Frequency Speed Control of Rotary Travelling Wave Ultrasonic Motor Using Fuzzy Controller

This paper proposes frequency speed control of rotary travelling wave ultrasonic motor (TWUSM), type Daimler Benz AWM90-X motor. The control characteristics of TWUSM are complicated, highly nonlinear and varying in time. This can lead to deterioration of the performance of conventional controller such as proportional integral (PI). In order to achieve high control performance of the TWUSM, fuzzy logic controller (FLC) has been designed and compared to the conventional PI controller. To validate the performance of the proposed FLC, simulation of the speed response has been performed and analyzed for a varying load. The simulation results show that the FLC has smaller settling time, smaller rising time and minimum error in steady state. Furthermore, the fuzzy controller provides good results for large load variations. The frequency output of the controller has been validated with experimental measurements of AWM90-X. Keywords-nonlinear control; fuzzy logic controller; travelling wave ultrasonic motor type Daimler–Benz; AWM90-X


INTRODUCTION
TWUSMs have structural and operational advantages compared to conventional electromagnetic motors, such as compact size, lighter weight, very low speed operation, high torque, nonmagnetic operation, freedom of constructional design, very low inertia, high speed response, possibility of electromagnetic noise reduction and miniaturization [1].Moreover, the settling time in ultrasonic motors is very short, fast response is one of the most important TWUSM advantages, which makes it suitable for applications with fast response demands such as robot actuators and auto-focus cameras.Several types of ultrasonic motors have been suggested and designed in the last 25 years [2].Their operating principle is based on piezoelectric vibrations that convert electric energy to mechanical energy in the form of elastic vibrations [3].It consists of two basic parts: the stator with piezoelectric ceramics, and the driven part (rotor).Apart from the advantages of TWUSM, deriving mathematical model for TWUSMs is a difficult task due to their complicated and highly nonlinear characteristics.In addition, the dynamic speed characteristics of TWUSM are time-varying due to the increase in temperature and depend on operating conditions such as driving frequency, source voltage and load torque [4,5].
The speed control of TWUSMs is one of the important issues under consideration [5][6][7].Three kinds of speed control strategies exist in the literature.The TWUSM speed can be controlled by the driving frequency, the phase difference or the voltage amplitude of the two excitation sinusoidal voltage of TWUSM.However, aimed at these problems, several researchers have opted for the driving frequency of the sinusoidal voltage as control variable [6][7][8].There is no perfect control scheme for TWUSMs.The ultrasonic motor speed controllers belong to several conventional and numeric controller types.A speed tracking control system using both neuro-fuzzy control and direct pulse width modulation for TWUSM has been developed in [8].In the same way, authors in [9] used also a fuzzy neural network controller.Authors in [4,10,11] studied a servo speed control system for travelingwave ultrasonic motor [4] and a servo position control for TWUSM [10] where fuzzy neural network was the control tool [11].Authors in [12] proposed a position control scheme for TWUSMs.Authors in [13] used H∞ strategy to control rotary travelling wave ultrasonic motor TWUSM while authors in [14] proposed an intelligent PID speed controller of TWUSM.In a nutshell, most of developed works for ultrasonic motors control focused on a motor of USR60 type.However, the speed control for Daimler-Benz AWM90-X motor has rarely been explored so far, which motivated us to make this study as a continuation of our previous works [15][16][17].The main contribution of this work consists to the design of a fuzzy logic controller to drive the speed of the Daimler-Benz AWM90-X motor.
ere ω=2πf and the mechanica reference link to be useful t h the travellin ordinate system ve: In the contac tor into the c
(6) (7) ncy.The mod r and the rotor t effective.It ate system mo ding to the in on of the trave nd (7) as: een stator and n in [16].In su h, x l and x r ar ngential speed (10) interpenetratin creation of no      Motor Conditio esponse chara Figure 8

D. The Third Load 3.3Nm
The motor has been loaded with a big load larger than the torque that can be supported by the PI controller.From the results given in the Figure 9(c), we note that the FLC can drive, similar to the previous cases, the TWUSM to track the reference speed.In the range 2.5 3 t ms ms   , the motor speed decreases under the load effect and then it increases to reach the reference speed.Noting that for a large load, greater than 3.3Nm, the TWUSM under FLC diverges.This phenomenon, called pull out phenomenon, has been signaled in the literature [24].It is due especially to the variation of the resonance frequency.Finally, FLC is more robust than the PI controller and it presents low error.Therefore, it is well suited to this actuator type which can be employed for varying loads.To ensure the good operation of the TWUSM, the motor should be driven with the antiresonance frequency to reach the maximum of its efficiency [25].To determine the antiresonance frequency of the TWUSM AWM90-X, the stator parameters were calculated with finite elemet method (FEM), obtained from the recorded admittance curve of the stator.The admittance curve of the free-vibrating stator was recorded with the impedance analyzer with a step size of 10Hz. Figure 10 shows the resulting measured curve.From Figure 10, the antiresonance frequency is equal to 43.425kHz [25].Regarding the proposed TWUSM model, we can note that it was validated in our previous work [17] without control speed.In fact, the simulation frequency of the proposed model was equal to 43.425kHz which is the same with the one obtained in experimental validation [25].Moreover, to validate the proposed controller, the frequency output of the FLC is illustrated in Figure 11.Hence, the simulation frequency has been found equal to 44.6kHz, which is closer to the experimental antiresonance frequency (43.425kHz).This result validates the efficiency of the FLC.In this paper, the frequency speed control of the TWUSM, type Daimler Benz AWM90-X, has been studied.The design of FLC and PI controllers has been presented and discussed.To illustrate the efficiency of the proposed controllers to drive the TWUSM to track perfectly the reference speed, thorough simulations were performed.The obtained results have revealed that the FLC controller performs better than the PI controller with very good tracking performance, namely low rise time, low settling time and high steady state accuracy.Likewise, the FLC controller seems more robust that the PI controller against load variations.Finally, the frequency output of the FLC has been validated with experimental measurements of the TWUSM AWM90-X.Frequency [HZ]

ng, Technology r.com
The robustness of the proposed controller against load

TABLE I .
.Structure of