Improving the Stability of Hydro Power Generator Using ......stability. The mechanical angle between rotor magnetic field and armature magnetic flux of a generator is known as the

International Journal of Advancements in Research & Technology, Volume 5, Issue 6, June-2016 113 ISSN 2278-7763

Copyright © 2016 SciResPub. IJOART

Improving the Stability of Hydro Power Generator Using Neuro-Fuzzy Techniques

Enebechi Chukwuemeka Theophilus and Prof I. I. Eneh Department of Electrical Electronic Engineering

Enugu State University of Science and Technology, (ESUT), [email protected]

Abstract In this work, an intelligent device that will continuously supervise and adjust any disturbances that might arise in a hydropower turbine system is presented. To this effect, a Neuro-fuzzy control device that will improve the stability of Hydropower generator is proposed. An effective control scheme in which a variable fuzzy based controller is cascaded with a Neural network identifier called Neuro-Fuzzy technique is demonstrated. This is achieved by using simulation package which is fuzzy logic Toolbox and SIMULINK in MATLAB Software. Water level and flow rate used as input variables are fed into MATLAB Software to design fuzzy controller to generate the fuzzy rule. The fuzzy rule is imbibed inside the fuzzy logic controller (FLC) and simulated to obtain output results for the release control and drain valve which represents the operation of the system overtime. The Augmentation of a classified Neural network to produce Neuro-fuzzy controller will enhance performance and significantly reduce limitations of rise time, settling time, and steady state error observed from conventional PLC-HMI based fuzzy scheme. The overall result is to satisfy the sudden load change and maintain constant turbine speed with less settling time. KEYWORDS: Hydropower, Hydro turbine, Fuzzy Logic Toolbox, Rule Editor, Stability, SIMULINK, Fuzzy Logic, Neural Network. 1.1 Introduction Since the advent of Hydropower generation, a characteristic common to electric power systems is that they operate under the influence of disturbances. As economic considerations continue to demand the operation of power systems closer to their stability limits, there is an increasing need for reliable and accurate means to determine limiting operating conditions. However, due to critical conditions, emotional and psychological stress, an engineer or operator in various scientific or industrial areas, where most of the real time systems are generally complex and difficult to be controlled, of which hydropower plant is one, operator may not be able to instantly, and continuously respond to making correct decisions as disturbances arise. Mistakes can damage very expensive power equipment or worst still lead to major emergencies and catastrophic situations. As a result, there is a strong need for an automated control device that can assist operators to improve on the robust control of hydropower turbine speed in order to stabilize and ensure optimal performances within the predetermined ranges using various intelligent techniques. [1]. In view of this, an intelligent robust power monitoring device that will continuously supervise and modulate their respective control action is proposed in this study. Thus, an Adaptive control

based scheme, in which a variable fuzzy based Controller is cascaded with a Neural network identifier called Neuro-fuzzy technique is presented in order to improve the stability of hydro power generation. Adaptive Neuro-fuzzy techniques are also referred to as “Adaptive Network based-fuzzy inference system” (ANFIS). [2] 1.2 Hydropower generation Hydropower plants capture the energy of falling water to generate electricity. A turbine converts the kinetic energy of falling water into mechanical energy. Then a generator converts the mechanical energy from the turbine into electrical energy.

IJOART



Fig. 1: General overview of Hydropower plant This scheme is aimed at designing an automated corrective procedure that can assist operator to improve on the robustness, stability and accuracy of input variables in order to stabilize and ensure optimal performances of turbine speed within expected range. The knowledge of Hydropower liquid level controller and controllers for flowing fluids, dam, Turbines, generator and the possible application of intelligent agents is needed to professionally discuss this phenomenon. Hydro power electric system is a main source of renewable energy for power generation. The aim of this control system is to keep the system within the predetermined ranges by controlling the flow through a control valve at the dam and outflow through drain valve in any condition for safety as well as efficient hydro electricity generation. Controlling reservoirs manually in dams is very difficult because it has nonlinear or time-variable behavior such as sudden changes in reservoir water level [3]. To manage the drawbacks in manually operated controls such as ecosystem damage, situation, flow shortage, etc., an efficient and accurate method based on intelligent agent controller is proposed for operating system in dam to solve that problem. Dams are reservoirs or lakes primarily for strong water or to produce electricity. The moving water causes the turbine to spin, which causes magnetic field inside a generator to rotate and create electricity. The mechanical energy produced as the water flows through the turbine is converted into electrical energy by the generator. The shaft, exciter, rotor and the stator are the major components of the generator. The exciter of the generator sends an electrical current to the rotor when the flow of water through the penstock to turbine turns the shaft as shown in figure 2.

