Novel Method for Voltage and Frequency Control for

Wind Driven Induction Generator

A.H.Besheer
Electrical Engineering Department
Faculty of Engineering – University of Tabuk
Tabuk - Kingdom of Saudi Arabia
abesheer@ut.edu.sa

Abstract— This paper introduces a new tuning technique for the controller [2], [3]. Since only a capacitor bank needs to be
PID control scheme that is adopted for regulating the magnitude connected to the stator terminals, the SESRIG provides a good
and frequency of the wind energy conversion system (WECS) quality ac source with little harmonic distortion to the stator
output voltage which presents interesting control demands and load. Another advantageous feature of the SESRIG is that
exhibits intrinsic non-linear characteristics. The developed
independent control of the voltage and frequency can be
technique is stem from blending the ant system optimization
method with Takagi Sugeno fuzzy system approach. The achieved easily. Even with a wide variation in speed, the
optimum relationship between the PID controller gains and the generator frequency can be maintained reasonably constant by
parameters of fuzzy modules is explored using Ant system rotor resistance control, while the voltage can be controlled by
algorithm; hence a satisfactory dynamic performance of the varying the excitation capacitance. The rating of the rotor
WECS is obtained. resistance controller is small compared with the generator
rating; hence, the cost saving is quite significant.
Keywords-Takagi - Sugeno fuzzy control; PID control; Ant Fuzzy control and tuning methodologies have emerged in
system optimization; Wind energy conversion systems
last years as promising ways to approach nonlinear control
I. INTRODUCTION problems and to deal with difficulties of conventional PID
In the world of today there is a need for alternatives to the controllers [4]. The area of auto-tuning of PID controller using
large coal and oil fired power plants. Renewable energy is one fuzzy systems has attracted many authors [5][6]. However, a
way to go, and in particular wind turbines have proven to be a common bottleneck encountered in fuzzy controller design is
solution [1]. The arrival of the new power devices that derivation of fuzzy rules is often difficult and time-
technologies, new circuit topologies and novel control consuming, and relies on expert knowledge. To overcome this
strategies are contributing to the success of the wind disadvantage, many automatic methods for fuzzy system
generation technology. design and optimization algorithms have been proposed, such
In wind energy standalone generation systems, the as neural fuzzy systems [7]and evolutionary fuzzy systems [8].
frequency and the magnitude of the terminal voltage for the Recently, the social inspired optimization algorithms
Self Excited Induction generators (SEIG) vary with load even become a successful alternative for the conventional tuning
when the rotor speed is maintained constant. An increase in method to adapt the PID controllers. Using these new
the rotor speed due to the increase of wind speed will result in methods, the global optimal or suboptimal solutions of the
a proportionate increase in frequency, often accompanied by optimized control scheme are found. These algorithms are
severe overvoltage and excessive current. adopted by many researchers for tuning PID controller in its
Although the slip-ring machine is more expensive and classical and intelligent forms. One such approach is Ant
requires more maintenance, it permits rotor slip-power control System Optimization (ASO) which is founded on the foraging
when driven by a variable-speed turbine. When a grid behavior of ants and their indirect communication based on
connection is permissible, the slip-ring machine may be pheromones see for instance [9]. ASO has been applied to
operated as a double-output induction generator (DOIG) using several combinatorial problems such as job scheduling,
the slip-energy recovery technique. In the case of Self Excited routing optimization in data communication networks and
Slip Ring Induction Generator (SESRIG), the system cost can
telephone networks [10].
be further reduced by the use of a simple rotor resistance
 One of the widely used methods for optimizing a fuzzy This paper is organized as follows: section II shows the
logic controller is to modify the rule base using this self- main feature of the control characteristic for SESRIG used in
organizing algorithm automatically according to previous the proposed standalone WECS. Takagi Sugeno fuzzy tuner
