International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 1

ISSN 2229-5518

Fuzzy Logic Based Hydro-Electric Power

Dam Control System

M. Abbas, M. Saleem Khan, Nasir Ali

Abstract — This research paper presents the construction design of Hydro-Electric Power Dam Control System using Fuzzy Logic. In this design two input parameters: water level and flow rate and two output parameters: release valve control and drain valve control are used. This proposed system uses a simplified algorithmic design approach with wide range of input and output membership functions. The hardware of control system for fuzzifiers and defuzzifiers is designed according to the need of system. The proposed simplified algorithmic design is verified using MATLAB simulation and results are found in agreement to the calculated values according to the Mamdani Model of the Fuzzy Logic Control System.

Index Terms—Fuzzy Logic Control, Hydro-Electric Power Plant, Inference Engine, Rule Selection.

——————————·——————————

1 INTRODUCTION

HE modern-day technologies in the areas of information storage and retrieval, web search, image processing, control, pattern recognition, bio- information and computational biology, e-markets, autonomous navigation, and guidance are benefited using fuzzy sets. An integrated framework sustaining a variety of facets of human-centric computing is developed by means of fuzzy sets. The current trends of information technology have proved that the increasing level of intelligence, autonomy and required flexibility comes true with the increased human centricity of resulting results. The holistic view covers concepts, design methodologies, and algorithms with interpretation, analysis, and engineering knowledge. The computing systems are based on predefined models of two-valued logic and human information processing, concerned with two distinct words. In order to communicate between these two words we need to develop an interface. This is the key motivation behind the emergence of human-centric systems and human-
centric computing [1].
The construction of a dam is necessary for the electric power generation, flood control, irrigation system, metropolitan and industrial water supply. Different kind of methods have been introduced and implemented to control the hydro- electric power dam due to non-deterministic behavior of water parameters such as flow rate and release etc. [2].
Fuzzy Set Theory along with its membership functions was implemented to the Fairbairn reservoir in Emerald,

————————————————

M. Abbas is working as Visiting Faculty in the Department of Electrical Engineering in National University of FAST Lahore and also Lecturer Internee in Physics, Government Islamia College, Lahore, Pakistan (e-mail: shairabbas@yahoo.com)

Dr. M. Saleem Khan is with the GC University Lahore Pakistan, working

as Director Computer Science Department

(e-mail: mskgcu@yahoo.com, s.khan@ed.ac.uk)

Nasir Ali is an Assistant Professor of Physics in Govt. DS College,

Lahore, Pakistan (e-mail nasirali8787@yahoo.com)

Central Queensland, Australia where fuzzy rules were generated with explicit recognition of storage volume non-specificity in the discrete Stochastic Dynamic Programming (SPD) [3].
Fuzzy dynamic programming model was used for Hirakud dam in the State of Orissa in India in which irrigation; hydropower generation and flood control were considered as fuzzy variables [4].
The neural network and fuzzy systems were also adopted for dam control in which a comparison was made between reservoir operations using the fuzzy and neural network systems and actual one by operator, using examples of floods during flood and non-flood seasons [5].
Reports show that hydroelectric dams produce 20 percent of the world’s total production of electrical energy. The development of a hydro-electric power dam control system based on fuzzy logic with two inputs and two outputs. Using water level and flow rate measuring devices for feedback control, and two control elements for draining and valve controlling (release), and formulated fuzzy rules for water level and flow rate has been achieved.
To control the water release, the controller reads the water level and flow rate after every sampling period. This proposed design work of Hydro-Electric Power Dam System is the application of fuzzy logic control system consisting of two input variables: water level and flow rate, and two output variables: Drain valve and (Releasing) Valve control used in a reservoir plant of Hydro-Electric Power Dam to monitor the system of Dam.
The basic structure of the proposed model is described in
Section 2. Section 3 gives the simplified design algorithm

IJSER © 2011

http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 2

ISSN 2229-5518

of fuzzy logic for Hydro-Electric Power Dam System. Section 4 describes the simulation results of this system. Conclusion and future work is given in Section 5.