Fig.2 Hydro turbine connected to the Shaft and Generator. Generators in hydropower plants work just like the generators in other types of power plants. The penstock is described as the tunnel, channel, pipeline

through which quantity of water flow for hydro power generation is passed or delivered to the turbine. The penstock is constructed inside the dam and starts at the upstream side of the turbine positioned inside the downstream to the powerhouse and the water pressure increases as it flows down through the penstock to the turbine. The released water from the dam reaches the turbine blade through the penstock. The proper slope and diameter of the penstock is important for the efficiency of the dam. A control gate is used for releasing reserved water from the dam. Depending upon the electricity requirements the gate is opened. The turbine consists of a number of large fan blades and a spindle which rotates when they strike the blades. Thus, the power of flowing water is converted to rotational power of the spindle. The spindle of the turbine is connected to the alternator while rotational power of the spindle is converted into electrical power. The produced electricity is then distributed to the grid. The outflow of water from the turbine is released to a river. In order to generate electricity from the kinetic energy moving water has to move with sufficient speed and volume to spin a propeller –like device called a turbine, which in turn rotates a generator to generate electricity.

Figure 3: Hydro power turbine Hydropower turbines are selected depending on factors such as head, the flow of water and site conditions. There are two types of turbine namely; Impulse turbine and Reaction turbine. 1.3 Power System Stability Stability of power system is related to stability of synchronous generator. The rotor angle stability is considered as the parameter affecting system stability. The mechanical angle between rotor magnetic field and armature magnetic flux of a generator is known as the load angle or power angle (𝜕). Basically, power system stability is a synchronism between rotating field flux and circulating armature flux [4]. Stability is the ability of a system to withstand the maximum power flow transfer from generation to

IJOART



load. Power system stability is best defined as the ability of an electric power system to regain a state of operating equilibrium after being subjected to a physical disturbance, when variables are bounded so that practically the entire system remains intact. 1.4 Neuro-Fuzzy

The Neuro-fuzzy integrated system can be trained by numerical data and linguistic information expressed by fuzzy if-the rules. Neuro-fuzzy system (NFS) incorporates the human-like reasoning style of fuzzy systems through the use fuzzy set and a linguistic model consisting of a set of if-THEN fuzzy rules. The main strength of neuro-fuzzy system is that they are universal appropriators with the ability to solicit interpretation of IF-Then rules. Fuzzy logic controller is an attractive choice when mathematical application is not possible. The three principal elements in the fuzzy controllers are;

a. Fuzzification module b. Rule based and inference engine c. Defuzzification module (Deffuzzifier )

This feature makes the incorporation of prior knowledge into the design of controllers possible. Another important feature of the neuro-fuzzy integrated system is that, without any given initial structure, the system can construct itself automatically from numerical training data. The application of neuro-fuzzy system to fast valving for stability improvement has two inputs which includes. Accelerating power and rotor speed. The Neuro-fuzzy system is stable, efficient and reliable. Fig.4 shows Sugeno’s fuzzy logic model while Fig. 5 shows the architecture of the ANFIS, comprising of input, fuzzification, inference and defuzzification layers. The network can be visualized as consisting of inputs, with ‘N’ neurons in the input layer and ‘F’ input membership functions for each input, with ‘F*N’ neurons in the fuzzification layer. There are FN rules with FN neurons in the inference and defuzzification layers and one neuron in the output layer. For simplicity, it is assumed that the fuzzy inference system under consideration has two inputs x and y and one output F as shown in Fig. 5. For a zero-order Sugeno fuzzy model, a common rule set with two fuzzy if-then rules is the following: Rule 1: If x is A1 and y is B1, Then f1 = r1 (1) Rule 2: If x is A2 and y is B2, Then f2 = r2 (2)

Fig. 4: Sugeno’s Fuzzy Logic Model

Fig. 5: The Architecture of the ANFIS Here the output of the 𝑖𝑡ℎ node in layer n is denoted as On1

2

Layer 1. Every node i in this layer is a square node with anode function.

O𝐼 1 = µAi(x), for i = 1, 2 (4)

Or O𝐼 1 = µBi - 2 (y), for i = 3, 4 (5)

Where x is the input to node-i, and Ai is the linguistic label (small, large, etc.) associated with this node function. In other words,O𝐼

1 is the membership function of Ai and it specifies the degree to which the given x satisfies the quantifier Ai. Usually µAi(x) is chosen to be bell-shaped with maximum equal to 1 and minimum equal to 0, such as the generalized bell function: Agus et al; 2012.

