would have lead to
permanent perturbation of the command sent to the
turbine gate.

2.4. Conclusions
The possibility of implementation of digital systems
for monitoring and control for power, frequency and
voltage in the cascade hydro power plant was
discussed. The simplified mathematical models,
capable to accurately describe dynamic and
stationary behaviour of the hydro units have been
developed and simulated. These aspects are
compared with experimental results. Finally, a
practical example was used to illustrate the design of
controller and to study the system stability.


3. IMPLEMENTATION OF DIGITAL SYSTEMS
FOR MONITORING AND CONTROL IN
CASCADE POWER PLANTS

3.1. Introduction

A series of hydro power plant systems in Romania
are capturing the river water in a geographical area
and store it in reservoirs connected through transport
pipes. Every lake has its own hydro power plant and
the water used through the turbines of one plant is
sent to the next reservoir, thus using the entire
hydraulic potential offered by the geographical area.
A similar hydro power system is built in the north-
west part of Oltenia region, in the south-west part of
Romania. The hydraulic profile of this cascade
system is presented in figure 6.

As it can be seen in this figure, there are two hydro
power plants, CH1 and CH2, with reservoirs and
surge tanks and one run-of-the-river power plant, CH
3, with a reservoir providing only a net head.

In this type of hydropower systems, provided with
small power units connected to the grid, the dynamic
of the hydraulic system has a great importance in the
stability of the system. The dynamics of the turbine
and electric generator have a very small influence
since the power grid has a great influence and will
maintain the generator rotational speed at a value
synchronized with the electrical frequency of the grid
and the generated power is dependent of the existing
unbalance. Therefore, this paper will try to study the
stability of the hydraulic system of the hydropower
system and the achievement of a maximum
efficiency of the water usage in the system.




0 0.5 1 1.5 2 2.5 3
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0 0.5 1 1.5 2 2.5 3
-2
-1
0
1
2
3
4
5
P[%]
a)
b)
t [s]
t [s]
0 0.5 1 1.5 2 2.5 3
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
P[%]
0 0.5 1 1.5 2 2.5 3
-4
-2
0
2
4
6
8
10
12
t [s]
t [s]
a)
b)
n[%]
-48-


3.2. The design of control systems for hydro power
plants with reservoir and surge tank.

Different construction of hydropower systems and
different operating principles of hydraulic turbines
make difficult to develop mathematical models for
dynamic regime, in order to design the automatic
control systems [1,2,5,6]. Also, there are major
differences in the structure of these models.
Moreover, there are major differences due to the
storage capacity of the reservoir and the water supply
system from the reservoir to the turbine (with or
without surge tank). The dynamic model of the
plants with penstock and surge tank is more
complicated than the run-of-the-river plants, since
the water feed system is a distributed parameters
system. This paper presents a solution for the
modelling of the hydraulic systems and the design of
the control system.

As it is know, the hydraulic power available at the
turbine is a function of the water flow trough the
turbine Q
t
and the net head H
n
.

P
h
= Q
t
H
n
(25)

In the long run operation of the hydro power plants
with reservoir and surge tank, it results from the
relation (25) that the power generated by the turbine
is maximum when the head H
n
is maintained at
maximum value, for variations of the river flow due
to meteorological conditions. In this case, the power
generated by the turbine will be only a function of
the water flow trough the turbine. In this case, the
controller is controlling the flow Q
t
in order to
maintain constant the water level in the reservoir,
thus the main feedback will be from the reservoirs
water level sensors.

There are cases when the hydro power plants with
high capacity reservoirs are used as peak load plants.
In these cases, the power dispatcher will impose the
plant generating power and the plant dispatcher or
the automatic control system will distribute the
required load over the operational power units and
the water level in the reservoir can vary between the
admissible maximum and minimum limits with an
admissible rate of variation. The set-point for the
water level controller is generated by the monitoring
system, has a certain variation profile and is a
function of the required power:


t
n
n
Q
P
H

*
*
= (26)
In figure 7 is presented the simplified hydraulic
diagram of a hydropower plant with reservoir and
surge tank. Using this structure, a case study is
presented, regarding the modelling, simulation and
the stability of the of the control structure.

3.2.1 Mathematical modelling of the control system

This control system is opening and closing the
control gates for the turbines in order to assure an
optimal operation and in the same time to maintain a
constant water level in the reservoir. In a similar
mode, a control system may be used for two plants in
a cascade system to maintain the water flows
between the plants.