responses until the desired control performance is achieved. for PID controller is presented in section III. Section IV
Many works studied this problem such as [11], [8] and [12]. In introduces ant system optimization algorithm for optimal
[11], a tracking control problem is assumed to develop a self parameter determination for Takagi Sugeno fuzzy systems.
tuning PID control scheme with application to antilock The proposed control strategy is applied to WECS in section
braking system via combinations of fuzzy and genetic V. Finally, a conclusion is set in section VI.
algorithms. While in [13], the approximation of analytical test
function is performed using ant colony algorithm to tune the II. SESRIG VOLTAGE AND FREQUENCY CONTROL
parameters of Takagi Sugeno (T-S) fuzzy rule base. The Normalized equivalent circuit of the steady state model of
authors simplified their work by assuming constant SESRIG can be proposed as seen in Fig. 1 [14]. Where a is per
consequence part instead of the ordinary linear function unit frequency (actual frequency/base frequency), b per-unit
consequent part of T-S fuzzy system. In [12], incremental speed (rotor speed/synchronous speed) corresponding to base
fuzzy PID controller is used which is slightly different from frequency, C excitation capacitance per phase, E1 air-gap
the T-S fuzzy PID controller where, the scaling factors for voltage per phase referred to base frequency, R1 & R2 stator
input/output Macvicar-Whelan based fuzzy system are and rotor winding resistances per phase. RL stator load
addressed in the ant colony optimization framework. In T-S resistance per phase. V1 Load voltage per phase, X1& X2
fuzzy systems design, one challenging design task is the Stator and rotor leakage reactances per phase, Xc Excitation
determination of the parameters of the consequent part. In this capacitive reactance per phase and XL Load reactance per
paper, ASO is employed to design the consequent part of a phase.
Takagi Sugeno fuzzy PID controller. The paper presents
design procedures for a model free T-S fuzzy PID controller.
Three independent T-S fuzzy systems are combined to tune
different gains of PID controller. The optimum relationship
between the PID controller parameters and the parameters of
T-S fuzzy modules is explored. Using Ant system algorithm,
the parameters of each T-S fuzzy system are optimally
Figure 1 Normalized equivalent circuit of SESRIG‎[14]
determined. The idea is to start with a tuned, conventional PID
controller, replace it with an equivalent Takagi-Sugeno fuzzy
PID controller and eventually fine-tune the T-S fuzzy tuner. Figure 1 Normalized equivalent circuit of SESRIG‎[14]
This is relevant whenever a PID controller is possible or
For convenience, all of the machine parameters, except the
already implemented.
excitation capacitance, are expressed in per unit. The solution
In the proposed T-S Fuzzy Ant System PID, structure of T-
of the SEIG equivalent circuit have been developed using
S fuzzy system is well adapted via defining the optimum various methods. Among them, the nodal admittance method
parameters of its consequence part. Once new rules are [15], is considered very simple and easy to follow, hence it
generated, the PID controller parameters are calculated from will be adopted in this paper. The main feature of the control
the defuzzification process. This paper has two main characteristic for SESRIG used in standalone WECS is
contributions. Firstly, a PID controller has been designed for deduced in this section by performing laboratory experiments
WECS using a blending tuning mechanism namely Takagi – on 1.1 KW machine. The effect of changing the external rotor
Sugeno fuzzy Ant system algorithm. Secondly, a satisfactory resistance with constant excitation capacitance and vice versa
closed loop performance is achieved with respect to the on the output terminal voltage and frequency at loading
conventional ZN frequency response tuning method. The condition‎is‎presented.‎Fig’s. 2 &3 show the effect of changing
objective of Ant System Optimization in this paper is to the external rotor resistance with constant excitation
improve both the design efficiency of Takagi – Sugeno fuzzy capacitance at loading conditions (RL = 8 p.u) on the SESRIG
systems and its performance to get optimal PID parameters. In output terminal voltage and frequency.
this paper the voltage and frequency control of a three phase
SESRIG is investigated by varying the external rotor
resistance and the excitation capacitance using the proposed
Ant Takagi Sugeno Fuzzy PID controller.
 Figure 2 Terminal Voltage (V1) versus Rotor Speed (b) at different external
resistance (Rx) at RL= 8 p.u.