2. BASIC STRUCTURE OF THE PROPOSED HYDRO-ELECTRIC POWER DAM

Fig.1 Arrangement of proposed hydro–electric power system

The main parts of the proposed hydro-electric power plant are shown in Fig. 1. Upper lake where water is stored presents the water level. The greater the vertical distance b/w the upper and lower lakes, the more is the generation of electricity. In order to release or block water, a control valve is used according to the need. Water on releasing from the dam gets to the blades of the turbine all the way through the penstock. Its slope and thickness determines the efficiency of the dam. Turbine produces electrical energy and water released from the turbine is released to lower lake where the drainage system is brought into action according to the requirements [6].
The schematic diagram of the proposed hydro-electric power plant is shown in Fig. 2. Water level and flow rate devices are used to monitor the status of water in the plant which is connected with the two fuzzifiers of the fuzzy logic control system after suitable amplification and voltage adjustment unit. Two outputs of defuzzifiers are the releasing control valve and drainage valve.

Fig. 2 Block Diagram of Hydro-Electric Power Dam fuzzy control system

3. DESIGN ALGORITHM

The algorithm designed for this system consists of two fuzzy input variables. Five triangular membership functions are equally determined over a scale range of 0 m to 20 m for the water level and 0(m3s-1) to100000(m3s-1) for flow rate inputs. The five fuzzy membership functions for water level input are termed as: very low 0-5 m, low
0-10 m, below danger5-15 m, danger10-20 m and above danger15-20 m. The five fuzzy membership functions for flow rate input are: very slow 0m3s-1 -25000m3s-1, slow
0m3s-1-50000m3s-1, normal 25000m3s-1-75000m3s-1, fast
50000m3s-1-100000m3s-1, and very fast 75000m3s-1-
100000m3s-1.Two outputs of this proposed system are:
(release) control valve and drainage valve. The control
valves for release and drainage output variables consist of
five membership functions: fully closed 0-5, 25% Opened
0-50, 50% Opened 40-60, 75% Opened 50-90 and Fully
Opened 70-100.

3.1. Fuzzifier

The input crisp values are compared by the fuzzifier with certain levels and generate linguistic values of each input variable for inference engine. The inference engine simulates human decision with fuzzy concepts, implication and rules of inference in fuzzy logic [7]. The occupied region description, membership functions and range for two input variables are given in Table 1 and Table 2.

TABLE 1

MEMBERSHIP FUNCTIONS AND RANGES OF INPUT VARIABLE WATER LEVEL (m)

Membership

Function (MF)

Ranges

Region Oc-

cupied

Very Low

0-5

1

Low

0-10

1-2

Below Danger

5-15

2-3

Danger

10-20

3-4

Above Danger

15-20

4

TABLE 2

MEMBERSHIP FUNCTIONS AND RANGES OF INPUT VARIABLE FLOW RATE (m3/s)

For each input variable, five membership functions are used as shown in Fig. 2 and in Fig. 3.

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 3

ISSN 2229-5518

Fig.3 Plot of membership functions for input variable, “WATER LEVEL”

The five membership functions, “very low”, “low”, “below danger”, “danger” and “above danger” are used to show the various ranges of input fuzzy variable “WATER LEVEL” in a plot consisting of four regions as shown in Fig. 2.
The five membership functions, “very slow”, “slow”, “normal”, “fast” and “very fast” are used to show the various ranges of input fuzzy variable “FLOW RATE” in a plot also consisting of four regions as shown in Fig. 3.

Fig. 4 Plot of membership functions for input variable, “FLOW RATE”

The linguistic values are the mapping values of the fuzzy input variables with the membership functions occupied in the regions. As we are using two variables, therefore four linguistic values are shown in Fig.4. The mapping of input fuzzy variables with the functions in four regions is listed in Table 3.

TABLE 4

RULE MAPPING FOR REGIONS OCCUPIED

TABLE 3

LINGUISTIC VALUES OF FUZZIFIERS OUTPUTS IN ALL REGIONS

Input

Variables

Linguistic Fuzzifiers Outputs

Region

1

Region

2

Region

3

Region

4

Water

Level

f1

f1[1]

f1[2]

f1[3]

f1[4]

Water

Level

f2

f1[2]

f1[3]

f1[4]

f1[5]

Flow

Rate

f3

f2[1]

f2[2]

f2[3]

f2[4]

Flow

Rate

f4

f2[2]

f2[3]

f2[4]

f2[5]

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 4

ISSN 2229-5518


Fuzzification process for two input variables need two separate fuzzifiers. Each fuzzifier consists of: input voltage to crisp value converter, operational region for a crisp value detector, fuzzy set membership value mapping and selection arrangements [8]. The design of such a fuzzifier is shown in Fig. 5.

