Adaptive Control Of Permanent Magnet Linear Synchronous Motor Using Non Linear Kalman Filtering Techniques

Authors

  • B. Arundhati, Madisa ,V G Varaprasad, R.S.Ravi Shankar

Abstract

This paper proposes speed/position sensor-less controlled Permanent magnet linear synchronous motor drive based on non-linearkalman filtering techniques. The limiting factor in case of  linear motors is its  speed sensors  as it is very expensive and is problematic not only for installation but also in the maintainance. Soft sensors are proposed in this paper to be used in the place of linear encoders to specifically combat this issue.Various methods are used to estimate  the mover speed/position from the information of theterminal voltage and currents of motor. Each method differs with respect to accuracy,sensitivity and robustness against model  inaccuracy and parameter variations. Using Non-Linear Kalman techniques like Unscented Kalman Filterand Extended Kalman Filter,accurate estimations  can be achieved.Extended Kalman Filter uses linearized model whereas,Unscented Kalman Filter usesNon-Linear unscented transforms to defend the stochastic characteristics of a Non-Linearsystem.Ability to capture the  systemNon-Linear behavior using Non-Linear unscented transforms makes Unscented Kalman Filter superior than Extended Kalman Filter.Analysis and comparison of dynamic performance of both the estimators are overviewed in this paper. Experimental results for the Linear synchronous motor drive confirmbetter performance of Unscented Kalman Filterwhen compared to Extended Kalman Filterat both low and high speed operartionas well as during transients like speed and load change conditions. Simulink / embedded MATLAB are used to perform extensive simulation which  is experimentallyvalidatedwith the help of  controller board  DS1104

Published

2020-12-01

Issue

Section

Articles