In order to design the control system, there is
necessary to obtain a mathematical model of the
plant, which contains the water tunnel between the
reservoir and the surge tank, the surge tank itself, the
penstock and the hydraulic turbine[4,5,6]. We will
analyze every component of the plant in order to
determine the model, using the energy balance
equations and the continuity equations for the water
flows, in dynamic regime.

For the hydropower plant we will consider the
simplified diagram in figure 7.

a) The continuity equation for the water in the
reservoir
w t r
B
L
Q Q Q
dt
dH
A = (27)
Where: -A
L
is the area of the horizontal surface of the
equivalent reservoir considered as having a constant
depth H
L
. H
L
.A
L
=V
real
, Q
r
is the water flow entering
the reservoir, Q
W
is the discharging water flow
(through the emptying channel or overspill).

b) The energy balance equation for the water tunnel
between the reservoir and the surge tank:

Surge Tank
Penstok
Generator

Water tunnel
Overspill

Figure 7. Simplified model of the hydropower plant
Turbine
Controller
H
n
*
Level
transducer
Q
t

Echivalent
reservoir

H
B
H
C

Q
cf
Q
B

L
t

Q
e
Q
r

Q
c
Turbine
Gate
H
L

-49-


c
H
B
H
c
Q
c
Q
C
c t
Q
t
Q
t
C
dt
t
dQ
gA
L
t
t
= + + (28)

Where: -L
t
is the length of the tunnel between the
reservoir and the surge tank, A
t
is the cross-section
of the tunnel, Q
t
is the water flow in the tunnel, Q
c
is
the water flow entering the surge tank, H
B
is the
water height in the reservoir, H
C
is the water height
in the surge tank,
dt
dv
g
L
dt
A Q d
g
L
dt
dQ
gA
L
t t t t t t
t
t
= =
) / (
is the
inertial force of the water mass in the tunnel
expressed in hydrostatic units (meters of water
column), C
t
/Q
t
/Q
t
is the hydraulic pressure loss due
to friction with the tunnel walls; the flow Q
t
is
considered positive if the water flows from the
reservoir to the surge tank and the loss coefficient C
t

will be positive in this case, C
t
/Q
c
/Q
c
is the hydraulic
pressure loss due to the cross-section reduction at the
entrance in the surge tank installed to avoid the
pressure shocks, -H
B
-H
C
is the driving force that
moves the water in the tunnel and the surge tank.

c) The continuity equation for the surge tank
cf t c c
Q Q Q
dt
dHc
A = = (29)
Where: A
C
is the cross-section of the surge tank, Q
C

is the water flow at the entrance of the surge tank
(we assume that there is no water flow through the
top of the surge tank), Q
cf
is the water flow through
the penstock.

d) The energy balance equation for the penstock and
the turbine
Considering the low water volume existing in these
two elements we can neglect the water mass
accumulation (water is considered incompressible).
Therefore, from the energy balance equation, the
behaviour of the penstock and the turbine can be
described by the water flow computed from the
following relation:

o
c c c c
o v
H
Q Q C H
Q x Q
+
= , (30)

Where: H
0
is the water level in the surge tank when
the gate is completely open, Q
0
is the water flow for
a completely open gate and -0 x 1 is the opening
factor for the gate.
We choose this form of the expression in order to use
non-dimensional values for the gate opening and to
avoid some very complex equations describing the
water flow through the gate. Therefore are necessary
some experiments to measure the values of H
0
and
Q
0
. The gate is open completely and the flow Q
0
in
steady conditions is measured. After the flow
stabilizes, the water height in the surge tank, H
0
, is
measured. It must be mentioned that in steady state
conditions (H
L
=constant, H
C
=constant,
Q
cf
=constant), the water flow into the surge tank
Q
C
=0.
e) The equation of the controller
The control system structure for the water level in the
reservoir is presented in figure 7. It is recommended
to use a PI controller if the gate servomotor is a
system with its own position controller (proportional
actuator) or a PD controller if the actuator is of the
integrating type (without its own position controller).
In our case we will use a PI controller.
In this control system, the water level in the reservoir
is compared with a set-point determined by the
operational conditions. If the level is different than
the set-point, the error is processed by the turbine
controller which will open or close the gates in order
to maintain the water level at the prescribed value.
The equation of the PI controller can be written in
differential form:
dt
H H d
K H H
T
K
dt
dx ref f
R ref f
i
R
) (
) (

+ = (31)
Where: H
ref
is the set-point level, H
f
is the water level
in the reservoir, K
R
, T
i
are the proportional constant
and the integral constant of the PI controller.
This equation is not considering some factors such
as: delays due to the servomotor and the controller,
and minimum and maximum variation rates for the
gates.