Fig. 5 Fuzzifier Block

Table 5 gives the working results of two fuzzifiers using the given values of input variables. These results are achieved using the fuzzifier design for water level and flow rate inputs shown in Fig. 6 (a) and Fig. 6 (b) [9].

Fig. 6 (a) Design of fuzzifier for Water Level Input

Fig. 6 (b) Design of fuzzifier for Flow Rate Input

TABLE 5

RESULTS OF FUZZIFICATION

Input

Variables

Values

Region Selec- tion

Fuzzy Set Calcu- lation

Water

Level

x=13

10x<15

Region-3

f1=(15-13)/5=0.4 f2=1-f1=1-0.4=0.6

Flow

Rate

x=95000

75000x<100000

Region-4

f3=(100000-

95000)/25000=0.2

f4=1-f3=1-0.2=0.8

3.2. Inference Engine

The inference engine contains four AND operators that select minimum value input for the output. This inference engine accepts four inputs from fuzzifier and applies the min-max composition to obtain the output R values. The min-max inference method uses min-AND operation

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 5

ISSN 2229-5518

between the four inputs. Fig. 7 shows this type of inference process.
variables in a region corresponds to mapping of two functions. The corresponding mapping values of f1 [3], f1
Number of active rules = mn
where m = maximum
[4], f2[2], f2[3] were used to establish the 4 rules. Here f1 [3]
means the corresponding mapping value of membership
number of overlapped fuzzy sets and n= number of in-
puts. For this design, m = 5 and n = 2, so the total number
of active rules are 25.The total number of rules is equal to the product of number of functions accompanied by the input variables in their working range [10]. The two input variables described here consisted of five membership
functions. Thus, 5 x 5 = 25 rules were required which are shown in Table 6.
function “Below Danger” of water level in region-3 and the similar definitions are for the others.
R1 = f1 ^ f3 = f1[3] ^ f2[4] = 0.4 ^ 0.2 = 0.2
R2 = f1 ^ f4 =f1[3] ^ f2[5] = 0.4 ^ 0.8 = 0.4
R3 = f2 ^ f3 =f1[4] ^ f2[4] = 0.6 ^ 0.2 = 0.2
R4 = f2 ^ f4 =f1[4] ^ f2[5] = 0.6 ^ 0.8 = 0.6

TABLE 6

TOTAL NUMBER OF RULES

Fig. 7 Block Diagram of Inference Engine

3.3. Rule Selector

The rule selector receives two crisp values of water level and flow rate. It gives singleton values of output functions under algorithm rules applied on design model. For two variables, four rules are needed to find the corresponding singleton values S1, S2, S3 and S4 for each variable according to these rules are listed in Table 7.

TABLE 7

ILLUSTRATION OF RULES APPLIED MODEL

Rule

No.

INPUTS

SINGLETON VALUES OF OUTPUTS

Sin- gleton Val- ues

Rule

No.

Water

Level

Flow

Rate

Release Control

Valve

Drainage

Valve

Sin- gleton Val- ues

1

Below

Danger

Fast

0.50

=50% opened

50% opened

=0.50

S1

2

Below

Danger

Very

Fast

0.50

=50% opened

75% opened

=0.75

S2

3

Danger

Fast

0.75

=75% opened

Fully

Opened

=1.0

S3

4

Danger

Very

Fast

0.75

=75% opened

Fully

Opened

=1.0

S4

In this case only 4 rules are required for the particular values of two variables because each value of two

The rule base accepts two crisp input values, distributes the universe of discourse into regions with each region containing two fuzzy variables, fires the rules, and gives the output singleton values corresponding to each output variable. Fig. 8 shows the main block diagram of the Rule Base.

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 6

ISSN 2229-5518

3.4. Deffuzifier

Fig. 8 Rule Base

The defuzzification process provides the crisp value outputs after estimating its inputs [11]. In this system 8 inputs are given to each of the two defuzzifiers. Four values of R1, R2, R3, R4 from the outputs of inference engine and four values S1, S2, S3, and S4 from the rule selector are shown in Fig. 10. Each defuzzifier estimates the crisp value output according to the center of average
(C.O.A) method using the mathematical expression, l:Si * Ri/ l:Ri, where i = 1 to 4. Each output variable membership
function plot consists of five functions with the same
range values for simplification.

