STUDY OF CHAOS IN INDUCTION MACHINE DRIVE SYSTEM

B.Tech. Project
By MIRZA ABDUL WARIS BEIGH (10289) AAKASH AGRAWAL (10288) GOPAL BHARADWAJ (10265) MOHAN LAL (09223)

DEPARTMENT OF ELECTRICAL ENGINEERING NATIONAL INSTITUTE OF TECHNOLOGY, HAMIRPUR - 177005 (INDIA) December, 2013

STUDY OF CHAOS IN INDUCTION MACHINE DRIVE SYSTEM

A PROJECT Submitted in partial fulfilment of the requirements for the award for the degree Of

BACHELOR OF TECHNOLOGY By MIRZA ABDUL WARIS BEIGH (10289) AAKASH AGRAWAL (10288) GOPAL BHARADWAJ (10265) MOHAN LAL (09223) Under The guidance Of Dr. Bharat Bhushan Sharma

DEPARTMENT OF ELECTRICAL ENGINEERING NATIONAL INSTITUTE OF TECHNOLOGY, HAMIRPUR - 177005 (INDIA) December, 2013

Copyright NIT HAMIRPUR, 2013

CANDIDATES DECLARARTION
We hereby certify that the work which is being presented in the project report entitled STUDY OF CHAOS IN INDUCTION MACHINE DRIVE SYSTEM, in partial fulfillment of the requirements for the award of degree of the Bachelor of Technology and submitted in the Department of Electrical Engineering, National Institute of Technology, Hamirpur H.P. is an authentic record of our own work carried out during a period from August 2013 to December 2013 under the supervision of Dr. Bharat Bhushan Sharma, Assistant Professor, Department of Electrical Engineering, N.I.T. Hamirpur. The matter presented in this project report has not been submitted by us for the award of any other degree of this or any other university/institute.
Sd/MIRZA ABDUL WARIS BEIGH (10289) AAKASH AGRAWAL (10288) GOPAL BHARADWAJ (10265) MOHAN LAL (09223)

This is to certify that above statement made by the candidate is correct to the best of my knowledge.

Date:

Sd/Dr. Bharat Bhushan Sharma Assistant Professor, EED

The project Viva-Voce Examination of the Candidates Mirza Abdul Waris Beigh (10289), Aakash Agrawal (10288), Gopal Bharadwaj (10265)Mohan Lal (09223) has been held on____________________.

Dr. Bharat Bhushan Sharma

Project Supervisor Electrical Engg. Dept.

---------------------------------External Examiner

ACKNOWLEDGEMENT
First things first we find it hard to express my gratefulness to Almighty GOD in words for bestowing upon us His deepest blessings and providing us with the most wonderful opportunity in the form of life of a human being and for the warmth and kindness he has showered upon us. We feel great pleasure in acknowledging our deepest gratitude to our revered guide and mentor, Dr. Bharat Bhushan Sharma, Assistant Professor, Electrical Engineering Department, National Institute of Technology Hamirpur, under whose firm guidance, motivation and vigilant supervision we succeeded in completing our work. He infused into us the enthusiasm to work on this topic. His tolerant nature accepted our shortcomings and he synergized his impeccable knowledge with our curiosity to learn into this fruitful result. We would sincerely thank Dr. Ravinder Nath HOD, Electrical Engineering Department who suggested many related points and is always very helpful and constructive. Words are inadequate to express our heartfelt gratitude to our affectionate parents who have shown so much confidence in us and by whose efforts and blessings we have reached here. We would also like to thank all the faculty members of Department of Electrical Engineering for their continuous moral support and encouragement. Last but not the least we wish to express heartiest thanks to our friends and colleagues for their support, love and inspiration.

Date: MIRZA ABDUL WARIS BEIGH (10289) AAKASH AGRAWAL (10288) GOPAL BHARADWAJ (10265) MOHAN LAL (09223)

Abstract
Our work brings attention to the nonlinear dynamics of an induction motor's drive system with indirect field controlled. To understand the complex dynamics of system, some basic dynamical properties, such as equilibrium, stability are rigorously derived and studied. Chaotic attractors are first numerically verified through investigating phase trajectories, Hopf bifurcation, and Lyapunov Exponents. Furthermore, a new sliding mode control method is studied to gain the synchronization with different initial values. It can control the system to an equilibrium point. After the control of chaos in the system, further variation in parameters is carried out to check for the events where chaos can creep into the system again. This is verified using the Lyapunov exponents and the Phase Plots. Numerical simulations are presented to demonstrate the effectiveness of the proposed controllers. Keywords: Induction motor, Chaos, Chaos control, Synchronization

Contents
1 Introduction 1 1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...........................1 1.1.1 Chaos in electric drives1 1.2 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Organization of the report . . . . . . . . . . . . . . . . . . . . .. . 2 2. Non Linear Dynamical model of Induction Motor................................ 1 3. Motor Model and Problem Formulation................................................. 2 3.1 Derivation of State Space Form... 3.2 Solution of Equations... 4. Bifurcations & Phase Plot analysis......................................................... 3 4.1 Phase plots........................................................................................... 3 4.2 Hopf Bifurcations.................................................................................4 5. Lyapunov Exponents................................................................................5 5.1 Introduction to Lyapunov exponent.................................................... 5 5.2 Simulation results................................................................................ 7 6. Sliding Mode Technique ......................................................................... 7 6.1 Introduction..........................................................................................7 6.2 Control Scheme...................................................................................8 8. Simulations, Analysis and Result............................................................ 8.1 Analysis and Solution of Controller.. 8.2 Results after parameter variation 8.2.1 Phase Plot and Bifurcations... 8.2.2 Lyapunov Exponent Plot.. 9. Conclusion.................................................................................................. 13 Bio-Data of Candidates.. Appendix... References....................................................................................................