Figure 5 Terminal voltage (V1) at rated frequency as the rotor speed varies at
different excitation capacitance (RL=8)

Fig.’s‎ 6&7,‎ show the external resistance and excitation

capacitance needed to maintain nominal voltage and frequency
at a constant rotor speeds (b) of 1.1 p.u. and 1.2 p.u. as the
load current increases. For example, when rotor speed (b) is
1.1 and load current (IL) is 0.25 p.u., the values of the external
resistance and excitation capacitance which are needed to
Figure 3 Output Frequency (a) versus Rotor Speed (b) at different external
resistance (Rx) at RL= 8 p.u.
maintain the frequency and voltage at 1 p.u. are 0.7 p.u. and
16.55 F respectively.
Fig. 4, shows that as the rotor speed (b) increase the external
rotor resistance (Rx) required to maintain the frequency
constant at 1 p.u. increases. It is noted that unity per-unit
frequency is only achievable at rotor speeds above 1 p.u.

Figure 6 Variation of external resistance (Rx) to operate at rated voltage and

frequency as the load current increases
Figure. 4 External Rotor Resistance (Rx) required to maintain rated frequency
as the rotor speed varies at different excitation capacitance (RL=8)

Fig. 5 shows the corresponding terminal voltage. The

terminal voltage (V1) is constant as the rotor speed increases.
At the same rotor speed, increasing the excitation capacitance
increases the voltage. At a certain value of excitation
capacitance, unity per-unit voltage can be obtained. For
example, at rotor speed b = 1.1 p.u. and load resistance RL= 8
p.u., the external rotor resistance required to achieve a = 1 p.u.
is 0.66 p.u. To keep the voltage at 1 p.u. at the same time, the
external capacitance is adjusted to 16.11 F.
 membership is designated as 𝜇𝑖 (𝑒(𝑡)) or ( 𝜇𝑖 (𝑒(𝑡)) ).
Figure 7 Variation of excitation capacitance (C) to operate at rated voltage and Throughout this work, a Gaussians membership functions for
frequency as the load current increases the two inputs of the premise part of T-S fuzzy system are
From the above figures and discussions the following points used. The error e(t) and the rate of change of the error 𝑒(𝑡) are
can be concluded: normalized using three Gaussian membership functions;
Remark 1: negative N, zero Z, and positive P, so that nine rules constitute
(1) Comparison between the computed and the rule base for each module. For simplicity, the consequent
experimental results reflects their close matching part has been chosen to be a first order function of e(t) and
that reflects the adequateness of the circuit model 𝑒(𝑡). The rule bases have the following form:
and the solution method proposed in [16] and [17]. Rule j: If e(t) is M and 𝐞(𝐭) is L then 𝒌𝒊 = 𝒂𝒊𝒋 𝒆(𝒕) + 𝒃𝒊𝒋 𝒆(𝒕)
(2) At constant load conditions, changing the external where 𝑘𝑖 = 𝑓(𝑒(𝑡), 𝑒(𝑡)) is the gain to be tuned, i.e. Kp , Ki
rotor resistance will affect the standalone WECS or Kd . aij and bij are the constants, and j=1, 2, …9 is the rule
terminal output voltage and frequency while the number. In this paper, the weighted average technique is
excitation capacitance remains unchanged (i.e. utilized to get the overall fuzzy output as follows:
increasing rotor speed (i.e. wind speed) should Given a pair of e(t), 𝑒(𝑡 ), the final output of the fuzzy
accompany with increasing the external rotor system is inferred as follows ‎[18]:
resistance while the excitation remains nearly 𝒓
𝒋=𝟏 𝒘𝒋 (𝒆 𝒕 , 𝒆(𝒕)) 𝒂𝒊𝒋 𝒆 𝒕 + 𝒃𝒊𝒋 𝒆(𝒕)
constant) 𝒌𝒊 = 𝒓
(3) At variable load conditions, simultaneous change 𝒋=𝟏 𝒘𝒋 (𝒆 𝒕 , 𝒆(𝒕))
in the external rotor resistance and the excitation 𝑟
𝑘𝑖 = 𝑗 =1 𝑕𝑗 𝑒 𝑡 ,𝑒 𝑡 𝑎𝑖𝑗 𝑒 𝑡 + 𝑏𝑖𝑗 𝑒 𝑡 (2)
capacitance is required in order to keep the
2
standalone WECS working at the rated voltage and 𝑤𝑗 𝑒 𝑡 , 𝑒 𝑡 = 𝑗 =1 𝜇𝐴𝑖 (𝑒 𝑡 , 𝑒 𝑡 ),
𝑗