3.2.2 Stability analysis

The equations (27)-(31), which describe the dynamic
behaviour of the control system for the hydropower
unit, are strongly non-linear. Moreover, some
parameters such as loss coefficients C
t
and C
c
can
change due to different flow regimes and due to
building parameters of the feeding system. Therefore
is necessary to study the system stability in the
operation point (small stability in Lyapunov way)
and to determine the stability domain in the
operational parameters plane.

a) Small stability of the control system
Equations (27)-(31) will be linearized around the
operation point. For simplicity, it is preferable to use
non-dimensional variables as follows: q
t
is the
relative variation of the flow through the penstock, h
B

is the relative variation of the water level in the
reservoir, h
c
is the relative variation of the water level
in the surge tank, is the relative variation of the
controller output and q
r
is the relative variation of the
river flow around the steady state value for constant
water level (Q
r0
=Q
t0
).

0
0
0
0
0
0
0
0
0
0
r
r r
r
c
c c
c
B
B B
B
t
t t
t
Q
Q Q
q ;
x
x x
;
H
H H
h
;
H
H H
h ;
Q
Q Q
q

=
(32)

We will introduce the following supplementary
reference parameters,
0
2
0 0
;
c
t
t
ref
i
H
Q C
p
H
T x
= = ,
where characterize the statism of the integral
-50-


controller (the slope of the controller output for a
step variation of the error) and p characterize the
ratio between the losses on the feeding pipes and the
hydraulic pressure at the base of the surge tank.
In steady state regime, H
B
=const., H
c
=const.,
Q
t0
=Q
cf
=const. and from equations (27)-(31) it
results that:
0
2
0
0 0
) 1 (
c
t
t c ref B
H p C H H H Q + = + = =

0
0
0
0
0
c
t
H
H
Q
Q
x = .
According with the previous relations and
considering that the spillover flow is zero (Q
B
=0),
following linearization of equations (27)-(31), the
linear state equations, expressed with non-
dimensional variables defined in (32), can be written
as:
-from (27) it results:

r t
B
t
ref L
q q
dt
dh
Q
H A
+ =
0
(33)
-from (28) it results:
c B
c
ref
c
c
c
c
t
c
t
t
t
c t
t t
h h
H
H
q
H
C
q
H
C
dt
dq
H gA
L Q
Q
Q
= +
+ +
0 0
2
0
0
2
0
0
0
2
2
(34)
It can be considered that the level control system
works with a zero steady state error, H
B0
=H
ref
and
Q
c0
=0 and from equation (28), written for steady
state, it results:


1 1
0
2
0
0
0 0 0
2
0
+ = + =
= =
p
H
C
H
H
or
H H H H C
c
t
t
c
ref
c ref c B
t
t
Q
Q

In this case, equation (34) becomes:

c B t
t
c t
t t
h h p q p
dt
dq
H gA
L Q
+ + = ) 1 ( 2
0
0
(35)
From the linearization of equation (29) it results:

0 0
0
t
cf
t
c
t
c c
Q
Q
q
dt
dh
Q
H A

= (36)
Where the variation of the water flow through the
penstock Q
cf
can be obtain from the linearization of
the equation (30) around the steady state values:

c
t
t
c
c c
cf
h
Q
Q
h
H
H Q x
H
H
Q x Q
+ =
= + =
2
2
0
0
0
0 0 0
0
0
0 0

In this case, equation (29) becomes:
=
c t
c
t
c c
h q
dt
dh
Q
H A
2
1
0
0
(37)
The equation system (33 -37) can be expressed in
canonical matrix-vectorial form:
Bu Ax
dt
dx
+ = or Bu Ax x + =
Where the following notations were used:
0 0
0
0
0
, ,
t
L ref
L
t
c c
c
t s
t t
t
Q
A H
T
Q
A H
T
A H g
Q L
T

=

=
.
And
(
(
(

=
c
B
t
h
h
q
x ;
(
(
(
(

=
1
1
0
L
T
B ;
(
(
(
(
(
(
(

=
c c
f
w w w
T T
T
T T
p
T
p
2
1
0
1
0 0
1
1 1 2
A ; | |
r
q 0 = u .
(39)

The block diagram of the installation is presented in
figure 8.