List of figures

List of Abbreviations

List of symbols

Introduction 1.1 Overview The Oxford English Dictionary defines Chaos as Behaviour of a system which is governed by deterministic laws but is so unpredictable as to appear random, owing to its extreme sensitivity to changes in parameters or its dependence on a large number of independent variables; a state characterized by such behaviour. The general perception on chaos is equivalent to disorder or even random. It should be noted that chaos is not exactly disordered, and its random-like behaviour is governed by a rule - mathematically, a deterministic model or equation that contains no element of chance. Actually, the disorder-like or random-like behaviour of chaos is due to its high sensitivity on initial conditions. Similar to many terms in science, there is no standard definition of chaos. Nevertheless chaos has some typical features: Nonlinearity: Chaos cannot occur in a linear system. Nonlinearity is a necessary, but not sufficient condition for the occurrence of chaos. Essentially, all realistic systems exhibit certain degree of nonlinearity. Determinism: Chaos must follow one or more deterministic equations that do not contain any random factors. The system states of past, present and future are controlled by deterministic, rather than probabilistic, underlying rules. Practically, the boundary between deterministic and probabilistic systems may not be so clear since a seemingly random process might involve deterministic underlying rules yet to be found. Sensitive dependence on initial conditions: A small change in the initial state of the system can lead to extremely different behaviour in its final state. Thus, the long-term prediction of system behaviour is impossible, even though it is governed by deterministic underlying rules. Aperiodicity: Chaotic orbits are aperiodic, but not all aperiodic orbits are chaotic. Almost-periodic and quasi-periodic orbits are aperiodic, but not chaotic.

1.1.1 Chaos in Electric Drive Systems The investigation of chaos in electric drive systems can be categorized as three themes, namely the analysis of chaotic phenomena, the control of chaotic behaviors, and the application of chaotic characteristics. Chaos in electric drive systems was firstly identified in induction drive systems in 1989. That is, the bifurcation of induction motor drives was studied (Kuroe and Hayashi, 1989), which was actually an extension of the instability analysis of pulsewidth-modulation (PWM) inverter systems. The bifurcation and chaos resulting from the tolerance-band PWM inverter-fed induction drive system was then investigated (Nagy, 1994; Suto, Nagy, and Masada, 2000). It was also identified that saddle-node bifurcation, or even Hopf bifurcation, might occur in induction drive systems under

indirect field oriented control (Bazanella and Reginatto, 2000) and, consequently, the control of chaos in induction drive systems was investigated. An attempt was made to use a neural network stabilizing chaos during speed control of induction drive systems (Asakura et al., 2000). On the other hand, an attempt was made to use periodic speed command to stimulate the chaotic motion of induction drive systems (Gao and Chau, 2003a). Without taking power electronic switching into consideration, it was identified that the permanent magnet (PM) brushless DC drive system could be transformed into a Lorenz system, which is well known to exhibit a Hopf bifurcation and chaotic behaviour (Hemati, 1994) The application of chaos in electric drive systems has focused on the practical use of the control of chaos, including the stabilization of chaos and the stimulation of chaos. For instance, chaotic vibration in an automotive wiper system not only decreases the wiping efficiency but also causes harmful distraction to the drivers (Suzuki and Yasuda, 1998). Thus, the corresponding chaos was directly stabilized by applying an extended time-delay auto-synchronization control to its DC drive (Wang and Chau, 2006; Wang and Chau, 2009a). This approach can be realized experimentally because the armature current of the DC motor can be easily measured by a Hall sensor and the perturbations on the feed-in motor voltage can be readily produced by a power converter. To control the undesirable chaos in the permanent magnet synchronous motor (PMSM), an adaptive dynamic surface control law was designed by Wei and his co-partners. However, there are few contributions to a current-driven induction motor, especially, the dynamical model for a whole induction motor system with indirect field controlled. While, it is a main drive device in modern industry, and its nonlinear vibration is catholic. Therefore, it is necessary to study the intrinsic quality of its nonlinear vibration via nonlinear dynamics theory. 1.2 Objective Chaos control is inquisitive in how to control the chaotic system to the periodic orbit or equilibrium point with the original parameters remained or only _ne-tuned, because the system parameters cannot be changed objectively, or the parameters change largely must pay a great price. Typical control methods have been proposed to achieve chaos control. For instance, two methods of chaos control with a small time continuous perturbation were proposed by Pyragas [16]. Ataei et al. [17] presented a chaos synchronization method for a class of uncertain chaotic systems using the combination of an optimal control theory and an adaptive strategy. Wang and his coworkers [18] used symbolic dynamics and the automaton reset sequence to identify the current drive word and obtained the synchronization. Nonlinear and linear feedback controllers were designed to control and synchronize the chaotic system by Rafikov et al. [19]. Golovin et al. [20] proposed a global feedback control method based on measuring the maximum of the pattern amplitude over the domain, which can stabilize the system. Based on OGY approach, a multiparameter semi-continuous method was designed to control chaotic behaviour by de Paula and Savi [21]. The united chaotic systems with uncertain parameters were synchronized based on the CLF method by Wang et al. [22]. Ataei et al. [23] presented a chaos synchronization method for a class of uncertain chaotic systems using the combination of an optimal control theory and an adaptive strategy. Among the control methods, sliding mode technique (SMT) is one of the best methods. Recently, many contributions have been published (see, for example, [24-28]). To our best knowledge, there is little

information about control method, which could bridge the chaos control and synchronization from the literature. And, it is a very valuable theory for its stable and synchronous operation with the power system. Considering all the above discussion, there are several advantages which make our approach attractive, compared with prior works. First, the nonlinear dynamical model for a whole induction motor system with indirect field controlled is proposed, and the nonlinear dynamics behaviors of the system model are analysed including bifurcation diagrams, phase plots. Moreover, we present a sliding mode control method. And the control method is effective to the chaos control and synchronization. Numerical simulations are demonstrated to the effectiveness of the proposed scheme. 1.3 Organisation of the Report This work is organised as follows. In Chapter 2 we present the nonlinear dynamical model of current-driven induction motor expressed in a reference frame rotating at synchronous speed. Chapter 3 discusses the nonlinear dynamical system and the problem formulation and its numerical results. Chapter 4 presents the Hopf Bifurcations and the phase plots of the model presented above. Chapter 5 introduces the Sliding Mode Technique (SMT) and its method of control. In Chapter 6 a sliding mode controller is presented. Chapter 7 discusses about the Lyapunov exponents and the various Lyapunov plots obtained for the system. In Chapter 8 we present the overall analysis and the results. Finally conclusion and future scope is discussed in Chapter 10.