frequency (i.e. increasing the loading condition
𝒘𝒋 𝒆 𝒕 ,𝒆 𝒕
(the output power required from the standalone 𝒉𝒋 𝒆 𝒕 , 𝒆 𝒕 = 𝒓 𝒘 𝒆 𝒕 ,𝒆 𝒕
𝒋=𝟏 𝒋
WECS) to be delivered to the load) should
accompany with decreasing the external rotor Where 𝜇𝐴𝑗 (𝑒 𝑡 , 𝑒(𝑡)) is the grade of membership of
𝑖
resistance and increasing the excitation (𝑒 𝑡 , 𝑒(𝑡)) in A and B. In this paper, we assume that,
capacitance). 𝑤𝑗 𝑒 𝑡 , 𝑒 𝑡 ≥ 0 𝑓𝑜𝑟 𝑗 = 1,2, … , 𝑟
𝑟
𝑗 =1 𝑤𝑗 (𝑒 𝑡 , 𝑒 (𝑡)) > 0 for all t. Therefore we obtain
III. T – S FUZZY TUNER FOR PID CONTROLLER 𝑕𝑗 𝑒 𝑡 , 𝑒 𝑡 ≥ 0 𝑓𝑜𝑟 𝑗 = 1,2, … , 𝑟 and
𝑟
𝑗 =1 𝑕𝑗 𝑒 𝑡 , 𝑒 𝑡 =1
Many PID controllers were presented in literature, the Remark 3: Our approach is different from those of [11],
expression of PID control law for a continuous time system is [13] and [12] in the following ways.
given as follows: i. A Takagi – Sugeno type of the fuzzy PID control system
is assumed instead of incremental fuzzy PID control
𝑲𝒑 𝒕 𝒅𝒆(𝒕)
𝒖 𝒕 = 𝑲𝒑 𝒆 𝒕 + 𝟎
𝒆 𝒕 𝒅𝒕 + 𝑲𝒑 𝑻𝒅 (1) system.
𝑻𝒊 𝒅𝒕
ii. The input Gaussian membership functions of T-S fuzzy
where e(t) is the error between the input and the output of
tuner are assumed to be constant (i.e. the membership
the system; u(t) is the control action generated by the PID
function’s‎ center “c” and its spread “σ” are constant
controller; Kp is the proportional gain; Ti is the integral time
constant; and Td is the derivative time constant. over each rules and fuzzy system).
In this section, three independent T-S fuzzy systems are iii. The total no. of free parameters to be tuned in is 117
used to tune different gains of PID controller. Each fuzzy parameters while in our proposed design procedures are
module is assigned to obtain each PID gains. The typical T-S 54 parameters (i.e. smaller computational burden).
fuzzy systems studied in this paper have two inputs and one
output. The input variables are error and rate of change of the
error. We denote e(t), 𝑒(𝑡) and y(t) as error, rate of change of
error and system output, respectively. A fuzzy set for e(t) (or
𝑒(𝑡) ) is denoted as M (or L) and the corresponding
 IV. ASO ALGORITHM BASED T-S FUZZY OPTIMAL Where 𝜁𝑜 , 𝑡𝑠𝑜 𝑎𝑛𝑑 𝑒𝑠𝑠𝑜 are the performance values obtained
PARAMETER DETERMINATION from ZN tuning formula. The parameters matrices for the
problem of tuning T-S fuzzy PID controller are defined as:
A. Ant System Optimization
𝒂𝟏𝟏 ⋯ 𝒂𝟏𝟗 𝒃𝟏𝟏 ⋯ 𝒃𝟏𝟗
The ant system is the first member of a class of algorithms ⋮ ⋱ ⋮ &𝐵 = ⋮
𝑨= ⋱ ⋮ (5)
called Ant Colony Optimization (ACO) that was initially 𝒂𝟑𝟏 ⋯ 𝒂𝟑𝟗 𝒃𝟑𝟏 ⋯ 𝒃𝟑𝟗
proposed by Colorni, Dorigo and Maniezzo [19]. This
technique is adopted in this paper due to its simplicity. The Now, the control problem in this paper can be formulated
main underlying idea, loosely inspired by the behavior of real as follows:
ants, is that of a parallel search over several constructive Given a plant G(s) to be controlled (Fig.8), determine the
computational threads based on local problem data and on a optimum values of the parameters in (2) using ant system
dynamic memory structure containing information on the algorithm, hence, find the optimal PID parameters Kp, Ti and
quality of previously obtained result. In this algorithm, Td so that the control system has the minimum value of given
computational resources are allocated to a set of artificial ants performance‎criterion‎“F” (3).
that exploit a form of indirect communication mediated by the Fig. 8 depicts the closed loop control system used in this
environment to find the shortest path from the ant nest to a set paper.
target. Ant algorithms have been proved to be a quick global
T-S Fuzzy T-S Fuzzy T-S Fuzzy
optimal solution finder when compared to other heuristic Sys1 Sys2 Sys3

methods such as simulated annealing and genetic algorithms.

