Comparison of Induction Motor Parameters Estimation Methods

Abstract

It is known that induction motors are mostly applied type of motors in industry. Therefore, their digital twin development remains relevant. One of the main issues during digital twin development is induction motor parameters determination. There are a number of methods to determine these parameters with varying degree of accuracy. The purpose of the study is to find an induction motor parameters estimation method which allows one to create a digital twin of actual induction motor. Methods. MATLAB software was used to create induction motor simulation model in phase coordinates. To determine induction motor parameters via catalogue data, the methods suggested by the following authors were used: Kravchik A.E. (1982), Moshchinskii Yu.A. (1998), Myasovskii V.A. (2020). The object of research is the induction motor AIRM80A6U3 whose actual performance characteristics were obtained during experiment. Results. Using the developed mathematical model, and methods to determine the parameters of the induction motor via catalogue data, simulation has been applied to compare actual performance characteristics and those obtained as a result of simulation. To improve results accuracy, a series of experiments was performed for stator winding phase resistance identification. The most accurate method was chosen and obtained data were included in it. The paper contains graphical images for visual comparison of simulated and actual induction motor performance characteristics. Maximum relative and integrated errors are the convergence criteria of the model and real motor. These errors are tabulated for visual reference. By analyzing simulated induction motor performance characteristics, current, speed and maximum relative and integrated errors, a method was determined which identifies induction motor parameters with highest accuracy. Moreover, using that method with precise resistances of the stator winding makes it possible to reduce the speed integral and relative error by 2.5 times. Conclusions. A comparison of the results shows good convergence, which allows the method to be used for modeling induction motor drive systems.