Research On Inverter Nonlinearity And Its Control Strategy In Induction Motor Drives

Variable frequency speed control system of induction motors has been increasingly popular in industrial areas, and higher requirement of its performance is still needed. On one hand, the accurate compensation for inverter voltage error is very important to improve the low-speed control performance of a variable frequency speed control system, because accurate compensation helps to reduce the current distortion of V/Hz control as well as increase the flux and speed observation accuracy of sensorless vector control. On the other hand, the suppression of inverter nonlinear error will improve the accuracy of offline motor parameter identification. Besides, high frequency characteristics of the output voltage are also nonliearity of the inverter, which lead to high frequency current ripple and acoustic noise. The alleviation of the acoustic noise is needed in some special applications.First, the composition of inverter nonlinear voltage error is analyzed. The relationship between the inverter voltage delay as well as the device voltage drop and current value is investigated deeply. Based on these analyses, an auto measurement method of inverter output voltage delay is proposed. With this method, the model of output voltage delay to current, which is used in inverter voltage error compensation, can be built up quickly and accurately. The method needs no other hardware except for the motor load. The accuracy of the measurement is assured by decreasing the influence of stator resistance variation at differenct temperatures. The compensation performance based on the auto-measured model is approaching to the manually measured models.What’s more, the measured voltage delay model can be used to compensate for inverters at different PWM frequencies, which is an advantage of this method. The effectiveness and accuracy is proved by measuring and compensation experiments.Secondly, to lower the impact of inverter nonlinear errors on motor parameters identification, the paper proposes a dc biased stimulation method for single phase ac stimulation in offline motor parameters identification. With the dc biased stimulation method the parameters identification can be operated without voltage error compensation. Simulations and experiments of the dc biased method implemented in two motor parameters indentificaiton techniquses (Impedence Equations Set technique and Recursive Least Square technique) are given respectively to prove the effectiveness of the propose method.At last, random SVPWM method is researched in this paper to suppress the acoustic noise of inverter-fed induction motors. Theoretical analyses of power spectrum density of both random postion SVPWM and random switching frequency SVPWM are carried out. Besides, several kinds of random SVPWM methods are realized in DSP programs. The experiments of these random SVPWM methods are implentend to verify and compare their performance. Then characteristics of those common random SVPWM methods are summarized.