ELG%204152%20MODERN%20CONTROL%20SYSTEM%20PROJECT

1
ELG 4152 MODERN CONTROL SYSTEM PROJECT
Predictive Feed forward Control for a
Hydroelectric Plant
by Amirul Bhuiya Norman Escobar Omar
Faroque Hung Pham
for Professor Riadh Habash TA Fouad F. Khalil
March 28, 2007
2
Introduction
The purpose for this project is to replicate the
work done by Dewi Jones and Saad Mansoor in
their paper Predictive Feedforward Control for
a Hydroelectric Plant, IEEE TRANSACTIONS ON
CONTROL SYSTEMS TECHNOLOGY, VOL. 12, NO. 6,
NOVEMBER 2004 Their work consisted of simulating
the Dinorwig hydroelectric plant in the U.K. and
improving its performance by designing a feed
forward loop to better respond to frequency
variations on the electrical grid. Their design
allowed them to improve the system response time
by up to 3 seconds depending on the operating
point of the hydroelectric plant and the load on
the system.
3
Introduction
The following references were used in the
development of our project D.I. Jones, Dynamic
system parameters for the National Grid, IEE
Proc.-Gener. Transm. Distrib., Vol. 152, No. 1,
January 2005 In this paper Dr. Jones developed a
generic model of the National grid Dewi Jones,
Estimation of Power System Parameters, IEEE
TRANSACTIONS ON POWER SYSTEMS, VOL. 19, NO. 4,
NOVEMBER 2004. In this paper Dr. Jones developed
the parameters for estimating the low order
dynamic model of the U.K. power system. German
Ardul Munoz-Hernandez,Dewi Jones, MIMO
Generalized Predictive Control for a
Hydroelectric Power Station, IEEE TRANSACTIONS
ON ENERGY CONVERSION, VOL. 21, NO. 4, DECEMBER
2006. In this paper Dr. Ardul applies generalized
predictive control to a multivariable model of
the Dinorwig power plant. Dewi Jones and Saad
Mansoor, Predictive Feedforward Control for a
Hydroelectric Plant, IEEE TRANSACTIONS ON
CONTROL SYSTEMS TECHNOLOGY, VOL. 12, NO. 6,
NOVEMBER 2004 In this paper the authors use a
feed-forward model to improve the Dinorwig power
plant. Sa'ad Mansoor , Thesis "Behaviour and
Operation of Pumped Storage Hydro Plants
University of Wales, Bangor, School of
Informatics, July 2000. In this thesis Dr.
Mansoor developed a model of the Dinorwing power
plant.
4
METHODOLOGY APPLIED
PROJECT SELECTION FROM IEEE PAPER
PAPER RESEARCH
EQUATION/MATERIAL ANALYSIS
DEVELOPMENT OF BASIC MODEL
DEVELOPMENT OF GUI AND TEST SCRIPTS
IMPLEMENTATION AND TESTING
ANALYSIS OF RESULTS
5
Work Distribution
Amirul Bhuiya Research power plant modeling,
select most appropriate model to use and develop
proper plant parameters for model
implementation Norman Escobar Implement the
system model, the GUI interface and perform
related tests using MATLAB and its Simulink and
GUIDE toolboxes Omar Faroque Research a viable
grid model, develop and simplify for
implementation Hung Pham Research feed forward
systems, develop a feed forward loop to be
incorporated into the system model. Each member
wrote the corresponding sections in the report
and in the development of the presentation.
6
Power Plant
General Background
DAM
Water
Generator
Transformer
Turbine sinning by water flow
7
General Background
Water Outlet
Water Entrance From Tunnel
8
General Background
Dinorwig Electric Generator
9
General Background
Primary Side
HV/
LV/Secondary side
Breather
10MVA Oil Filled Power Transformer
Oil Tank
10
Power Grid
General Background

• The power grid consists of many interconnected
  power generating stations
• Grid circuit load is shared by interconnected
  power station.
• Interconnected power generating stations have to
  maintain identical Voltage, Phase difference, and
  Frequency with each other.