µA (x) = 1

1+�𝑥− 𝑐𝑖𝑎𝑖

�2𝑏𝑖

(6)

Parameters in this layer are referred to as premise parameters.

IJOART



Layer 2. Every node in this layer is a circle node labeled II which multiplies the incoming signals and sends the product out. For instance,

O𝐼 2 =𝑤𝑖 = µAi(x) x µB(y), for i = 1, 2 (7)

Each node output represents the firing strength of a rule. (In fact, other T-norm operators that perform generalized operations/functions and can be used as the node function in this layer.)

Layer 3. Every node in this layer is a circle node labeled N. The i-th node calculates the ratio of the i-th rule’s firing strength to the sum of all rules firing strengths:

O𝐼 3 = 𝑤�= 𝑤𝑖

𝑤1+ 𝑤2, i = 1, 2 (8)

For convenience, outputs of this layer will be called normalized firing strengths.

Layer 4. Every node i in this layer is a square node with anode function:

O𝐼 4 = 𝑤�1𝑓1 = 𝑤�1(𝑝1 𝑥 + 𝑞1 𝑦 + 𝑟1) (9)

Where 𝑤��1is the output of layer 3, and (𝑝1,𝑞1,𝑟1 is the parameter set. Parameters in this layer will be referred to as consequent parameters.

Layer 5. The single node in this layer is a circle node labeled Σ that computes the overall output as the summation of all incoming signals, i.e.,

O𝐼 5 = ∑𝑤�1𝑓1 (10)

In flow process fuzzy logic is taken as the controlling unit. The fuzzy controller implements the control algorithm, compares the output with the set pit and then gives necessary command to the final control element via actuator unit (Sa) neural network equipped with the capability of handling fuzzy information [termed fuzzy neural network (FNN) and a fuzzy system augmented by neural networks to enhance some of its characteristics like flexibility, speed, and adaptability [termed neural-fuzzy system [NFS]. An important feature of Neuro-fuzzy integrated system is that, without any given initial structure, the system can construct itself automatically from numerical training. However, this study did not only demonstrate an appreciable improvement in hydropower system stability, efficiency and reliability but also optimize the sudden load change and also maintain the expected constant speed.

2.0 Related work The huge venture of providing electrical energy suffers various engineering difficulties that poses the engineer with a range of challenges. These various difficulties motivate the engineer's technical abilities to use their knowledge and experience in designing a better system. To be able to guess the outcome of such complicated systems, the engineer is dragged to look for highly powerful tools of assessment and synthesis.

[5], studied Neural Cascaded with Fuzzy scheme for control of hydroelectric power plant using pump storage to manage the high load demand. Back propagation algorithm was used to excite the neural network. Which involves repeating the input data to the neural network. It was observed that rise time is faster in proposed scheme than in conventional scheme because of dead time. Similarly, the performance evaluation criteria in the conventional method have large ISE (Integrated system error) and IAE (Integrated angle error) errors when compared with the proposed scheme. This emphasizes the better performance of Programmable Logic Controller- (PLC-HMI) based Neural network cascaded with fuzzy scheme than the conventional Fuzzy method. [6], in his work developed a two area thermal-hydro power system interconnected with fuzzy logic and adaptive neural network controller to illustrate the performance of load frequency control using MATLAB/SIMULINK package for better dynamic behavior. By his proposed method the responses obtained are better than the previous work already reported by Ram et al., 2010. [7], showed that in hydrothermal power plant an exact signal is necessary to manipulate the gate that has to be identified and it will match the requirement of the sudden large change in the loads. Also he described that the maximum sudden increase and decrease in load system has to be expected and consequent manipulation in gate position has to be well known. With the intention of enhancing, the performance of the hydroelectric power plant, the PLC-HMI based fuzzy controller scheme was developed. [8], highlighted on controlling the variables water level and flow rate with real time implementation of gate control in hydroelectric power plant using Programmable Logic Control for automation in small hydroelectric plants. The water level in the reservoir is measured by capacity level sensor. Comparison between the desired water flow and actual flow rate is done and the deviation is controlled by the PLC