Charter 2 Model of induction Motor

Chapter-3

Motor Model and Problem formulation

3.1 Derivation of State Space Form The nonlinear dynamical model of a current-driven induction motor expressed in a reference frame rotating at synchronous speed is given as follows:

where Rr is rotor resistance, Lr is rotor self-inductance, Lm is mutual inductance in a rotating reference frame, np is the number of pole pairs, sl is slipping frequency, J is inertia coefficient, TL is load, qr is quadrature axis component of the rotor flux, dr is direct axis component of the rotor flux, r is rotor angular speed, Rr is rotating resistance coefficient. The parameters are introduced as follows:

(1)
Therefore, the nonlinear dynamical model of induction motor system with indirect field controlled can be rewritten as follows:

(2)

In speed regulation applications, the indirect field oriented control is usually applied with a proportional integral (PI) speed loop, and this control strategy is described as follows:

(3)

Where c^1 is the estimate for the inverse rotor time constant c1, is the constant reference velocity, u02 is the constant reference for the rotor flux magnitude, Kp is the

proportional of the PI speed regulator, Ki is the integral gains of the PI speed regulator. The rotor time constant varies widely in practice IFOC system of IM. One sets c^1 = c1. That is to say, if it has a perfect estimate of the rotor time constant, the control is tuned; otherwise it is said to be detuned. Therefore, the degree of tuning is set to k = c^1/c1. Obviously, the controller is tuned and one sets k = 1. Let x3 =ref - r and x4 = u3, and thus a new fourth dimensional system can be written as follows, based on the model of the whole closed-loop system (2) and the control strategy (3).

(4)

3.2 Solution of the equations. The equilibria of system (4) can be found by solving the following algebraic equations:

where, c1 = 13:67, c2 = 1:56, c3 = 0:59, c4 = 1176, c5 = 2:86, u02= 4, kp = 0:001, ki = 1, k = 1:5, TL = 0:5, ref = 181:1 and the initial state is set to x1 = 0, x2 = 0:4, x3 = -200, x4 = 6. The system has three equilibria, which are respectively described as follows: O(-0:017; 0:455; 0; 0:304), E+(-0.022 0.182 *i; 0.184 + 0.021*i, 0,0.187 - 3:981 * i), E+(-0.022 + 0.182 *i; 0.184 - 0.021*i, 0,0.187 + 3:981 * i), The system has a unique equilibrium O(-0:017; 0:455; 0; 0:304). Linearize the system at O, and the Jacobian matrix is obtained as follows:

For gaining its eigenvalues, we have |I-J0| These eigenvalues at equilibrium O are respectively obtained as follows: 1= 1.65 + 40.39i; 2= 1.65 + 40.39i; 3= -18.98; 4= -13.78 1, 2 are complex conjugate pair and their real parts are positive, and 3 and 4 are negative real numbers. Therefore, the equilibrium O is a saddle point. It is unstable. The other two equilibrium points E+ and E do not belong to the real space. Thus, it is not necessary to discuss stability of these points. The theory of dissipative systems is a basic tool to describe the system characteristics. And dissipative analysis of system (4) is presented as follows. For system (4), it is noticed that

Obviously, system (4) can have dissipative structure, with an exponential contraction rate: That is, a volume element V0 is contracted by the flow into a volume element V0 e- (2c1 + c3 + kpc4c5x2)t in time t. This means that each volume containing the system orbit shrinks to zero as t at an exponential rate - (2c1 + c3 + kpc4c5x2). Therefore, all system orbits are ultimately confined to some subset of zero volume, and the asymptotic motion settles on some attractors.

Chapter-4

Bifurcations & Phase plot analysis

4.1 Phase Plots The behaviour of the system can be analysed by observing phase plots, Hopf bifurcation plots and calculating the value of the maximum Lyapunov exponent. The detailed description of each of the following is given below. A phase space is a space in which all possible states of a system are represented, with each possible state of the system corresponding to one unique point in the phase space. The concept of phase space was developed in the late 19th century by Ludwig Boltzmann, Henri Poincare, and Willard Gibbs. A plot of position and momentum variables as a function of time is sometimes called a phase plot or a phase diagram. The parameters of the motor are listed as c1 = 13.67, c2 = 1.56, c3 = 0.59, c4 = 1176, c5 = 2.86, the parameters of the system are given u02= 4, ref = 181.1 rad/s, TL =0.5, kp = 0.001, ki = 0.5, and k = 1.5. Phase portraits of chaotic system shown here illustrate the existence of only one-wing. In the below figures the phase portrait are plotted between various states for the various parameters listed above. We observe that the states are doesnt have any fixed equilibrium points and are settled on certain attractors known as strange attractors. Therefore phase portrait serves as one of the approaches to judge the systems chaotic

Fig(a). Phase plot between X1,X2 and X3.

Fig(b). Phase plot between X2 and X4.

Fig(c) . Phase plot between X2 and X3.