It also states that the ant algorithms have the quality to find
new optimal solution without reinitiating the computations
from scratch. PID Controller
Ant Sys. Fine Tuner
𝒅
B. T-S Fuzzy PID optimization problem SP
𝒅𝒕
Kd
y(t)
e(t)
Kp G(S)
Usually, the optimization process consists of finding the
𝒅𝒕 Ki
controller parameters such as to minimize or maximize a given
cost function of the closed loop system consisting of a fuzzy
PID controller and an unknown plant. The optimization of step
Figure 8 The Proposed Closed loop Control System
response of the system under control by minimizing a suitable
performance criterion is the aim of this work. A. Proposed Algorithm
Each T-S fuzzy module consists of constant premise Using the same technique used in [20] or the ant system
parameters‎ (that‎ is‎ needn’t‎ to‎ be‎ optimized)‎ and‎ 18‎ free‎ optimization, we can modify the design algorithm presented
consequence parameters (that will be optimized). The total in ‎[12] to take into account the presence of T-S fuzzy system.
number of free parameters to be optimized for the overall The proposed AS Algorithm based T-S Fuzzy optimal PID
fuzzy system is 54. The AS algorithm is applied to optimize parameter determination can be summarized as follows:
two different parameter matrices (A & B). These parameters Step 1: Determine Kpo, Kio and Kdo using the classical
matrices are driven by the consequence parts of fuzzy rule Ziegler-Nichols tuning formula for a given control system
base.‎The‎dimension‎of‎both‎parameter‎matrices‎is‎“3x9”.‎The‎ with a PID controller and compute the system's performance
effectiveness of the proposed T-S Fuzzy controller is indexes 𝜁𝑜 , 𝑡𝑠𝑜 𝑎𝑛𝑑 𝑒𝑠𝑠𝑜 .
quantified by the following performance criteria that are Step 2: Specify the initialization parameters of our proposed
problem that contain the following:
evaluated at the end of a step response experiment. It includes
the overshoot 𝜁, settling time ts and steady state error ess of the  Values of 𝛼, 𝛽, 𝜌, 𝜏𝑜 𝑎𝑛𝑑 𝑚.
system unit step response. The performance criterion of the  The maximum number of iterations tmax.
system F is designed as follows:  A one-dimensional array Pathk with 108 elements.
Array Pathk can be used to denote the moving path
𝑭 = 𝝀𝜻 𝒇𝟏 + 𝝀𝒕𝒔 𝒇𝟐 + 𝝀𝒆𝒔𝒔 𝒇𝟑 (3) of ant k.
where 𝜆𝜁 , 𝜆𝑡 𝑠 𝑎𝑛𝑑 𝜆𝑒𝑠𝑠 are three weighting coefficients and f1, f2 Step 3: Set the iteration counter t=1 and then place all of the
& f3 are defined as follows: m ants at the origin O.
𝑒𝑠𝑠 Step 4: Set i=1.
𝜁 𝑡 𝑖𝑓 𝑒𝑠𝑠𝑜 ≠ 0
𝑓1 = , 𝑓2 = 𝑠 𝑎𝑛𝑑 𝑓3 = 𝑒𝑠𝑠𝑜 , (4) Step 5: Set k=1.
𝜁𝑜 𝑡 𝑠𝑜
0 , 𝑖𝑓 𝑒𝑠𝑠𝑜 = 0
 Step 6: Compute the transition probability of each node on Where 0 < ρ < 1 is the pheromone decay parameter; 𝜏(𝑖, 𝑗) is
line Li using the following formula the amount of pheromone laid at nij by ant k in the iteration
𝜏 𝛼 𝑖, 𝑗, 𝑡 𝜂 𝛽 (𝑖, 𝑗, 𝑡) just completed and computed by the following formula:
𝑃𝑘 𝑖, 𝑗, 𝑡 = 9 𝛼 𝛽 𝑄
𝑗 =0 𝜏 𝑖, 𝑗, 𝑡 𝜂 (𝑖, 𝑗, 𝑡) ∆𝜏𝑘 𝑖, 𝑗 = 𝐹𝑘
𝑖𝑓 𝑎𝑛𝑡 𝑘 𝑝𝑎𝑠𝑠𝑒𝑑 𝑡𝑕𝑟𝑜𝑢𝑔𝑕 𝑛𝑖𝑗 𝑖𝑛 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑖𝑡𝑒𝑟𝑎𝑡.
Where 𝜂(𝑖, 𝑗, 𝑡) is the visibility of node (i, j) and defined as: 0 𝑂𝑡𝑕𝑒𝑟𝑤𝑖𝑠𝑒
10 − 𝑗 − 𝑗 ∗ Where Q is a positive constant, Fk is the value of performance
𝜂 𝑖, 𝑗, 𝑡 =
10 criterion of ant k in the iteration just completed and computed
Where the values of j* are set in the following way: by (3).
