Sliding Mode Control of Suspended Pendulum Hakan Kizmaz Department of Electrical and Electronics Engineering Sakarya University Sakarya, 54187 Turkey kizmaz@sakarya.edutr Abstract—This paper presents a robust control technique application which is called sliding mode control. Sliding mode ‘control algorithms is a useful algorithms because It is low order and usable for every control system. As a matter of fact the system which was used for thie study, has a nonlinear function such as sine. However the linear part of the mathematical model ‘of suspended pendulum system was used in this study in order to have both simulation and real time results corresponded with ‘each other, Finally sliding mode control algorithm was observed as a robust control ofthe system, Keywords—Suspended Pendulum System, Control, Position Control Sliding Mode 1. INTRODUCTION A simple pendulum system is a mechanical system that exhibits periodic motion. It consists of particle like bob of ‘mass m suspended by a light string of lenght that is fixed at the upper. The motion occured in the vertical plane is driven by a gravitional foree[!]. Pendulums have been known to be @ good resonators and suitable for use as a reference in high ‘quality clocks for along time[2]. A compound pendulum is @ standart topic in most physics courses because it includes some physical subjects such as the simple harmonic motion, the period of oscillation, the acceleration of gravity, the center of mass, the moment of the inertis, momentum etc.[3]. That's why the compound pendulum system can be used in many field like determination ‘of acceleration of the gravity, determination of the mass distibution of something, the measurement of time ete. Furthermore the mathematical equation related to torque foree input and pendulum angle variable is a good model in order to examine the behaviours of control slgoritms which applies to control systems. So pendulum systems can be used for analyzing control algorithms, ‘Variable Structure Systems (VSS) have been described in the literature by numerous Soviet authors including Emel’yanov [4,5], Utkin [6,7,8] and Itkis [9]. In their pioneer ‘works, the considered plant was a linear second-order system. ‘modelled in phase variable form. The most distinguished feature of VSC is its ability on result in very robust control systems; in many cases invariant control systems result. The term “invariant” means that the system is completely insensitive to parametric uncertainty and external Modem Electric Power Systems 2010, Wroclaw, Poland Saadcttin Aksoy Department of Electrical and Electronies Engineering Sakarya University Sakarya, 54187 Turkey saksoy@sakarya.cdu.tr Aydin Mahirct Department of Electrical and Electronies Engineering Sakarya University Sakarya, 54187 Turkey amuhureu@sakarye.edvetr disturbances. The basic philosophy of sliding mode control is that the structure of the feedback control is altered as the state crosses discontinuity surfaces in the state space with the result that certain desirable properties are achieved. The switching of the control function yields total invariance to system parameter variations and disturbances, and closed-loop eigenvalue placement in time-varying and uncertain systems, ‘VSS design entails the choice of the switching control and the associated switching logic to yield sliding motion on the switching surfaces. Although sliding mode control method provides a robust controller for nonlinear dynamic systems, it inherits a discontinuous control action and hence chattering phenomena will take place when the system operates near the sliding surface[ 10-13], ‘The sliding mode control of suspended pendulum has been purposed in this study. Firstly, the mathematical equations of the proposed suspended pendulum is given and then obtained linear transfer function. Secondly, some basic knowledges about the sliding mode control which is a kind of control techniques of variable structure control systems are examined. After that the experimental setup which is related to suspended pendulum prepared in the lab and then the approximate mathematical model of the suspended pendulum vwas obtained fiom the input/output measurements. Finally, codes in MATLAB/Simulink and in Labview has been developed to obtain simulation and real time experiment results which is comesponded to various reference angle inputs. The proposed control algorithm was separately implemented in MATLAB/Simulink and Labview as real time and it was applied to different cases. Experimental results which were obtained by the proposed contro] algorithm are in harmony with simulation results. I SUSPENDED PENDULUMSYSTEM ‘The schematic diagram of the suspended pendulum control system is given in Fig. 1. It i clear that in order to analysis and control of a physical system, it is necessary to be known the mathematical model of it. On the pendulum system given in Fig. 1. there is a DC motor with a propeller on the lead of a suspended stick. Here, u(t) is the control input and the angle which is between pendulum arm and vertical axis is the control variable. Pendulum is driven by DC motor. From Fig, 1. we can write the tansfer fonction