Fig(d). Phase plot between X1 and X2

Fig(e). Phase plot between X1,X2 and X4

Fig(f). Phase plot between X2,X3 and X4

4.2 Hopf Bifurcation The appearance or the disappearance of a periodic orbit through a local change in the stability properties of a steady point is known as the Hopf bifurcation. The following theorem works with steady points with one pair of conjugate nonzero purely imaginary eigenvalues. It tells the conditions under which this bifurcation phenomenon occurs. The term Hopf bifurcation (also sometimes called Poincare-Andronov-Hopf bifurcation) refers to the local birth or death of a periodic solution (self-excited oscillation) from an equilibrium as a parameter crosses a critical value. A bifurcation diagram summarizes the essential dynamics of a system, and thus is a useful tool to observe its nonlinear dynamical response. Under the occurrence of Hopf bifurcation, the dynamical system may demonstrate a complicated behavior that is, chaos. To further identify the chaotic behavior, the calculation of Lyapunov exponents plays an important role The bifurcation diagram is shown below for different load TL with other parameters kept constant.

Fig(a). Hopf-Bifurcation plot between T and X1

Fig(a). Hopf-Bifurcation plot between T and X2

Fig(a). Hopf-Bifurcation plot between T and X3

Fig(a). Hopf-Bifurcation plot between T and X2

It can clearly be seen that there exist a bifurcation in the values of state variables when the range of load is (-4,4).

Chapter 5

Lyapunov exponents
5.1 Introduction to Lyapunov exponent If all points in a neighbourhood of a trajectory converge toward the same orbit, the attractor is a fixed point or a limit cycle. However, if the attractor is strange, any two trajectories x(t) = f t(x0) and x(t) + x(t) = f t(x0 + x0) that start out very close to each other separate exponentially with time, and in a finite time their separation. attains the size of the accessible state space. This sensitivity to initial conditions can be quantified as:

where , the mean rate of separation of trajectories of the system, is called the leading Lyapunov exponent. In the limit of infinite time the Lyapunov exponent is a global measure of the rate at which nearby trajectories diverge, averaged over the strange attractor.

The mean growth rate of the distance

between neighbouring trajectories is

given by the leading Lyapunov exponent which can be estimated for long (but not too long) time t as

(1) For notational brevity we shall often suppress the dependence of quantities such as = (x0, t), x(t)= x(x0, t) on the initial point x0. One can use (1) as is, take a small initial separation x0, track the distance between two nearby trajectories until ||x(t1)|| gets significantly big, then record t11 = ln(||x(t1)|| / ||x0||), rescale x(t1) by factor x0/x(t1), and continue ad infinitum, with the leading Lyapunov exponent given by

(2)

Deciding what is a safe linear range, the distance beyond which the separation vector x(t) should be rescaled, is a dark art.
We can start out with a small x and try to estimate the leading Lyapunov exponent from (2).

5.2 Simulation Result To further identify the chaotic behaviour, the calculation of Lyapunov exponents has been done and the results are plotted below. The lyapunov exponent have been plotted for load TL= 0.5 and TL = 8.5 , it can clearly be observed that the highest lyapunov exponent is positive for Tl= 0.5 which indicates the systems chaotic behaviour .

Fig(a). Dynamics of lyapunov exponent for T= 0.5

Fig(a). Dynamics of lyapunov exponent for T= 0.5

Chapter-6

Sliding Mode Technique

6.1 Introduction: In control theory, sliding mode control, or SMC, is a nonlinear control method that alters the dynamics of a nonlinear system by application of a discontinuous control signal that forces the system to "slide" along a cross-section of the system's normal behaviour. The state-feedback control law is not a continuous function of time. Instead, it can switch from one continuous structure to another based on the current position in the state space. Hence, sliding mode control is a variable structure control method. The multiple control structures are designed so that trajectories always move toward an adjacent region with a different control structure, and so the ultimate trajectory will not exist entirely within one control structure. Instead, it will slide along the boundaries of the control structures. The motion of the system as it slides along these boundaries is called a sliding mode[1] and the geometrical locus consisting of the boundaries is called the sliding (hyper)surface. In the context of modern control theory, any variable structure system, like a system under SMC, may be viewed as a special case of a hybrid dynamical system as the system both flows through a continuous state space but also moves through different discrete control modes.

Figure shows an example trajectory of a system under sliding mode control. The sliding surface is described by , and the sliding mode along the surface commences after the finite time when system trajectories have reached the surface. In the theoretical description of sliding modes, the system stays confined to the sliding surface and need only be viewed as sliding along the surface. However, real implementations of sliding mode control approximate this theoretical behaviour with a high-frequency and generally nondeterministic switching control signal that causes the system to "chatter" in a tight neighbourhood of the sliding surface. This chattering behaviour is evident in Figure 1, which chatters along the surface as the system asymptotically approaches the origin, which is an asymptotically stable equilibrium of the system when confined to the sliding surface.

Intuitively, sliding mode control uses practically infinite gain to force the trajectories of a dynamic system to slide along the restricted sliding mode subspace. Trajectories from this reduced-order sliding mode have desirable properties (e.g., the system naturally slides along it until it comes to rest at a desired equilibrium). The main strength of sliding mode control is its robustness. Because the control can be as simple as a switching between two states (e.g., "on"/"off" or "forward"/"reverse"), it need not be precise and will not be sensitive to parameter variations that enter into the control channel. Additionally, because the control law is not a continuous function, the sliding mode can be reached in finite time (i.e., better than asymptotic behaviour). Under certain common conditions, optimality requires the use of bangbang control; hence, sliding mode control describes the optimal controller for a broad set of dynamic systems.