For the controller, from (7) and noting:
H
H
H H
h
ref
ref ref
r
0
0

= ,
0
0
x T
H K
K
i
ref R
I
= , (integral
constant of the controller
rationalized),
0
0
x
H K
K
ref R
P
= , the proportional
constant (rationalized) and e(t)=(h
r
-h
BT
), we obtain:

dt
t de
K t e K
dt
d
P I
) (
) ( + =

(40)
Adding equations (33), (35), and (37) and using the
notations from (14), we obtain the complete equation
system for the controller in figure 7:
H
R
(s) H
IT
(s)
H
p
(s)
h
B
*
h
B
(s) (s) e(s)

q
r
(s)
q
t
(s)
-
+
+
-
Figure 8. The block diagram of level controller
Notes: H
R
(s)-controller transfer function;
H
IT
(s)-the transfer function on the direct channel
H
P
(s)-transfer function on the perturbation channel
-51-


r
L
t
L
B
q
T
q
T dt
dh
+ =
1 1
(41)


c
t
B
t
t
t
t
h
T
h
T
p
q
T
p
dt
dq

|
|
.
|

\
| +
+ =
1 1 2
(42)
=
c
c
c
t
c
c
T
h
T
q
T dt
dh 1
2
1 1
(43)
Using Laplace transform, for zero initial conditions,
from equations (40) to (43) it results:

) s ( h
sT
) s ( h
sT
p
) s ( q
sT
p
) s ( q
); s ( q
sT
) s ( q
sT
) s ( h
c
t
B
t
t
t
t
r
L
t
L
B

|
|
.
|

\
| +
+ =
+ =
1 1 2
1 1
(44)
) s ( e K ) s ( e
s
K
) s (
) s (
sT
) s ( h
sT
) s ( q
sT
) s ( h
P
I
c
c
c
t
c
c
+ =
=
1
2
1 1
(46)
These equations can be processed to obtain the
transfer functions on the direct channel and on the
perturbation channel. From (45), the flow q
t
can be
expressed as:
) (
2
1
) (
2
) 1 (
) ( s h
p s T
s h
p s T
p
s q
c
t
B
t
t
+

+
+
= (48)
This will be used in equation (46):
) s ( ) s ( h ) s ( h
p s T
) s ( h
p s T
) p (
) s ( h sT
c c
t
B
t
c c

+

+
+
=
2
1
2
1
2
1
(49)

And the variation of the water level in the surge tank
can be determined as:

) s (
) p ( s ) T pT ( s T T
p s T
) s ( h
) p ( s ) T pT ( s T T
) p (
) s ( h
t c t c
t
B
t c t c
c

+ + + +
+

+ + + +
+
=
1 2 4 2
2
1 2 4 2
1
2
2
(50)

Let note H
n
(s)=2T
c
T
t
s+(4pT
c
+T
t
)s+2(p+1) to
simplify the calculations and let use (50) in (48) to
determine the water flow through the penstock,
which is equal with the flow through turbine.

1 2 4 2 : where
1
2
1
2
+ + + + =
+
+
+
=
p s ) T pT ( s T T ) s ( H
), s (
) s ( H
) s ( h
) s ( H
) s ( H
p s T
) p (
) s ( q
t c t c m
n
B
n
m
t
t
(51)
And after simple processing from (51) and (44) it
results:
( )
) (
) (
) ( 2
) (
) (
2
) ( s q
s P
s H p s T
s
s P
p s T
s h
r
n t t
B

+
+
+
=
(52)
Where the characteristic polynomial of the open-loop
system is:
( )
| |
| | ) p p ( s T ) p ( ) T T )( p ( p
s T T ) p ( T T ) p ( T T p
s T pT T T s T T ) s ( P
t c L
t c t L L c
t c t L t c L T
1 3 2 1 1 4
1 2 1 2 2 8
8 2
2
2 2
3 4 2
+ + + + + + + +
+ + + + +
+ + + =

(53)