 In the first iteration of the AS algorithm, the values of Step 16: Set t←t +1. If t< tmax and all of the m ants do not
j* (i=1-108, j=0-9) are assigned from the different make the same tour, place all the ants at the origin O and go to
values of the one dimensional array Pathk-initial. Step 4; if t< tmax but all of the m ants make the same tour or t
 In each of the following iterations, j* are determined = tmax, ; stop.
from the calculated one dimensional array Pathk (it
contains different entities of matrices A & B) that is V. STANDALONE WECS SIMULATION USING THE PROPOSED
corresponding to optimal traveling path generated in TECHNIQUE
the previous iteration by the ant In this section, based on the results drawn from the above
α&β represent respectively the relative importance of the sections, the proposed technique namely Ant System based
pheromone concentration and visibility in transition Takagi Sugeno Fuzzy PID (ASTSFPID) Control is applied to
probability the standalone WECS configured in Fig. (9) to automatically
Step 7: Based on the calculated transition probability in the control the terminal output voltage and frequency of SESRIG
previous step and using Roulette wheel selection method, when either the stator load impedance or the rotor speed
select a node on line Li for ant k and move ant k to this node, changes. The goal is to keep the voltage and frequency at the
then save the y coordinate of the node into the ith element of value of 1 p.u. This is achieved by simultaneously changing
Pathk. the excitation capacitor and the added rotor resistance.
Step 8: Set k←k+1. If k ≤m, go to Step 6; Otherwise, A. Closed loop control
continue. The stator terminal voltage and the output frequency are
Step 9: Set i←i+1. If i ≤ 108, go to Step 5; Otherwise, chosen as the feedback variable since any change in speed and
continue. stator load impedance will result in a corresponding change in
Step 10: For each ant k (k=1, 2, …, m): according to Pathk, the terminal voltage and output frequency. The stator terminal
compute different entities values of matrices A & B using voltage signal is compared with the voltage reference signal.
formula (5). The voltage error signal is fed to a proportional-plus-integral
Step 11: For each ant k (k=1, 2, …, m): evaluate the PID (PI) controller that outputs a signal for the firing angle of the
gain values from the T-S fuzzy system according to formula thyristor controlled inductor (TCR). The output frequency
(2). signal is compared with the frequency reference signal. The
Step 12: For each ant k (k=1, 2, …, m): perform a simulation frequency error signal is fed to another PI controller that
experiment using the calculated matrices A & B and compute outputs a signal for the duty cycle of the chopper controlled
the system's performance indexes 𝜁 𝑘 , 𝑡𝑠𝑘 𝑎𝑛𝑑 𝑒𝑠𝑠 𝑘
then compute resistance.
the performance criterion Fk using formula (3).
Step 13: Compare all of the obtained Fk values and find the
optimal ant path of this iteration (Pathk-optimal), that has
minimum value of the performance criterion (i.e. 𝑚𝑖𝑛𝑘 Fk,
k=1,2,…,m). Save the Pathk-optimal and the corresponding
optimal PID gains.
Step 14: Set each element of Pathk to zero, k=1, 2,…, m.
Step 15: Update the pheromone concentration of each node
on the moving ant path using the following updating rule:
𝜏 𝑖, 𝑗, 𝑡 + 1 = 𝜌𝜏 𝑖, 𝑗, 𝑡 + Δ𝜏(𝑖, 𝑗)
𝑚