6.2 Control Scheme: In sliding-mode control, the designer knows that the system behaves desirably (e.g., it has a stable equilibrium) provided that it is constrained to a subspace of its configuration space. Sliding mode control forces the system trajectories into this subspace and then holds them there so that they slide along it. This reduced-order subspace is referred to as a sliding (hyper)surface, and when closed-loop feedback forces trajectories to slide along it, it is referred to as a sliding mode of the closed-loop system. Trajectories along this subspace can be likened to trajectories along eigenvectors (i.e., modes) of LTI systems; however, the sliding mode is enforced by creasing the vector field with high-gain feedback. Like a marble rolling along a crack, trajectories are confined to the sliding mode. The sliding-mode control scheme involves 1. Selection of a (hyper) surface or a manifold (i.e., the sliding surface) such that the system trajectory exhibits desirable behaviour when confined to this manifold. 2. Finding feedback gains so that the system trajectory intersects and stays on the manifold. Because sliding mode control laws are not continuous, it has the ability to drive trajectories to the sliding mode in finite time (i.e., stability of the sliding surface is better than asymptotic). The sliding-mode designer picks a switching function :Rn Rm that represents a kind of "distance" that the states x are away from a sliding surface.

A state x that is outside of this sliding surface has (x)0 A state that is on this sliding surface has (x)=0

The sliding-mode-control law switches from one state to another based on the sign of this distance. So the sliding-mode control acts like a stiff pressure always pushing in the direction of the sliding mode where (x)=0. Desirable x(t) trajectories will approach the sliding surface, and because the control law is not continuous (i.e., it switches from one state to another as trajectories move across this surface), the surface

is reached in finite time. Once a trajectory reaches the surface, it will slide along it and may, for example, move toward the x=0 origin. So the switching function is like a topographic map with a contour of constant height along which trajectories are forced to move. The sliding (hyper) surface is of dimension n x m where n is the number of states in x and m is the number of input signals (i.e., control signals) in u. For each control index 1<k<m, there is an n x 1 sliding surface given by

The vital part of SMC design is to choose a control law so that the sliding mode (i.e., this surface given (x) = 0 by exists and is reachable along system trajectories. The principle of sliding mode control is to forcibly constrain the system, by suitable control strategy, to stay on the sliding surface on which the system will exhibit desirable features. When the system is constrained by the sliding control to stay on the sliding surface, the system dynamics are governed by reduced-order system obtained from above equation. To force the system states to satisfy (x) = 0, one must: 1. Ensure that the system is capable of reaching (x) = 0 from any initial condition 2. Having reached (x) = 0, the control action is capable of maintaining the system at (x) = 0.

Chapter-7

Controller Design
Consider the drive system Dx = Ax + g(x) (5) where x(t)= R4 denotes the state vector of the 4-dimensional system, A = R4*4 represents the linear part of the system and g : R4 R4 is the nonlinear part of the system. Considering y(t)= R4 as the response of state vector of the 4-dimensional system, we can rewrite the response system as Dy=Ay+g(y) The controller u(t) = R4 is added to system (6), so it can be rewritten as: Dy = Ay + g(y) + u(t) (7) (6)

Here, we define the synchronization errors e = y- x. The aim is to choose a suitable controller u(t) = R4 such that the states of the master and slave systems can reach synchronization (i.e., lim||e|| = 0, where||.|| is the Euclidean norm). Now, one sets the controller u(t) as u(t) = u1(t) + u2(t) (8)

where u1(t)= R4 is a compensation controller, and u1(t) = Dx-A(x)-g(x). u2(t)= R4 is a vector function, and will be designed later. Using (8), response system (7) can be rewritten as De(t) = Ae + g(y)- g(x) + u2(t) (9)

In accordance with the procedure of designing active controller, the nonlinear part of the error dynamics is eliminated by the following the following input vector: u2(t)= g(x) - g(y) + Kw(t) Error system (9) is then rewritten as follows De(t)=Ae + Kw(t) (11) (10)

where K = [k1; k2; k3; k4]T is a constant gain vector and w(t) = R is the control input that satisfies

(12)

As a choice for the sliding surface, we have

s(t)=Ce (13) where C = [c1; c2; c3; c4]T is a constant vector. For sliding mode method, the sliding surface and its derivative must satisfy the following conditions. s(t)=0; s.(t) = 0 One sets: s.(t)=CD,e(t)=C(Ae+Kw(t))=0 (15) (14)

To satisfy the above condition, the discontinuous reaching law is chosen as follows Ds(t) = -psign(s) - rs Where p > 0, r > 0 are the gains of the controller. Considering (15) and (16), we have w(t) = -(CK)-1 [C(rI + A)e + psign(s)] Now, the total control law can be de_ned as follows u(t) = Dx - Ax - g(y) - K(CK)-1 [C(rI + A)e + psign(s)] Using (19) and (9), the error dynamics can be obtained De = [A - K(CK)-1C(rI + A)]e - K(CK)-1psign(s) (20) (19) (16)

(18)

For the sliding term, a linear system is a bounded input (-K(CK)-1p, when s > 0 and K(CK)-1p, when s < 0). The system (20) is stable, if |arg(eig([A - K(CK)-1C(rI +A)]))| > /2. It can be shown that choosing appropriate K, C and r can make the error dynamics stable. Hence, the synchronization is realized. Similarly, if the drive system (5) is modified as Dx=0 (21) Thus, the response system can be controlled to the initial values of drive system. If the initial values are changed, the controlling to any stable point can be achieved.

Chapter-8 Simulations, Analysis and Results 8.1 Analysis and Solution of Controller The numerical simulation results are carried out to verify the applicability and effectiveness of the proposed sliding mode control method. It should be noticed that the controller is in action at t = 10. The ode45 solver of MATLAB. Software is applied to solve different equations. By taking the parameters as these in Section 3, system (4) can be rewritten as:

According to 4.1, we get

Let system (22) with initial conditions [xd1; xd2; xd3; xd4]T = [0; 0:4;-200; 6]T as a drive system, and system (22) with initial values [xr1; xr2; xr3; xr4]T = [0:3; 0:5; 0:2; 0:4]T as a response system. The parameters of the controller are set as K = [-2;-6;-2;2]T,C = [5; 5; 5; 5], r = 5, and p = 0:2. This selection of parameters results in eigenvalues (1;2; 3; 4) = (-2247:3;-14:072;-5;-2:1495) which are located in the stable region. According to (19), the control signals are obtained as

Where e1 = xr - xd, e2 = yr - yd, e3 = zr - zd, e4 = wr - wd. The numerical simulation results are given in Figure 5. One can see, the errors converge to zero immediately after the controller was applied, which implies that the chaos synchronization between the two systems is realized. Keep the parameters of the controller fixed, while set the drive system as system (21) to investigate the effectiveness of the controller. And we still use system (23) as the controller. Fortunately, Figure 6 illustrates the response states, which show that the response states follow initial values of the drive system immediately.

