Speed and Motor Position Estimation of Electronically Commutated Motor

Abstract

The paper presents a method for estimating the rotation speed and position of the rotor of an electronically commutated motor (EKM). When designing the method, known estimation procedures for EKM, synchronous motor with permanent magnets on the rotor and asynchronous motor were analyzed from the literature. The presented method is based on the application of the extended Kalman filter (PKF) and a discrete motor model with line voltages. When applying the PKF, it was necessary to solve a number of specific problems, which are mainly related to the characteristic waveforms of voltages and currents of EKM. The measured quantities are line voltages and motor currents that are not filtered using analog filters, which is otherwise common with similar methods. This is made possible by synchronizing the measurement and control of the transistors of the converter so that the influence of noise on the measured values ​​is minimized at the time of measurement. The mean voltage value during the discretization period is determined by combined measurement and calculation procedures. In this calculation, the intermittency of the transistor is used, which is calculated by the predictive current regulator at the beginning of the discretization period. This controller is based on a mathematical model of the engine. Due to the application of PKF, special attention was paid to the selection of covariance matrices of system and measurement noise. The measurement noise covariance matrix depends on the accuracy of the current measurement and its members are constant throughout the estimation process, while the members of the system noise matrix depend on the accuracy of the motor model and the accuracy of the voltage determination, and change adaptively depending on the motor operating mode. Optimization of the coefficients of the system noise matrix was carried out by simulation and experimental procedures. In order to obtain basic knowledge about the quality of the presented estimation method, a simulation was carried out on a digital computer. For this purpose, a computer program was created that simulates the electric motor drive with EKM and the estimation procedure. Two variants of the program for simulating the estimation procedure were created: (1) with constant engine parameters and (2) with simultaneous estimation of state variables and engine parameters. The simulation results of stationary and dynamic modes of operation with constant engine parameters showed that it is possible to estimate the speed and position of the EKM rotor with high estimation accuracy. The sensitivity of the estimation results to the motor parameters was tested by simulation, and significant sensitivity of the results to the stator winding resistance and the electromotive force constant was determined, while the results were not significantly sensitive to the stator winding inductance. The influence of the covariance matrices of the system and measurement noise on the estimation results was also analyzed. The simulation The possibility of simultaneously estimating state variables and each motor parameter individually was investigated, and satisfactory results were obtained in estimating the stator winding resistance and the electromotive force constant. Before experimentally verifying the estimation results, the motor parameter identification procedure was carried out. This procedure consists of experiments and numerical calculations. , and is based on minimizing the difference between measured and calculated quantities in the time domain. To determine the parameters of the stator winding, it is important to conduct an experiment that ensures that the current waveforms in that experiment are similar to those that occur in real operation. The results obtained show that the resistance of the stator winding winding depends on the amplitude and frequency of current pulsations, while inductance is practically independent of these quantities. The generator idle test determined the electromotive force constant, which differs from the nominal value by approximately 5%. For the purpose of experimental verification of estimation results, a laboratory facility was built. All control functions and the estimation algorithm are implemented using a DSP, type TMS320C50. Thanks to the phase symmetry of the stator winding, the matrix equations of the PKF were simplified, which significantly simplified the estimation program in the DSP and shortened its calculation time. In stationary modes of operation, the results were experimentally checked at different speeds and loads. It was shown that the estimation results are better at higher speeds and with a loaded engine. The results were experimentally verified in dynamic conditions with a sudden change in speed reference and shock load. These results are practically equal to those in the stationary state due to the relatively large inertia constant. The possibility of simultaneous estimation of state variables and resistance of the stator winding was tested experimentally. In contrast to the simulation results, it was determined that the stator resistance cannot be estimated using this method. Finally, special attention should be paid to filtering the estimated speed, and additional delay due to the application of PKF should be taken into account when synthesizing the speed control circuit. The procedure for determining the time constant of the PKF is presented. The use of the estimated speed as a feedback signal has been experimentally verified, whereby satisfactory indicators of the quality of regulation have been obtained. The accuracy of maintaining the stationary speed (around the reference value) in the amount of ±1% of the nominal speed has been obtained at all loads and at all speeds greater than 5% of the rated value.

Božo Terzić

Full Professor | Department of Electrical Drives and Industrial Control

Full professor at the Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture in Split, with significant contributions in the field of industrial development projects including the design of prototypes of electronic converters used in industrial plants around the world. His research interests are focused on the application of electronic converters in electric drives and renewable energy sources.