IJOART



which gives manipulated variables to the gate valves. The water outflow from the gate valve is taken to the turbine. Due to the sudden increase and decrease in load, the gate opening/closing operation has limitation in speed deviation but it resulted in damages to turbine/generator shaft. Since slow and smooth operation of goals are preferred to prevent such limitations in speed deviations observed from [8] research, using Programmable Logic Control, it was then necessary to get a suitable signal to operate the gate which will satisfy the sudden load change and also maintain the constant speed. [

In this case, a cascaded PLC-HMI based fuzzy scheme is proposed by [9].The study dwell on controlling the variables water level and flow rate with real time implementation of control valves. Here, the maximum sudden in and out load change is expected and equivalent manipulation in valve position is obtained using the proposed scheme. Comparisons of time domain specifications such as rise time, peak time and setting time for conventional and the proposed PLC-HMI based fuzzy scheme were demonstrated. Similarly, analyzing performance evaluation criteria, the conventional control method has large integral square error (ISE) and integral absolute error (IAE) when compared with the proposed scheme. This highlights the better performance of PLC HMI based fuzzy over the conventional control scheme. The result of this real time work emphasize the robustness of the PLC-HMI based Fuzzy controller scheme for step variation in loads. It was observed that there were some limitations such as time delay and low speed deviation which formulates damages to turbine/generator shaft, and the pressure across the turbine reduced significantly making power to reduce during dynamic load change. As a result of the various researches due to drawbacks of the existing techniques this research is proposed with the aim of developing a modern type of control technique that can be used for better performance of power stability. The foregoing review calls for the need to introduce an intelligent monitoring device that will continuously supervise and modulate the expected control action for a real time robust and accurate system performance in this study. The intelligent control based scheme, neural network identifier cascaded with the Fuzzy PLC-HMI is called Neuro- fuzzy technique

3.0 Method of design Several milestones have been recorded but the quest to improve standard has continued. In a bid to make an enduring professional input in this regard, it has become quite pertinent to scientifically propose an intelligent technique controller that will guarantee the stability of the plant as the frequency returns to its steady state in a very short time. [10], revealed that encoded fuzzy sets, fuzzy rules, procedures, the rule base and the membership functions have great influence on the performance and efficiency of the plant. Hence, the task is to design a simulation of fuzzy logic controller that will stabilize the level of water and the rate of flow whose force of flow spins the turbine at predetermined speed. The result of the simulation will be displayed using Rule viewer which is part of the graphical user interface (GUI) tools in fuzzy Logic toolbox in MATLAB software. The purpose is to find the best way to get the result as close as the required stability as possible so that the force of water spins the turbine at an accurate and stable revolution per minute. In the design two important inputs parameters: namely the water level and flow rate ultimately play important roles in turbine and shaft rotation, Also stator, rotor and excitation are taken into consideration. The two output parameters: Release valve control and drain valve control are used for efficient and robust hydropower stability. To maintain these parameters within the prescribed optimum level using water level and flow rate input variables, a measuring device for feedback control (Fuzzy inference Editor) for the two output, the drain and control release valves of the controller, is required. The Fuzzy Inference Editor or the rule Editor which allows us to interpret the entire fuzzy inference process at once and also shows how the shape of certain membership functions influences the overall result. Nevertheless, the rule in rule editor provides the control strategy. FLC uses the rules for a straight forward implementation proposed to solve a class level of control problems with unknown dynamics or variable time delays commonly found in industries, given attention to various parameters, such as the time of response, the error of steadying and overshooting. This study, discusses the new controller based on a hybrid Neural network identifier cascaded with a Fuzzy logic based controller called Neuro-fuzzy controller designed to classify and predict the system stability when disturbances occur in a hydropower generator. Hence, the purpose of this present work is to create such a controller that has to be vigorous

IJOART



enough to withstand the required robustness, stability, accuracy and maintain the expected constant speed of the proposed Neuro-fuzzy controller, than when compared with the conventional Fuzzy logic cascaded with Programmable logic control based scheme. The first step is to apply defuzzification to form a crisp output. In the defuzzification step, the Water level and Flow rate are selected as input variables. Generally, these input variables are divided into three fuzzy sets and they are linguistically named as HIGH, LOW and OK as shown in Figures 6 and 7 below. The Gaussian membership functions with the appropriate ranges have been used for these fuzzy sets. The values of the valve have been selected as Fuzzy output variables.

Figure 6: Membership function of fuzzy set characterizing input variables.

Figure 7: Membership function of fuzzy set characterizing output variables (Change in error). Like input variables, the output variables are divided into five fuzzy sets with linguistic names; OpenFast, OpenSlow, NoChange, CloseSlow and CloseFast as shown in Figure 8 below.