Fig. (a)

Fig. (b)

Fig. (c)

Fig. (d) The state variables of the response system in the presence of controller (the controller u(t) is activated at t = 10)

8.2 Results after parameter variation. 8.2.1 Phase Plot and Bifurcations The phase portrait analysis was done again with certain change in parameters which are given below. The parameters of the motor are listed as c1 = 136.7s-1, c2 = 15.6H.s-1, c3 = 0.59s-1, c4 = 1176kg-1 _m-2 and c5 = 28.6, the parameters of the system are given u02= 4A, wref = 181:1rad/s, T =8.5; kp = 0:001, ki = 0:5, and k = 1:5 The phase plots are shown below, these phase plots gives us idea about the behavior of the system.

Fig(d). Phase plot between X1 and X4

Fig(d). Phase plot between X1 and X2

Fig(d). Phase plot between X1,X2 and X3

Fig(d). Phase plot between X2 and X4

With the value of load TL > 5 , it was observed that the chaos in the systems reappears if the value of rotor inductance falls by a certain ratio r. To analyse the systems behaviour in accordance to the inductance of rotor windings , bifurcation diagram is plotted below by varying the values of r.

Fig(a). Hopf-Bifurcation plot between r and X1

Fig(a). Hopf-Bifurcation plot between r and X2

Fig(a). Hopf-Bifurcation plot between r and X3

Fig(a). Hopf-Bifurcation plot between r and X4

It is clear from the above figures that the chaos occurs for the values of r within range (5, 15). 8.2.2 Lyapunov Exponent Plot The chaotic behavior of the system due to the decrease in the inductance value is confirmed by plotting the Lyapunov exponents.

Fig(a). Dynamics of lyapunov exponent for r= 10

Fig(a). Dynamics of lyapunov exponent for r= 20

APPENDIX
1)Code used for plotting phase portrait.
function [ dx ] = phaseplots( t,x) %This function is used for plotting the phase portrait for the given %system. % The function is called using ode45 command and following conditions % were taken into account. % tspan = 0:0.1:100; % options = []; % initvalue = [ 0 .4 -200 6 ]; % The values of the various constant used are given below. c1 = 13.67*10; c2 = 1.56*10; c3 = 0.59; c4 = 1176; c5 = 2.86*10; u= 4; kp = 0.001; k1 = 1; k = 1.5;W=181.1; T= 8.5; dx=zeros(4,1); dx(1)= -c1*x(1)+c2*x(4)- (k*c1/u)*x(2)*x(4); dx(2)= -c1*x(2) + c2*u + (k*c1/u)*x(1)*x(4) ; dx(3) = -c3*x(3)-c4*(c5*(x(2)*x(4)-x(1)*u)-T-(c3/c4)*W) ; dx(4) = (k1-kp*c3)*x(3)-kp*c4*(c5*(x(2)*x(4)-x(1)*u)-T(c3/c4)*W) ; end

2) Code for plotting hopf bifurcation

function [ dx ] = chaos_bifurcation( t,x,r) %This function is used for plotting the hopf-bifurcation for the given %system. % The values of the various constant used are given below. % r is that ratio by which the value of the inducatance has been reduceed. c1 = 13.67*r; c2 = 1.56*r; c3 = 0.59; c4 = 1176;c5 = 2.86*r ; u= 4; kp = 0.001; k1 = 1; k = 1.5;W=181.1; T= 8.5;

x1 = 0; x2 = 0.4; x3 = -200; x4 = 6; dx=zeros(4,1); dx(1)= -c1*x(1)+c2*x(4)- (k*c1/u)*x(2)*x(4); dx(2)= -c1*x(2) + c2*u + (k*c1/u)*x(1)*x(4) ; dx(3) = -c3*x(3)-c4*(c5*(x(2)*x(4)-x(1)*u)-T-(c3/c4)*W) ; dx(4) = (k1-kp*c3)*x(3)-kp*c4*(c5*(x(2)*x(4)-x(1)*u)-T(c3/c4)*W) ; end

2.1) Calling the bifurcation function

% Specifying the initail conditions required for running the ODE function. tspan=0:0.01:10; initvalue=[0 0.4 -200 6]; option=[]; figure % recursively calling the ODE expression for different values of 'r'. for r= 1:.2:30 [t,x]=ode45(@chaos_bifurcation,tspan,initvalue,option,r); hold on; plot(r,x(j,4),'marker','+') xlabel('r') ylabel('X4') end

3) Code for plotting the lyapunov exponents. 3.1 Main lyapunov function
function [Texp,Lexp]=lyapunov(n,rhs_ext_fcn,fcn_integrator,tstart, stept,tend,ystart,ioutp); % % Lyapunov exponent calcullation for ODE-system. % % The alogrithm employed in this m-file for determining Lyapunov