of the suspended pendulum as follows: MEPS'10 - paper P149) 1 —a i 7) ~ Js?4estmiga o Drive JA/D_ PC D/A F-] = £ Figure 1 Sytematic diagram ofthe sspended pendulum control sytem where: T(e) = the thrust which is provided by DC motor saght of pendulum m= weight of pendulum 1¢ distance from suspending point fo center of mass J inertia moment sceleration of gravity iscous damping coefficient ‘The rational equation between V veltage which applied to ‘DC motor and thrust T, ean be wnitten as follows: T(s) = Ky VS) 2 els the pendulum get the stability. Then, one ean obtain £.6(¢) = 0 and #-9(0) = 0. tn this way, fom Eq, 2) can be Spritten as TH my-gsd-sindgs @ where 0., is the angle at the stable situation. Thus, substituting Eq. (2) in Eq. (3) lead to mygdsinOss v Kn a Finally from Eq. (1) and Eq. (2) the open-loop block diagram ‘of the pendulum system can be obtained as follows: va] [re] wes) med safse Figure 2 The cper-oop block diagram of suspended pendulum Now, fiom above figure the transfer function of the suspended pendulum can be written as follows: 8 6 v(s) se Ges IL, SLIDING MODE CONTROL Sliding mode control is a kind of control technique of variable structure control systems. Sliding mode control technique is based on forcing the error vector into desired dynamic and holding in this dynamic, Variable structure systems, was defined in Soviet Russia in 1964s for the first time by Emelyanov [4] and in that study, defining a line at the phase space of second order systems and forcing and holding the states of system to that line under any initial situation was discussed. The line whereof was discussed was called as surface, switching surface, sliding surface over time. The basic philosophy at the sliding mode contol is to generate control strategy atthe sliding surface. ‘The position, angular velocity and phase curves related to a suspended pendulum which is defined in the phase space is siven in Fiz. 3. © (degree) / « (degree/second) 8 a) a Time (second) @ 0) (degree!seoond) © (degree) x » Figwe3 The curve: related toa aupended pendulum sytem (2) Position and ‘angular elect cues () Phase plane cure,Here, x; ve x) state variables are angular position and angular velocity, respectively. The trajectory of this state ‘vanables is called “state trajectory”. The coming closer of the state trajectory to origin means that decreasing the error. The slate trajectory is tried to hold on the sliding surface in sliding ‘mode control technique. In the variable structure systems theory, state trajectory ‘can be a stable system which is coming eloser to origin. Then, Ietus define a switching function as follows s(y,%2) = men $3 © ‘This fonction determines the subsystems which is going to be used in the phase space. KO) -Kxs(0) s(x) +x, 50 . ute) f.uco| If the system is changed according to the switching rule like the Eq, (7), @ structure which is similar in the Fig. 4. is obtained. The state trajectories of the system is oriented to region which is biased to come closer towards origin, Then, examined system converges to origin asymptotically. Figue 4 Th stato usjectory ofthe variable sructue sytem ‘A new system was obtained by associating two subsystems each other according to a convenient decision rule through variable structure control theory. From chosen under any’ initial condition, the phase trajectory of the new obtained system is oriented to sliding surface, After the state trajectory reaches the sliding surface, it shows a motion approaching to origin asymptotically. This motion is called as sliding mode. ‘The m parameter at the Eq. (6) is @ positive constant which provided that 0 < m < VK. The different state trajectories can be illustrate related to different m parameters[8, 11] IV, MODELLING The experimental static characteristic curve related to the pendulum system is given in Fig. 5. It is clear that the proposed pendulum system in the experimental setup has a nonlinear behavior. But, from Fig. 5., itis easy to show that the relation of between DC supply and 9 is approximately linear between 0° and 60°. In order to obtain aq, Ky and ¢ parameters in Eq. (S) experimentally, the response of the pendulum which corresponds to fiee release from any initial point is enough, The experimental response curve which corresponds to fige release from 9 = 60° is given Fig. 6. @ (degree) De motor Supply) ‘Figure 5. The expernental static characteristic curve elated tothe pondluen S885 od anguar Postion (sere) BEERS 0: 2 4 6 8 1 2 4 18 te 2 Time (send) Figure 6 The positon change curve ofthe pendslum system which hs 60° inal conden Thus, if we consider that open-loop transfer function in Fig. 2. isa second order for only between 0° and 60°, it can be rewritten as 863) _ __Km(oon)? V(S) — S242¢0n+(n)? (8) By using the parameter calculating techniques of @ second order system [14,15], . Ky, and is obtained as follows: y= 3.1621 rad/sn 01137 ‘The characteristic polynomial related to the transfer function given at the Eq, (8) can be written easily as P(s) = s? + 0,719s + 9,998 0 Due to the fact that the Eq. (9) is going to be equal to parametric polynomial, Eq. (9) can be written as follows: mudg 7 (a0) P(s)=s? 4 ist