Δ𝜏 𝑖, 𝑗 = Δ𝜏(𝑖, 𝑗)
𝑘=1 Figure 9 Schematic Diagram for the system configuration
In [17], since no electrical dynamics is involved the SESRIG
was approximated as a first-order delayed system, the Figure 10 Dynamic Response of the voltage as the rotor speed increases from
parameters of the proposed transfer function were determined 1.2 to 1.3 p.u at RL= 4 p.u. with the proposed and the traditional control
using the open-loop step response method. Proper tuning of
the PI controller is required in order to give satisfactory
dynamic performance of the proposed standalone WECS.
Hence, same approximation for SESRIG performed in [17]
will be adopted in this paper.
Among different traditional tuning methods, the Ziegler-
Nichols (ZN) technique has been very influential. The idea is
to start the proposed algorithm with ZN PI gains then enhance
the dynamic response of the closed loop system according to
the performance index in (3) via the steps of the algorithm.
Finally, the algorithm will reach to the optimal parameters of
the PI controller that can be used in controlling the voltage and
the frequency of the standalone WECS.
Figure 11 Dynamic Response of the frequency as the rotor speed increases
B. Closed Loop Control Simulation Results from 1.2 to 1.3 p.u at RL= 4 p.u. with the proposed and the traditional control
This section contains the results of three simulation
experiments using MATLAB/SIMULINK environment to
study the dynamic response of SESRIG based standalone
WECS with Takagi Seguno Fuzzy based Ant System PI closed
loop control. For the sack of comparison, the results with ZN-
PI controller designed in [17] are also presented. Firstly the
response for changing the rotor speed (b) while the stator load
resistance kept constant is presented. Secondly the response
for changing the stator load resistance (RL) while the rotor
speed (b) kept constant is studied. Thirdly the response for
changing the rotor speed (b) followed with a change in the
stator load resistance (RL) is introduced. Fig’s.‎ (10-15) show
the dynamic response (output voltage and frequency) of the
standalone WECS subjected to the proposed and traditional
control method. The proposed PI controllers (ASTSFPI) after Figure 12 Dynamic Response of the voltage as the load increases from 4 to 4.3
a short transient time restored both the output terminal voltage p.u at b= 1.2 p.u. with the proposed and the traditional control
and frequency to the reference value (i.e. 1 p.u.) with
reasonable overshoot and zero steady state error irrespective of
changing the rotor speed and the stator load resistance. As
seen from the following figures, the proposed control method
outperforms the traditional control method.