% exponents was proposed in % % A. Wolf, J. B. Swift, H. L. Swinney, and J. A. Vastano, % "Determining Lyapunov Exponents from a Time Series," Physica D, % Vol. 16, pp. 285-317, 1985. % % For integrating ODE system can be used any MATLAB ODE-suite methods. % This function is a part of MATDS program - toolbox for dynamical system investigation % See: http://www.math.rsu.ru/mexmat/kvm/matds/ % % Input parameters: % n - number of equation % rhs_ext_fcn - handle of function with right hand side of extended ODE-system. % This function must include RHS of ODEsystem coupled with % variational equation (n items of linearized systems, see Example). % fcn_integrator - handle of ODE integrator function, for example: @ode45 % tstart - start values of independent value (time t) % stept - step on t-variable for Gram-Schmidt renormalization procedure. % tend - finish value of time % ystart - start point of trajectory of ODE system. % ioutp - step of print to MATLAB main window. ioutp==0 - no print, % if ioutp>0 then each ioutp-th point will be print. % % Output parameters: % Texp - time values % Lexp - Lyapunov exponents to each time value. % % Users have to write their own ODE functions for their specified % systems and use handle of this function as rhs_ext_fcn - parameter. % % Example. Lorenz system: % dx/dt = sigma*(y - x) = f1 % dy/dt = r*x - y - x*z = f2 % dz/dt = x*y - b*z = f3 % % The Jacobian of system: % | -sigma sigma 0 |

% J = | r-z -1 -x | % | y x -b | % % Then, the variational equation has a form: % % F = J*Y % where Y is a square matrix with the same dimension as J. % Corresponding m-file: % function f=lorenz_ext(t,X) % SIGMA = 10; R = 28; BETA = 8/3; % x=X(1); y=X(2); z=X(3); % % Y= [X(4), X(7), X(10); % X(5), X(8), X(11); % X(6), X(9), X(12)]; % f=zeros(9,1); % f(1)=SIGMA*(y-x); f(2)=-x*z+R*x-y; f(3)=x*yBETA*z; % % Jac=[-SIGMA,SIGMA,0; R-z,-1,-x; y, x,-BETA]; % % f(4:12)=Jac*Y; % % Run Lyapunov exponent calculation: % % [T,Res]=lyapunov(3,@lorenz_ext,@ode45,0,0.5,200,[0 1 0],10); % % See files: lorenz_ext, run_lyap. % % ------------------------------------------------------------------% Copyright (C) 2004, Govorukhin V.N. % This file is intended for use with MATLAB and was produced for MATDS-program % http://www.math.rsu.ru/mexmat/kvm/matds/ % lyapunov.m is free software. lyapunov.m is distributed in the hope that it % will be useful, but WITHOUT ANY WARRANTY. % n=number of nonlinear odes % n2=n*(n+1)=total number of odes % n1=n; n2=n1*(n1+1); % Number of steps

nit = round((tend-tstart)/stept); % Memory allocation

y=zeros(n2,1); cum=zeros(n1,1); y0=y; gsc=cum; znorm=cum; % Initial values y(1:n)=ystart(:); for i=1:n1 y((n1+1)*i)=1.0; end; t=tstart; % Main loop for ITERLYAP=1:nit % Solutuion of extended ODE system [T,Y] = feval(fcn_integrator,rhs_ext_fcn,[t t+stept],y); t=t+stept; y=Y(size(Y,1),:); for i=1:n1 for j=1:n1 y0(n1*i+j)=y(n1*j+i); end; end; % % % construct new orthonormal basis by gram-schmidt znorm(1)=0.0; for j=1:n1 znorm(1)=znorm(1)+y0(n1*j+1)^2; end; znorm(1)=sqrt(znorm(1)); for j=1:n1 y0(n1*j+1)=y0(n1*j+1)/znorm(1); end; for j=2:n1 for k=1:(j-1) gsc(k)=0.0; for l=1:n1 gsc(k)=gsc(k)+y0(n1*l+j)*y0(n1*l+k); end; end; for k=1:n1 for l=1:(j-1) y0(n1*k+j)=y0(n1*k+j)-gsc(l)*y0(n1*k+l); end; end;

znorm(j)=0.0; for k=1:n1 znorm(j)=znorm(j)+y0(n1*k+j)^2; end; znorm(j)=sqrt(znorm(j)); for k=1:n1 y0(n1*k+j)=y0(n1*k+j)/znorm(j); end; end; % update running vector magnitudes for k=1:n1 cum(k)=cum(k)+log(znorm(k)); end; % normalize exponent for k=1:n1 lp(k)=cum(k)/(t-tstart); end; % Output modification if ITERLYAP==1 Lexp=lp; Texp=t; else Lexp=[Lexp; lp]; Texp=[Texp; t]; end; if (mod(ITERLYAP,ioutp)==0) fprintf('t=%6.4f',t); for k=1:n1 fprintf(' %10.6f',lp(k)); end; fprintf('\n'); end; for i=1:n1 for j=1:n1 y(n1*j+i)=y0(n1*i+j); end; end; end;

3.2 Function where the equations are defined

function f=lyapunov_chaos(t,X) % Values of parameters c1 = 13.67*20; c2 = 1.56*20; c3 = 0.59; c4 = 1176; c5 = 2.86*20;

u= 4; kp = 0.001; k1 = 1; k = 1.5;W = 181.1; T= 8.5; Y= [X(5), X(6), X(7), X(8), X(9), X(13), X(17); X(10), X(14), X(18); X(11), X(15), X(19); X(12), X(16), X(20)];

f=zeros(16,1); % Differential equation f(1)=-c1*X(1)+c2*X(4)-(k*c1/u)*X(2)*X(4); f(2)=-c1*X(2)+c2*u+(k*c1/u)*X(1)*X(4) ; f(3)=-c3*X(3)-c4*(c5*(X(2)*X(4)-X(1)*u)-T-(c3/c4)*W) ; f(4)=(k1-kp*c3)*X(3)-kp*c4*(c5*(X(2)*X(4)-X(1)*u)-T(c3/c4)*W) ; %Linearized system Jac=[ -c1, (k*c1/u)*X(2); (k*c1/u)*X(4), (k*c1/u)*X(1); c4*c5*u, c4*c5*X(2); kp*c4*c5*u, kp*c4*c5*X(2)]; %Variational equation f(5:20)=Jac*Y; %Output data must be a column vector -(k*c1/u)*X(4), -c1, -c4*c5*X(4), 0, c20, -c3, -

-kp*c4*c5*X(4), (k1-kp*c3),

3.3 Calling function:

[T,Res]=lyapunov(4,@lyapunov_chaos,@ode45,0,0.5,50,[0 .4 -200 6],10); figure plot(T,Res); title('Dynamics of Lyapunov exponents for T=8.5 and ratio for C1,C2 % C5 is 20'); xlabel('Time'); ylabel('Lyapunov exponents');

4) Code used for controller design.