From (52), the two transfer functions in the block
diagram of the control system presented in figure 2
can be calculated:
) (
2
) (
s P
p s T
s H
t
IT
+
= and
( )
) (
) ( 2
) (
s P
s H p s T
s H
n t
P
+
=
(54)
From (47), using the Laplace transform, we get:

) ( ) ( ) ( s h
s
s K K
s h
s
s K K
s
BT
R I
r
R I

+

+
= (55)


Replacing (55) in (52), after several simple
transforms, it results:

) s ( q
) s K K )( p s T ( ) s ( P s
) s ( H ) p s T ( s
) s ( h
) s K K )( p s T ( ) s ( P s
) s K K )( p s T (
) s ( h
r
R I t
n t
r
R I t
R I t
B

+ + +
+
+
+
+ + +
+ +
=
2
2
2
2

(56)
Where the characteristic polynomial of the closed-
loop system is:
( )
| |
| |
I R I t
R t t c L
t c t L L c
t c t L
t c L R I t
pK s ) p p pK K T (
s K T T ) p ( ) T T )( p ( p
s T T ) p ( T T ) p ( T T p
s T pT T T
s T T ) s K K )( p s T ( ) s ( P s ) s ( L
T
2 1 3 2 2
1 1 4
1 2 1 2 2 8
8
2 2
2
2
3 2
4
5 2
+ + + + + +
+ + + + + + +
+ + + + + +
+ + +
+ = + + + =


For an exact analysis of the hydropower unit
dynamic behaviour it can be considered the example
of a hydropower plant with reservoir and surge tank
having a low installed power but a high storage
capacity in the reservoir.

Experiment. It is considered a hydropower system
with the following parameters:
The water volume in the reservoir V
L
=4,8.10
6
m
3
;
-The equivalent depth of the reservoir (considered
constant) H
L
=60m; The equivalent reservoir surface
area is A
L
4,8.10
6
/60=8.10
4
m
2
; The length of the pipe
between the reservoir and the surge tank L
t
=9650m,
the diameter of the pipe D
t
=3,6m and the cross-
section A
t
=10m
2
; The surge tank has a diameter
D
c
=5,4m with a cross-section A
c
=23m
2
; The gross
nominal head H
B
=260m, H
Bmax
=266m, H
Bmin
=245m;
The gross nominal head at the surge tank H
c0
=230m;
The nominal flow through the penstock Q
t0
=36m
3
/s;
The length of the penstock L
cf
=205m.
-52-


Using the previous data, the time constants and
equations coefficients can be determined using
relation (14):

s .
.
T
;
.
T ;
,
T
L
c t
5
10 6
36
80000 260
150
36
23 234
15
10 230 81 9
36 9560
= =
= = =

=

023 0
36 36
230 13 0
13 0
230
230 260
2
0
0
0
0
0
2
,
.
. , pH
C
; ,
H
H H
H
C
p
Q
Q
t
c
t
c
c ref
c
t
t
= =
=

= =
(56)
34 2
230
36
0
0
0
0 0
0
0
0 0 0 0
,
c
H
t
Q
H
Q x
;
H
c
H
Q x
t
Q
v
Q = = = = =

The steady state value for the controller output x
0
is
determined using the maximum flow through the
penstock Q
vmax
=56m
3
/s and is obtained for x
0
=1:

7 3
64 0
374 2
iar 64 0
56
36
0
0
0
,
,
,
; , x
H
Q
= = = =
Replacing these values in equations (48) and (50),
we obtain the operational expressions for the flow q
t

and the level in the surge tank:
) s ( h
s
,
) s ( h
s
.
) s ( q
) s ( q
s
. ,
) s ( q
s
. ,
) s ( h
c B t
r t B
1 58
92 1
1 58
35 4
10 166 0 10 166 0
5 5
+

+
=
+ =

(57)

) s (
s s
) s ( ,
) s ( h
s s
,
) s ( h
B c

+ +
+

+ +
=
1 41 1991
1 58 26 0
1 41 1991
5 0
2
2
(58)
( )
| |
| |
I R I t
R t t c L
t c t L L c
t c t L t c L
R I t
pK s ) p p pK K T (
s K T T ) p ( ) T T )( p ( p
s T T ) p ( T T ) p ( T T p
s T pT T T s T T
) s K K )( p s T ( ) s ( P s ) s ( L
T
2 1 3 2 2
1 1 4
1 2 1 2 2 8
8 2
2
2
2
3 2
4 5 2
+ + + + + +
+ + + + + + +
+ + + + + +
+ + + =
= + + + =