Figure 13 Dynamic Response of the frequency as the load increases from 4 to

4.3 p.u at b= 1.2 p.u. with the proposed and the traditional control
 rotor speed and load impedance both the frequency and the
output voltage of the SESRIG can be maintained constant by
rotor resistance control and excitation capacitance control over
a wide range of speeds without exceeding the stator current
limit. The proposed algorithm is easily implemented, has a
good convergence property and has an efficient searching
ability for the optimal PID controller. Moreover the proposed
technique is very effective and useful for making the self
excited induction generator feasible for remote windy areas.

REFERENCES
[1] J.F.Manwell, J.G. McGowan, A.L. Rogers, Wind Energy Explained
Theory, Design and Application, John Wiley & sons LTD, England,
2002.
Figure 14 Dynamic response for the voltage as rotor speed & Stator load [2] A. A. Shaltout and A. F. El-Ramahi , " Maximum power tracking for a
resistance change from b = 1.15 to 1.3 p.u and from RL= 4.5 to RL= 4.3 p.u. wind driven induction generator connected to a utility network", Applied
with the proposed and the traditional control Energy, Volume 52, Issues 2-3, pp. 243-253, 1995.
[3] Badrul H. Chowdhury and Srinivas Chellapilla, " Double-fed induction
generator control for variable speed wind power generation ", Electric
Power Systems Research, Volume 76, Issues 9-10, pp. 786-800, June
2006.
[4] S.R.Vaishnav and Z.J.Khan, "Design and performance of PID and fuzzy
logic controller with smaller rule set for higher order system",
Proceedings of the World Congress on Engineering and Computer
Science 2007, WCECS 2007, San Francisco, USA October 24-26, 2007.
[5] A.‎Visioli,‎“Tuning‎PID‎controllers with fuzzy logic,”‎IEE‎Proc.‎Control‎
Theory and Applications, Vol. 148, No. 1, pp. 1-8, 2001.
[6] S.‎ Bhattacharya,‎ A.‎ Chatterjee‎ and‎ S.‎Munshi,‎ “An‎ improved PID-type
fuzzy controller employing individual fuzzy P, fuzzy I and fuzzy PID
controllers,”‎ Transaction‎ of‎ the‎ Institute‎ of‎ Measurements‎ and‎ Control‎
Vol. 25, pp. 352-372, No. 4, 2003.
[7] Lin,‎ C.T.‎ and‎ Lee,‎ C.S.G.,‎ “Neural‎ Fuzzy‎ Systems:‎ A‎ Neural-Fuzzy
Synergism‎ to‎ Intelligent‎ Systems”,‎ Englewood‎ Cliffs:‎ Prentice‎ Hall,‎
1996.
[8] Cordon,‎ O.,‎ Herrera,‎ F.,‎ Hoffmann,‎ F.‎ and‎ Magdalena,‎ L.,‎ “Genetic
fuzzy systems: evolutionary tuning and learning of fuzzy knowledge
bases”‎(Singapore:‎World‎Scientific),‎2001.
[9] A.H.Besheer,‎ “Tuning‎ method‎ for‎ PID‎ control scheme: A blended ant
system optimization technique with Takagi Sugeno fuzzy system”,‎ICFC‎
2010 - International Conference on Fuzzy Computation, Valencia, Spain
24-26 October, 2010.
[10] Karl‎O.‎Jones,‎ André‎Bouffet’‎“Comparison of ant colony optimisation
Figure 15 Dynamic response of the frequency as rotor speed change &stator and‎ differential‎ evolution”,‎ International‎ Conference‎ on‎ Computer‎
load resistance change from b= 1.15 to 1.3 p.u and from RL= 4.5 to RL= 4.3 p.u. Systems and Technologies – CompSysTech, 2007.
with the proposed and the traditional control
[11] Abdel Badie Sharkawy, "Genetic fuzzy self tuning PID controllers for
antilock braking systems", Alexandria Engineering Journal, Vol. 45, No.
6, pp. 657-673, 2006.
VI. CONCLUSION [12] Tan Guan-zheng and Dou Hong-quan,‎ “ACS‎ algorithm-based adaptive
In this paper, an optimal Takagi Seguno fuzzy PID fuzzy PID controller and its application to CIP-I‎intelligent‎leg”,‎J.Cent.‎
South Univ. Technol. (2007)04-0528-09, Springer, DOI:
controller is determined using ant system algorithm. The 10.1007/s11771-007-0103-3.
proposed controller is used to control the output voltage and [13] Hadi Nobahari, Seid H. Pourtakdoust, "Optimization of fuzzy rule bases
frequency for standalone WECS at the sending end of self- using continuous ant colony system", Proceeding of the first
excited slip-ring induction generator. The ant system International Conference on Modeling, Simulation and Applied
optimization, Sharjah, U.A.E, February 1-3, 2005.
algorithm search for the optimal parameters of the consequent
[14] M. G. Say, Alternating Current Machines, 5th ed. London, U.K.: Pitman
part of the Takagi Seguno fuzzy system that is in turn Publishers (ELBS), 1983.
produces the optimal PID gains. It is shown that with varying
[15] T.F.Chan,‎ “Analysis‎ of‎ self‎ excited‎ induction‎ generators‎ using‎ an‎
iterative‎method”,‎IEEE‎Transaction‎on‎Energy‎Conversion‎Vol.‎10,‎No.‎
3, pp. 502-507, September 1995.
[16] T. F. Chan, "Steady state analysis of self excited induction generators",
IEEE Transactions on Energy Conversion, Vol. 9. No. 2, June, 1994.
[17] T. F. Chan, K. A. Nigim, and L. L. Lai,‎“Voltage‎and‎frequency control
of self-excited slip-ring induction generators”,‎ IEEE Transactions on
Energy Conversion, vol. 19, No. 1, pp. 81-87, March 2004.
[18] A.H.Besheer‎ and‎ H.R.Emara‎ “Relaxed‎ LMI‎ based designs for Takagi
Sugeno fuzzy regulators and observers; poly-quadratic lyapunov
function‎ approach”,‎ Proceedings of the 2009 IEEE International
Conference on Systems, Man, and Cybernetics San Antonio, TX, USA –
October, 2009.
[19] DORIGO M, V. Maniezzo‎and‎A.‎Colorni,‎“Ant‎system‎optimization by
a‎colony‎of‎cooperating‎agents”,‎IEEE‎Transaction‎on‎Systems‎Man‎and‎
cybernetics – part B, 26 (1), pp. 29-41, 1996.
[20] Guanzheng Tan, Qingdong Zeng, Shengjun He and Guangchao,
“Adaptive‎ and‎ robust‎ design‎ for‎ PID‎ controller‎ based‎ on‎ ant‎ system‎
algorithm”, Advances in Natural Computation, Lecture Notes in
Computer Science, 2005, Volume 3612/2005, 915-924, DOI:
10.1007/11539902_113