Figure 8: Output membership function of valve.

The six rules are: 1. If (Flow level HIGH) then (Valve is CLOSEFAST) 2. If (Flow level is OK) then (Valve is NOCHANGE) 3. If (Flow level is LOW) then (Valve is OPENFAST) 4. If (Flow level is OK) and (Flow rate is POSITIVE) then (Valve is CLOSESLOW) 5. If (Flow level is OK) and (Flow rate is OK) then (Valve is NO CHANGE) 6. If (Flow level is OK) and (Flow rate is NEGATIVE) then (Valve is OPENSLOW) The centroid method has been used to obtain the crisp value [10]. INVESTIGATING THE STABILITY OF HYDROPOWER SYSTEM USING PLC WITH CASCADED FUZZY LOGIC SCHEME In stability control using conventional PLC-HMI scheme, the first step taken was to design the Fuzzy Inference System.. The FIS has two Inputs. One is Level of the liquid in the DAM defined as "level" and the other one is flow of change of liquid in the dam denoted as "rate". Both Inputs are applied to the Rule Editor. According to the Rules written in the Rule Editor the controller takes the action and governs the opening of the Valve which is the Output of the controller' and are denoted by "control valve and drain valve" [ 8]. This is shown figure 9.

Fig. 9: Hydro Power Plant Rule Editor Fuzzification transform a numerical variable (real number or crisp variable) into a linguistic variable (fuzzy number). The occupied region description, Triangular membership functions and range for two input variables are given in table 2, 3,4 and 5 The Rule editor also shows how the shape of certain membership functions influences the overall result. Rules shown in Rule editor provide inference mechanism strategy and producing the control signal output. In this research total number of active rules obtained is equal to 25 rules (=52 ) as shown in figure below on table 1 and 2

IJOART



Table 1: Total Number of Rule for Release control valve Flow rate/water Level

Very Slow

Slow Normal Fast Very Fast

Above Danger

Close Fast

Close Fast

Close Fast

Close Fast

Open Slow

Danger Close Fast

Close Fast

Close Fast

No Change

No Change

Below Danger

Open Slow

No Change

Close Slow

Close Slow

Open Slow

Low Open Fast

Open Fast

Open Fast

Open Fast

Open Fast

Very Low Open Fast

Open Fast

Open Fast

Open Fast

Open Fast

X X1 X2 X3 X4 X5 The table above shows the 25 variable numbers of Rules generated for Release control valve.

Table 2: Total Number of Rule for Drain Valve Flow rate/water Level

Very Slow

Slow Normal Fast Very Fast

Above Danger

Close Fast

Close Fast

Close Fast

Close Fast

Open Slow

Danger Close Fast

Close Fast

Close Fast

No Change

No Change

Below Danger

Open Slow

No Change

Close Slow

Close Slow

Open Slow

Low Open Fast

Open Fast

Open Fast

Open Fast

Open Fast

Very Low Open Fast

Open Fast

Open Fast

Open Fast

Open Fast

X X1 X2 X3 X4 X5 Table 2 shows the 25 variable numbers of Rules generated for Drain and Release control valve see, the first 10 generated membership functions in figure 10, and their triangular functions shown in figure14. The triangular membership functions for water level, Flow rate and the output results (Drain and Release control) are displayed in figures 11,12 and 13.

Fig.10: Membership Functions Generated for 1 - 10 by fuzzy Rule Editor.

The inference engine involves four and operators that select minimum input values for the output. Four inputs (F) of fuzzifier were accepted by inference engines and applied to the min- max composition to obtain the output value (F). Inference Engine for four inputs (F) of Fuzzifiers are calculated, as shown in figure 3. In arriving at our values, appropriate domains are chosen on the x axis, since it gives us the best level of stability as desired in this work. The ranges are 0 < x < 10, A domain of x is chosen from the number line as shown below: x < 0 x < 10 0 2 4 6 8 10 x1 x2 x3 x4

From the inequality F1 = 10 To find F2, two members are chosen from the domain to give us a steady state flow in the system. Thus, we have x1 + x4 and x2 + x3. Therefore, F2 = x1 + x4 = 2 + 8 =10. Similarly, F3 = x2 + x3 = 4 + 6 = 10. Using alternate formula, F2 = F4 = 10. Thus, the four fuzzification results F1, F2, F3, and F4 have the same value of 10. Input Variables

Region selection

Fuzzy set Calculation

Water level 0 < x < 10 0 < y < 1

F1 = 10 F2 = x1 + x4 = 2 + 8 =10. F3 = x2 + x3 = 4 + 6 = 10. F2 = F4 = 10.