L(s) = 405.10
8
. s
5
+ 153.10
7
.s
4
+ 348,53.10
5
.s
3
+
(35,27.10
4
+15.K
R
).s
2
+ (15.K
I
+0,26.K
R
+4238).s +
0,26.K
I
(59)

Using the stability criteria Routh-Hurwitz for the
characteristic polynomial (59) we obtain a series of
inequalities for the tuning parameters K
i
and K
R
as
follows:
-from the block diagram of the control system
presented in figure 8, it can be seen that the transfer
coefficient on the direct channel is negative,
therefore the controller shall have inverse output and
therefore K
R
<0;
-from the inequality a
0
=0,26.K
I
>0 it results that the
integral parameter shall be positive;
-from the Hurwitz determinant of second order
results:
a
0
. a
1
= 0,26 K
I
.( 15 K
I
+ 0,26 K
R
+ 4238) > 0

16300 69 , 57 + <
I R
K K
this gives us a very large range for the tuning
parameters of the controller;

-from the Hurwitz determinant of third order results:
K
I
< (14950000000 + 152270* K
R
+3,9*
K
R
2
)/(3072000 - 225* K
R
)
Solving previous inequality graphically, we obtain
K
R
(0 -100), 0<K
i
<(4866 -4826) and again, the
range for the tuning parameters is large.

During the real operation of the hydropower unit,
even if the control system remains stable, the
oscillations in the hydraulic unit shall be avoided
since they can generate huge over pressure in the
penstock. In order to highlight these oscillations, a
simulation of the hydropower unit was performed
using different values for the tuning parameters.
The block diagram in figure 9 was used for
simulation. The transfer functions in the diagram
were computed using the equations (57)-(58).

Notes: |
.
|

\
|
+ =
s
K
K s H
I
R R
1 ) ( ,
1 58
92 . 1
+
=
s
H
Qt

s
s
s H
qth
5 10 . 166 . 0
) (

= ,
1 58
35 . 4
+
=
s
H
hqt
,
1 41 1991
5 . 0
) (
2
+ +
=
s s
s H
hhc
,
1 41 1991
1 58
) (
2
+ +
+
=
s s
s
s H
C

s
5
10 . 166 , 0

h
B
(s)
q
r
(s)
q
t
(s)
-
+
Figure 9. The block diagram of the level control system, used for simulation
H
R
(s)
h
B
*
(s)
e(s)

+
-
H
C
(s)
H
hhc
(s)
H
hqt
(s)
H
qth
(s)
H
Qt
(s)
h
c
(s)
-
-
+ +
-53-




































This diagram was preferred for simulation since it
allows highlighting the variation of some specific
parameters of the system, such as the level in the
surge tank and the flow variation in the penstock Q
t
.
The simulation results are represented in the figures
10 and 11.

In figure 10 is presented the system output for a 10%
variation of the set-point for the reservoir water
level, with no limitation on the flow control channel.
Oscillations of the flow can be observed, due to the
big differences between the time constants in the
simulation scheme.

To avoid these flow oscillations, the variation of the
flow control output is limited in order to avoid big
differences between the subsequent commands sent
to the gate. The simulation results are presented in
figure 11.

3.3. Conclusions

Specifically, a hydropower plant may be composed
of several turbine-generator units, referred to in this
paper as groups. Each hydropower plant must be
controlled by a local SCADA system that will be
connected at the Central SCADA system, which
performs the control of the plants from the hydro-
valley, for optimal generation scheduling and control
for the reservoir stream-flows.


































4. MONITORING SYSTEM FOR CASCADE
HYDRO POWER PLANTS.

The optimal generation scheduling is an important
daily activity for hydroelectric power generation
companies. The goals are to determine which
hydropower groups are to be used in order to
generate enough power to satisfy demand
requirements, with various technological constraints
and with minimum operating cost. In the same time
hydropower systems must consider the stream-flow
equations for reservoirs that couple all reservoirs
along a hydro-valley, because the amount of outflow
water released by one of hydropower plant affects
water volumes in all the plants downstream.
Furthermore, other conditions must be imposed as:
water travel times, alternative uses (irrigation, flood
control, navigation). To solve the optimal scheduling
of hydropower plants, a highly sophisticated
modelling for the operation is required.

