Improved induction machine torque observers and virtual loading in industrial applications, based on real-time embedded controllers

Abstract

This thesis presents an extension of the use of standard industrial command and control equipment in different applications. Three types of programmable logic controllers (PLC) are investigated as the main controller for electrical machine torque estimation and also for frequency spectrum analysis. The major drawback of the investigated low-cost PLCs related to this topic is highlighted. With high applicability in condition monitoring and fault detection, three types of electrical machines' electromagnetic torque estimators are presented, simulated, and experimentally tested on a high-speed data acquisition and control platform based on cRIO-9086 chassis. Online and offline-obtained results are given in the context of steady-state and dynamic load opperation. An artificial loading method for electrical machines is given and investigated as a niche application of the standard low-cost equipment. Two identical induction machines without mechanical coupling driven by two identical dc link-connected variable frequency drives are artificially loaded at phase-rated RMS current. A low-cost PLC is used for command and control. A simple openloop method with different loading reference types validates the synthetic loading principle. A closed-loop control method is successfully used for automatized thermal testing of both machines. Frequency spectrum analysis presents the inverters and the machine windings' connection influences. Comprehensive acceptable results related to the loss equivalency are obtained from simulation. A high-speed data acquisition and control system is used as a torque estimator for online and offline analysis of a novel active torque pulsation reduction method for rotary electromechanical systems with position-dependent loading torques. A mathematical two-step method used for preliminary total system inertia and loading torque detection is presented in detail and experimentally validated.