Flow rate 0 < x < 10 0 < y < 1

F1 = 10 F2 = x1 + x4 = 2 + 8 =10. F3 = x2 + x3 = 4 + 6 = 10. F2 = F4 = 10.

Table 3: Fuzzification results The triangular membership functions are used to represent the linguistic terms (VS Very small; S small; M medium; I. large; VI. Very large). Rule base is developed and an inference mechanism is designed using Mamdani (min-max) method that gives the output signal to the final control elements. The maximum number of overlapped fuzzy sets generates the membership functions for the Drain valve and Release control valve is shown in Figure 13. The total number of rules is equal to the product of number of functions accompanied by the input variables in their working range. The two input

IJOART



variables described here consisted of five membership functions. While the generated membership functions for the Water level, Flow rate, Release control, and Drain valve are displayed below

Fig.11: Triangular Membership Functions Generated for Flow rate Two crisp values inputs - Water level and Flow rate received by the rule selector generate variables according to the rules for each triangular membership function as shown below i.e. Water level- Above danger, Danger, Below Danger Low and Very low and Flow rate – Very slow, Normal, Fast and Very fast in Figure 11 and 12 Respectively.

Fig.12: Triangular Membership Functions Generated for Water Level

It could be observed that from the triangular membership functions of output variable for Drain and Release control valves both have the same output variables of Closed fast, closed slowly, No change, opened slow and opened fast.

Fig. 13 membership functions Generated for Release Control and Drain control valves. The rules are based on ‘Mamdani inference method’. The simulation results are obtained using a 25 rule Fuzzy Logic control. Rules shown in Rule editor provide the control strategy. The parameters of the FLC system designed for the hydro electric power plant are presented as shown above.

Fig. 14: The triangular membership functions for 25 Rules.

Using a range of 0 to 1 for Water level and 0 to 10 for Flow rate 5 membership functions for water level and Flow rate, the output format is delegated to each of the functions, as written below; Output Format; Close Fast = 0, Close Slow = 2, No change = 4 , Open Slow = 6, Open Fast = 8; The table 3 shows the member functions for Control Valve.

Table 4. Design value for Flow Rate membership functions and Range of input Variable

Membership function (MF)

Ranges

X (axis)

Region

Y (axis)

Panam

Closed fast 0 - 10 0 - 1 -404 Close Slow 0 - 10 0 - 1 1.0004274 No Change 0 - 10 0 - 1 -3.9967

.996 Opened slow 0 - 10 0 - 1 1.8975

.899

.897 Opened fast 0 - 10 0 - 1 0.96159

.8822096 The table above describes the design data for the input variable of Flow rate, see fig. 11 for the method of design and analysis.

IJOART



The act of simulating something first requires that a model be developed; which represents the key characteristics of the selected physical or process.

A module for the PLC-HMI scheme Fuzzy logic based is designed as shown below 15. Recall that the fuzzy rule is imbibed inside the fuzzy logic controller to enhance the efficiency of controlling the expected stability. Fuzzy rule helps to guard the liquid level of the Dam. Without imbibing the rule inside the fuzzy logic controller, the hydropower generator will NEVER run, let alone giving an accurate result.

The input variables of water level and Flow rate are fed into the Fuzzy logic controller which is inserted inside the hydropower generator model. This is then simulated, to obtain the output result for both the Release control and Drain valves, which represents the operation of the system overtime.

The graph so obtained shows the level of instability using fuzzy logic controller alone as shown in Figure 17 and 18.

Table 5. Design value for Water level membership functions and Range of input Variable

PARAM Membership function (MF)

Ranges Region Occupied

-404 Above Danger 0 - 10 0 - 1 -126273674 Danger 0 - 10 0 - 1 1.064 5.0649 .064

Below Danger 0 - 10 0 - 1

-2.897 .06217

Low 0 - 10 0 - 1

0.7522 .9892073

Very Low 0 - 10 0 - 1

The table above describes the design data for the input variable of water level, see fig. 12, for the analysis.

Table 6. Design value for Release control Valve membership functions and Range of output Variable


Ranges Region Panam

Closed fast 0 - 10 0 - 1 -5368

-011752622

Close Slow 0 -10 0 - 1 -0.0042743996798

No Change 0 - 10 0 - 1 4.72615137

Opened slow 0 -10 0 - 1 -8.827938112

Opened fast 0- 10 0 - 1 1.19710032121

The table above describes the design data for the input variable of Release control valve, see fig. 13 for the method of design and analysis.