11
General Background
12
What causes grid frequency variations?
General Background

1. Load variation on the GRID overloads the
   generators causing their rotation (RPM) to slow
   down.
2. Lightning creates tremendously high voltages on
   the Transmission lines with unknown frequencies
   producing resonant frequencies on the HT line..

Why the need to maintain a constant grid
frequency?

1. Frequency variations can damage or disrupt
   sensitive electronic equipment.
2. Motors, variable speed drives, switching power
   supplies etc rely on constant frequency.

13
Frequency deviation to power demand relationship
Freq. Deviation
Power Demand
14
MODEL DEVELOPMENT

• Model of Hydroelectric Plant and Electric Grid

15
MODEL DEVELOPMENT

• PI Controller

Guide Vane Servo model
16
MODEL DEVELOPMENT
Turbine/Generator

• Hydrodynamic model

17
MODEL DEVELOPMENT
Grid
Feedback
18
MODEL DEVELOPMENT
Feed forward model
19
FEED FORWARD MODEL
Required Feed Forward Transfer Function Inverse
of Gp (According a paper propose by G. C.
Goodwin)
20
RESULTS
GUI developed to perform tests on system model
21
RESULTS

• The GUI was designed to control the
  turbine/generator parameters
• It provides control of the amount of generators
  running at any given time
• It allows the addition of disturbance that models
  the grid power demand change
• It allows control of feed forward and feedback
  loops to test performance

22
RESULTS

• Simulation of system with one generator running,
  standard turbine/generator parameters and no grid
  disturbance with feedback only

23
RESULTS

• Simulation of system with all generators running,
  standard turbine/generator parameters, 15MW grid
  disturbance, feedback and feed forward enabled

24
RESULTS
of Generators Response time (s) Response time (s) Response time (s)
With Feedback only With Feed Forward With Disturbance
1 2.24 2.24 2.27
2 3.36 3.36 3.3
3 4.43 4.43 4.12
4 5.44 5.44 5.42
5 6.36 6.36 5.5
6 7.22 7.22 5.9
25
RESULTS
Response time (s) Response time (s) Response time (s)
With Feedback only With Feed Forward With Disturbance
Normal Operation 3.15 3.15 0
High Load 9.85 9.85 0
26
RESULTS
Normal Operation with grid incorporated
27
RESULTS
High Demand with grid incorporated
28
Conclusion

• The objective of this project was to simulate the
  hydroelectric plant model by Dewi Jones and Saad
  Mansoor which incorporated a feed forward path
  in an existing feedback control system. Here, we
  have successfully simulated the idea behind the
  feed forward control system of 6 hydro power
  generator.
• As the results indicate, with feedback alone, the
  system is stable and works within requirements
  and specifications. However, the electrical
  power response to demand variations is very
  important with respect to delivery time and
  frequency variations (every second counts). This
  is where the feed forward theory comes into the
  design. For this particular control system,
  feedback improved the steady state error, the
  settling time, the systems overall stability,
  and the systems overshoot but not the systems
  delivery time (when the power is actually being
  deliver to the consumer).
• The clear advantages of feed forward theory for
  this particular system are
• Improve delivery time
• 2. Cheap to implement into existing system
  (avoiding a new design from scratch)
• The disadvantage of this particular feed forward
  model is that every new Grid (city) needs a new
  feed forward path because different cities have
  different load or demand characteristics.

The main problem that we encounter during the
investigating of this project is the simulation
of the more practical feed forward design propose
by the paper. This is due the fact that we do
not have the exact parameter values of the
Turbine generator and the exact transfer function
for the Grid (city) that the paper has access to.
However, by simulating our developed feed
forward design it is easy to demonstrate that
feed forward theory is something that should be
consider in all control system designs.