In this system, two defuzzifiers control the actuators; Release (Valve Control and Drainage Valve. The member- ship functions of the two output variables are shown in Fig. 7 to Fig.9, and the detail of each plot is given in Table 8.

TABLE 8

OUTPUT VARIABLES MEMBERSHIP FUNCTIONS

MFs

Range

Release(Valve)

Drain Valve

MF1

0-5

Fully Closed

Fully Closed

MF2

0-50

25% Opened

25% Opened

MF3

40-60

50% Opened

50% Opened

MF4

50-90

75% Opened

75% Opened

MF5

70-100

Fully Opened

Fully Opened

Fig. 9 Plot of Membership Functions for Output Variable, “Drain Valve”

Fig. 10 Plot of Membership Functions for Output Variable, “Release (Valve Control)”

Fig. 11 Defuzzifier Block

Fig. 11 shows the design arrangement of a defuzzifier.
One defuzzifier consists of: one adder for l:Ri, four multipliers for the product of Si*Ri, one adder for l:Si * Ri, and one divider for l:Si * Ri/ l:Ri. Finally a defuzzifier
gives the estimated crisp value output.

4. RESULTS AND DISCUSSION

The designed values for two outputs; Release (Valve Control) and Drain Valve are given in the Table 9 and Table 10. According to the results of inference engine
l:Ri = R1+R2+R3+R4 = 0.2+0.4 +0.2 +0.6 = 1.4

TABLE 9

DESIGNED VALUE FOR CONTROL VALVE (RELEASE)

l:Si * Ri= 0.90

l:Si * Ri/ l:Ri= 0.90/1.4 = 0.6428= 64.28%

of the Valve will be Opened.

TABLE 10

DESIGNED VALUE FOR DRAIN VALVE

l:Si * Ri= 1.2

l:Si * Ri/ l:Ri= 1.2/1.4 = 0.8571= 85.71 % of Drain Valve.

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 7

ISSN 2229-5518

Using mathematical expression l:Si * Ri/ l:Rithe crisp

values for output variables were determined and the
results were found according to the MATLAB simulation
as shown in Fig. 12. These results are compared in Table
12 and found correct according to the design model.
MATLAB simulation was adapted according to the arrangement of membership functions for four rules as given in Table 11.

TABLE 12

COMPARISON OF SIMULATED AND CALCULATED RESULT

TABLE 11

ARRANGEMENT OF MEMBERSHIP FUNCTIONS FOR- SIMULATION

Rule

No.

INPUTS

OUTPUTS

Rule

No.

Water

Level

Flow Rate

Release(Valve)

Drain Valve

1

Below

Danger

Fast

50% opened

50% opened

2

Below

Danger

Very Fast

50% opened

75% opened

3

Danger

Fast

75% opened

Fully

Opened

4

Danger

Very Fast

75% opened

Fully

Opened

In Fig. 12 the same values of input variables, Water Level
= 13, and Flow Rate = 95000 are shown. Various values of
input and output variables match the dependency scheme
of the system design. The simulated values were checked
using MATLAB-Rule viewer as shown in Fig. 12.

The correctness of results shows the validity of the simplified design work for processing system using control system.

Fig. 12 MATLAB-Rule Viewer

4.1. Simulation Graphs Discussion

This system was simulated for the given range of input variables. The given value of: Water Level = 13 lies in region 3 of the range 10-15 and Flow Rate = 95000 lies in region 4 of the range 75000-100000. The four rules were applied for MATLAB simulation according to this range scheme. In this design model, the release and drain control valves depends upon the selected values of water level and flow rate. The simulated and calculated results are according to the dependence scheme.

Fig. 13(a) shows that the control valve is directly proportional to water level and it does not depend upon the flow rate. Fig. 13(b) shows that the drain valve system is directly proportional to flow rate.