Ranges Region Panam

Very slow 0 - 10 0 - 1 -346900913 .651

Slow 0 - 10 0 - 1 133449 Normal 0 - 10 0 - 1 326854171127 Fast 0 - 10 0 - 1 -6.3257

.7251364 Very fast 0 - 10 0 - 1 1.339

.95213 Table 7: Design value for Drain Valve membership functions and Range of input Variable The table above describes the design data for the input variable of Drain valve, see fig. 3.14 for the method of design and analysis. The analysis of the output result for both the Release control and Drain valves is shown in figure 18.

The MATLAB simulation is shown below L = [0 1000 1000 1000 1000 1000 1000]; T = [0 0.2 1 4 6 8 10]; plot(T,L,'r') grid on Ylabel ('RELEASE CONTROL VALVE');Xlabel('TIME(S)')

Fig. 16: Graphical representation of Fuzzy scheme output for Release control valve. The programs for the MATLAB simulation are of the conventional fuzzy scheme for both Release control and Drain Valves respectively: D = [0 1000 1000 1000 1000 1000 1000] T = [0 0.2 1 4 6 8 10] Plot [T D “b”] Grid on Y label [‘Release Control Valve] The MATLAB simulation for the Drain valve is shown below:

IJOART



%T = X D = [0 500 500 500 500 500 500] T = [0 0.2 1 4 6 8 10] Plot [T D “b”] Grid on Y label [‘Drain Valve level]

tank 2

WATERTANK

fLOW RATE

84 .85

Subtract

Subsystem2

VALVE

Subsystem1

VALVE

SignalGenerator

Scope 2

Scope

S-Function

animtank

QUANTITY OF WATER

1500

Product

OUTLET PRESSURE

1000

NEURAL NETWORK SUBSYSTEM

In1 Out1

Mux

Mux

MathFunction

sqrt

INLET PRESSURE

6000

Fuzzy Logic Controller 2


FLUID SPECIFIC

1

FLOW COEFFICIENT

1.2

Divide

Display

84 .85

du /dt

DRAIN TANK

In1 Out1

DRAIN LEVEL GRAPH

DRAIN LEVEL

500

CONTROLTANK

In1 Out1

CONTROL LEVEL GRAPH

CONTROL LEVEL

1000

Add

Figure 15: The Model for Fuzzy Cascaded with PLC-HMI

tank 2

WATERTANK

fLOW RATE

84 .85

Subtract

Subsystem2

VALVE

Subsystem1

VALVE

SignalGenerator

Scope 2

Scope

S-Function

animtank

QUANTITY OF WATER

1500

Product

OUTLET PRESSURE

1000

NEURAL NETWORK SUBSYSTEM

In1 Out1

Mux

Mux

MathFunction

sqrt

INLET PRESSURE

6000



FLUID SPECIFIC

1

FLOW COEFFICIENT

1.2

Divide

Display

84 .85

du /dt

DRAIN TANK

In1 Out1

DRAIN LEVEL GRAPH

DRAIN LEVEL

500

CONTROLTANK

In1 Out1

CONTROL LEVEL GRAPH

CONTROL LEVEL

1000

Add

Figure 16: The Model for Fuzzy Cascaded with PLC-HMI

IJOART



Fig 17: Graphical representation of Fuzzy scheme output for Drain valve.

The MATLAB simulation program is shown below: %T = X D = [0 500 500 500 500 500 500] T = [0 0.2 1 4 6 8 10] Plot [T D “b”] Grid on Y label [‘Drain Valve level]

Fig. 18 Graphical representation of Fuzzy scheme Release Control valve output The MATLAB simulation program is shown below: %T = X D = [0 1000 1000 1000 1000 1000 1000] T = [0 0.2 1 4 6 8 10] Plot [T D “b”] Grid on Y label [‘Release Control Valve]

INVESTIGATING THE STABILITY OF HYDROPOWER GENERATOR USING NEURO - FUZZY TECHNIQUE As a result of limitations observed at the outputs of the two earlier cases above; (i.e. the conventional Fuzzy logic scheme, and the Neural network scheme), there is need to improve on the stability of hydropower generator by cascading the conventional Fuzzy logic scheme with a classified Neural network, known as “Neuro- Fuzzy” scheme is shown below. However, the neural network model was inserted into the conventional scheme of fuzzy base model as shown in figure 17 and 18. The result obtained after simulation of the hybrid shows a constant speed with a reduction of settling time. This emphasizes the better performance evaluation of PLC-HMI based neural cascaded with fuzzy scheme from the conventional method on different load variations.