4.1)Uncontrolled system function [ dx ] = controller( t,x)

%This function is used for obtaining the state response for the given %system. % The function is called using ode45 command and following conditions % were taken into account. % tspan = 0:0.1:100; % options = []; % initvalue = [ 0 .4 -200 6 ]; % The values of the various constant used are given below. c1 = 13.67; c2 = 1.56; c3 = 0.59; c4 = 1176; c5 = 2.86; u= 4; kp = 0.001; k1 = 1; k = 1.5;W=181.1; T= 0.5; dx=zeros(4,1); dx(1)= -c1*x(1)+c2*x(4)- (k*c1/u)*x(2)*x(4); dx(2)= -c1*x(2) + c2*u + (k*c1/u)*x(1)*x(4) ; dx(3) = -c3*x(3)-c4*(c5*(x(2)*x(4)-x(1)*u)-T-(c3/c4)*W) ; dx(4) = (k1-kp*c3)*x(3)-kp*c4*(c5*(x(2)*x(4)-x(1)*u)-T(c3/c4)*W) ; end

4.2) Controlled system

function [de]= smt_error(t,e) % de(t) = Ae + K*w(t) % calling function. is [t,e]= ode45(@error1,[0:0.5:10],e_initial); de=zeros(4,1); de(1) = -2256.6339*e(1) + 1.445*e(2) + -0.90157*e(3) + .4667*e(4) - .0067*sign(5 * e(1)); de(2) = -6728.8917*e(1) + -9.335*e(2) + -2.7047*e(3) + 3.28*e(4) - .02*sign(5 * e(2)); de(3) = 11210.0361*e(1) + 1.445*e(2) + -1.4916*e(3) + 1.093*e(4) - .0067*sign(5 * e(3)); de(4) = -2229.511*e(1) + 1.445*e(2) + .0978*e(3) + 1.093*e(4) - .0067*sign(5 * e(4)); end

4.3) Final calling function used for calling above 2 functions.

initvalue = [ 0 .4 -200 6 ]; tspan = 0:0.147:10; options= []; [t,x]= ode45(@chaos_phaseplots,tspan,initvalue,options); initvalue = [ 0.3 0.5 0.2 0.4 ]; [t,y]= ode45(@controller,tspan,initvalue,options); %finding out the error of the system without controller. e= x-y; initvalue = e(69,:); [t1,z]= ode45(@smt_error,10:0.1:30,initvalue,options); % Plotting the results obtained. plot(t,a(:,4)) hold on plot(t1,z(:,4)) xlabel('Time T/(sec)') ylabel('e4') title('Synchronization error between 2 systems')

References
[1] Diyi Chen, Peng Shi and Xiaoyi Ma, Control and synchronization of chaos in an induction motor system, International Journal of Innovative Computing, Information and Control, Volume 8, Number 10(B), October 2012. [2] D. Y. Chen, Y. X. Liu, X. Y. Ma and R. F. Zhang, Control of a class of fractionalorder chaotic systems via sliding mode, Nonlinear Dynamics, vol.67, no.1, pp.893901, 2012. [3] D. Y. Chen, W. L. Zhao, X. Y. Ma et al., No-chattering sliding mode control chaos in Hindmarsh Rose neurons with uncertain parameters, Computers and Mathematics with Applications, vol.61, no.8, pp.3161-3171, 2011. [4] D. Y. Chen, C. Wu, C. F. Liu et al., Synchronization and circuit simulation of a new double-wing chaos, Nonlinear Dynamics, vol.67, no.2, pp.1481-1504, 2012. [5] M. Rafikov and J. M. Balthazar, on control and synchronization in chaotic and hyper chaotic systems via linear feedback control, Communications in Nonlinear Science and Numerical Simulation, vol.13, no.7, pp.1246-1255, 2008. [6] A. S. de Paula and M. A. Savi, A multiparameter chaos control method based on OGY approach, Chaos, Solitons & Fractals, vol.40, no.3, pp.1376-1390, 2009. [7] H. Wang, Z. Z. Han, W. Zhang and Q. Y. Xie, Synchronization of united chaotic systems with uncertain parameters based on the CLF, Nonlinear Analysis: Real World Applications, vol.10, no.2, pp.715-722, 2009. [8] M. Ataei, A. Iromloozadeh and B. Karimi, Robust synchronization of a class of uncertain chaotic systems based on quadratic optimal theory and adaptive strategy, Chaos, vol.20, no.4, pp.043137, 2010. 7248 D. CHEN, P. SHI AND X. MA [9] B. Jiang, P. Shi and Z. Mao, Sliding mode observer-based fault estimation for nonlinear networked control systems, Circuits Systems and Signal Processing, vol.30, no.1, pp.1-16, 2011. [10] D. Y. Chen, R. F. Zhang, X. Y. Ma et al., Chaotic synchronization and antisynchronization for a novel class of multiple chaotic systems via a sliding mode control scheme, Nonlinear Dynamics, vol.69, no.1-2, pp.35-55, 2012. [11] K. T. Chau and Z. Wang. Chaos in electric Drive Systems Analysis, Control and Application. Singapore: Wiley-IEEE Press 2011. [12] The Wikipedia online encyclopedia: www.wikipedia.com .