In the figure 12 is presented the SCADA system
architecture for monitoring and control of hydro
power plants in a cascade system along a river,
implemented in order to optimize the use of the river
resources [7]. This structure is implemented at the
central dispatcher. At the level of hydro power plant
there is a local monitoring and control system,
interconnected with central dispatcher via modems
and radio communication buses.
0 2000 4000 6000 8000 10000
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
0 2 0 0 0 4 0 0 0 6 0 0 0 8 0 0 0 1 0 0 0 0
-0 .0 2
0
0 .0 2
0 .0 4
0 .0 6
0 .0 8
0 .1
0 .1 2
0 .1 4
t [s]
h
B
[%]
Figure 10. The response of the control system for a 10% variation of the set-point
t [s]
q
t
[%]
0 2000 4000 6000 8000 10000
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
Figure 11. The response of the control system with flow limitation
0 2000 4000 6000 8000 10000
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
t t
h
B
[%] q
t
-54-


The monitoring system presented above is
performing the following functions:
-Data acquisition from the local monitoring systems
installed in each power plant, using routers
connected to telephone lines or radio
communications; the data is sent tot the SCADA
server for processing, monitoring and optimization;
-The optimal load distribution between the
operational units in the power plants, depending of
the available water flow, in order to ensure an
efficient usage of the water for the required system
power;
-Monitoring of the water levels in the reservoirs and
water flow control in order to maintain the maximum
water level in the reservoirs;
-In the case of extreme rain conditions, control the
water flows in order to avoid flooding of the areas
neighbouring the plants.


5. CONCLUSIONS

The possibility of implementation of digital systems
for monitoring and control for power, frequency and
voltage in the cascade hydro power plant was
discussed. The simplified mathematical models,
capable to accurately describe dynamic and
stationary behaviour of the hydro units were
developed and simulated and these results were
compared with the experimental results. Finally, a
practical example was used to illustrate the design of
controller and to study the system stability.
REFERENCES

[1] G., D Ferrari-Trecate, Mignone, D. Castagnoli, M.
Morari, Hybrid Modeling and Control of a
Hydroelectric Power Plant, CH 0802, Institut fur
Automatik, ETH- Zurich, (2003).
[2] C. Henderson, Yue Yang Power Station The
Implementation of the Distributed Control System,
GEC Alsthom Technical Review, Nr. 10, 1992.
[3] G. I. Krivchenko, Hydraulic machines: Turbine and
Pumps,2nd ed. ISBN 1-56670-001-9, CRC Press,
London (1994).
[4] M. Vntoru, E. Iancu, C. Vntoru, Control,
Monitoring and Protection of the Turbine and
Generator System, International Symposium on System
Theory, Robotics, Computers and Process Informatics,
SINTES 9, Craiova , (1998).
[5] O. F. Jimenez, M. H. Chaudhry Water level Control in
Hydropower Plants, Journal of Energy Engineering,
Vol118, No. 3 Dec. 1992.
[6] H. Weber, V. Fustik, F. Prillwitz, A. Iliev, Practically
oriented simulation model for the Hydro Power Plant
Vrutok in Macedonia, Balkan Power Conference, 19.
21.06.2002, Belgrade
[7] I.C.E Felix -Bucharest S.A, Arhitectura sistemului de
control supervizor i achiziie de date (SCADA) al
cascadei de hidrocentrale de pe Oltul Mijlociu, http://
www.felix.ro/apl.html.

[8] M. Vinatoru, Conducerea automata a proceselor
industriale, vol. II, Editura Universitaria Craiova, 2005.

To CHE
System
ER
Radio
Engineering
Pules Message NMEA
Sincron

mGPS
Local Sincron
Server

Ethernet Switch
Pulse Message NMEA

Operator 1
Operator 2
Engineering Server
SCADA 1
Server
SCADA 2
Hist. Date
Server
Ethernet Bus
Ethernet Switch
Simulator CHE
Router
Asyncron 2
Router
Asyncron 1
- - - - - - Modem 10 Modem 1
Ethernet Bus
Sever Radio
comunication
Fig. 12. SCADA system architecture - Dispatcher
Hydro power plants
Wall Display (4 modules)
-55-