Time (s)

Fig 19: Graphical representation of the Output for release control valve for Neuro-fuzzy scheme.

Time (s)

Fig 20: Graphical representation of the Output for drain valve for Neuro-fuzzy scheme.

Vol

ume

of w

ater

leve

l

Vol

ume

of w

ater

leve

l

IJOART



The overall qualitative and quantitative operation of the proposed technique was investigated and the output result shown in Figure 19 and 20 depicts a faster Rise time, a constant speed with reduction of settling time.

Simulation Result Analysis

Result of the simulation in both output Release the control and Drain valves for the Neuro-Fuzzy scheme and the output of the Release control and Drain valves for the conventional PLC-HMI Fuzzy scheme and on the other hand were critically analyzed. Comparing the output release control valve for the conventional fuzzy scheme with the output release control valve for the Neuro-fuzzy scheme, it was observed that in Figure 17 & 18 and 19 & 20, while the Neuro-Fuzzy scheme shown in the two graphs. It could thus be concluded that the Neuro-Fuzzy scheme provides a better real time robust and accurate stability for a hydropower generator than the PLC Fuzzy logic cascade.

CONCLUSION

We have been able to establish that compared to the conventional PLC-HMI Fuzzy scheme, this Neuro-Fuzzy scheme shows a better control action for a real time robust and accurate stability of a hydropower generator. This has provided us with an obvious better and more reliable method for solving commonly known problems of time delay and settling time commonly found in industries.

Figure 20: Visual basic display of a neuro-fuzzy hydropower generator stability.

REFERENCES [1]Pratibha Srivestar, Manojjha, M.F.Qureshi: ‘Co-active Neuro Fuzzy inference system for covering control & excitation control of power system stability’; International Journal of Innovation Research in Science, Engineering and Technology, Vol. 3, Issue 6, June 2014. [2]Mahusudhana, G. Rao, Anwesha Kumar, B. V. Sankar Ram., Design of Neuro fuzzy Inference distributed power flow controller for transient stability improvement, Word Journal of Modeling and simulation, Vol. 10(2014) No. 4 pp. 252-2 [3]Rajeswari, V., Y. Rajeshwari and L.P. Suresh, (2012). Real time implementation of hydroelectric power plant using PLC SCADA. Int. J. Eng. Res. Applic., 2:899-902. [4]A. G. Abro & J. M. Saleh, Control of power system stability {reviewed solutions based on intelligent systems (International Journal of InnovativeComputing, Information and Control ICIC International c2012 ISSN 1349-4198Volume 8, Number 10(A), October 2012 pp. 6643 - 6666) [5] Selwin Mich Priyadharson A., Saravanan M.S., Gomathi N., Vinson Joshua S. and Mutharsan A. (2014). Energy efficient flow and level control in a hydro power plant using fuzzy logic. Journal of computer Sceince 10(9): 1703 - 1711 [6] Surya P., Sinha, S.K., (2011) Application of artificial intelligence in load frequency control of interconnected power system; International Journa of Engineering , Science and Technology Vol. 3, No, 4 pp. 264 – 275. [7] Singh, G. and D.S. Chauhan, 2011 Simulation and Modeling of hydro power plant to study time response during different gate states. Int. J. Advanced En. Sci. Technology, 10: 042-047. [8]Falqueto, J. and M.S. TeIles. 2007. Automation of diagnosis of electrical power transformers in Itaipu Hydropower plant with a fuzzy expert system. Proceedings of IEE Conference on Emerging Technologies and Factories Automation, Sept. 25-28, IEEE Xplore Press. Patras, pp: 577-584 DOI: 10.1109/EFTA., 2007. [9]Rajeswari, V., Y. Rajeshwari and L.P. Suresh, (2012). Real time implementation of hydroelectric power plant using PLC SCADA. Int. J. Eng. Res. Applic., 2:899-902. [10]Dharamniwas and AZIZ Ahmad and Varun Redhu and Umesh Gupta, July, 2012, Liquid level control using Fuzzy Level Controller

IJOART



IJOART

Documents

Improving the Stability of Hydro Power Generator Using ......stability. The mechanical angle between rotor magnetic field and armature magnetic flux of a generator is known as the