Fig. 13(a) Plot between Water Level - Flow Rate Release Valve

Opened/Closed

Fig. 14 Plot between Water Level and Flow Rate Drain Valve

IJSER © 2011 http://www.ijser.org

International Journal of Scientific & Engineering Research Volume 2, Issue 6, June-2011 8

ISSN 2229-5518

5. CONCLUSION

Both the design model and simulation results are same. The designed system can be extended for any number of inputs and outputs. The drain valve control output can be utilized further for land irrigation according to the need and water release control valve for electric generation to fulfill the dire need of this system in automation.
The design work is being carried out to design state of the art fuzzy logic Hydro-Electric control system in future using FPGAs.

REFERENCES

[1] E. Frias-Martinez, G. Magoulas, S. Chen and R.

Macredie,“Modeling human behavior in user-adaptive systems: Recent advances using soft computing techniques” Expert Syst. Appl.29:2 (2005) 320-329.

[2] V. Ramani and M. Rom, “Fuzzy Logic Model on Operation and Control of Hydro-Power Dams in Malaysia” ICCES vol.4 2007, p.p 31-39.

[3] Suharyanto and Goulter, “Use of fuzzy set theory for consideration of storage non-specificity in stochastic dynamic programming for reservoir operation “Proceeding of Seventh IAHR International Symposium on Stochastic Hydraulics, Mackay/Queensland/Australia, 129-36.

[4] Umamahesh and Chandramouli. Fuzzy dynamic programming model for optimal operation of a multipurpose reservoir, Journal of Water Resources Planning and Management, 18(2): 14-20.

[5] Hasebe and Nagayama (2002). Reservoir operation using the neural network and fuzzy systems for dam control and opera- tion support, Advanced in Engineering Software, 33(5): (1996) 245-

260.

[6] Suvo, “How does a Hydroelectric Dam work” http://www.brighthub.com/engineering/civil/articles/42794.asp x 2009.

[7] Shabiul Islam, Shakowat,”Development of a Fuzzy Logic

Controller Algorithm for Air-conditioning System”, ICSE 2006

Proc 2006 IEEE.

[8] M. Saleem Khan and Khaled Benkrid, “A proposed Grinding admixing System using Fuzzy Time Control Discrete Event Model for Industrial Application”, Lecture Notes in Engineering and Computer Science vol. 2175 2009, p.p. 1231-

1236, Directory of Open Access Journals (DOAJ)

[9] M. Saleem Khan and Khaled Benkrid, “Multi-Dimensional Supervisory Fuzzy Logic Control DEV, Processing System for Industrial Applications” Lecture Notes in Engineering and Computer Science vol. 2175, p.p. 1208-1217, Directory of Open Access Journals (DOAJ)

[10] M. Berthold, D. Hand, “Intelligent Data Analysis, 2nd ed.”,

Springer-Verlag, Heidelberg (2006).

[11] M. Saleem Khan and Khaled Benkrid, “Design of Liquids Mixing Control System using Fuzzy Time Control Discrete Event Model for Industrial Applications”, World Academy of Science, Engineering and Technology vol. 72 2010, p.p. 545-553, Directory of Open Access Journals (DOAJ).

Mr. M. Abbas is working as Visiting Faculty in the Department of Electrical Engineering in National University of FAST Lahore and

also Lecturer Internee in Physics, Government

Islamia College Lahore, Pakistan. He com- pleted his Master Degree in Physics with Specialization in Advanced Electronics from Government Islamia College Civil Lines,

Lahore. His research interests include FPGAs and Fuzzy Logic

Control System and Industrial Applications. (e-mail: shairabbas@yahoo.com)

Dr. M. Saleem Khan is an Assistant Professor at GC University Lahore Pakistan. He availed research fellowship from The School of Elec- tronics & Engineering, University of Edin- burgh, UK and completed his Ph.D thesis in the field of control systems: design, simulation and analysis in local and distributed environ-

ment. He contributed his services on various projects in the field of Advanced Electronics and Communication. His research interests include control systems design and industrial applications. He pro- moted a large team of Electronics researchers and organized this field in his country. Mr. Khan had also been served as a Senior Scien- tific officer in a classified defense research organisation in his coun- try.

(e-mail mskgcu@yahoo.c om)

Mr. Nasir Ali is an Assistant Professor in DS College Lahore. Mr.Nasir also served as a Scientific Officer in a classified de- fense research organisation in his coun- try. He has been involved in the various research projects of Electronics.

(e-mail nasirali8787@yahoo.c om

IJSER © 2011 http://www.ijser.org