Torque Control in a Two-Mass Resonant System: Simulation and Dynamic Analysis

EasyChair Preprint 9665

Abstract

A multi-mass system is a mechanical system that consists of several masses such as a motor, load, and gear that are connected by a flexible shaft. The mechanical fluctuations in drives with flexible coupling between the motor and the driven device can no longer be ignored, as they could in the past when requirements for speed control dynamics were low. These prerequisites pertain to the regulation of the rotational speed and position of the servo drive and are quantified by the magnitude of speed step response time and speed fault elimination time caused by the step change of load torque, which, in modern drives, can be measured. In this study, the dynamical behavior of a two-mass resonant system with a three-term controller for control of torque and speed is investigated using eigenvalues analysis. The proposed control strategy aims to eliminate the rotational fluctuations of the motor shaft, dampen the load torque disturbance impact, provide a quick response to changes in the base speed while avoiding increases in the load speed, and be resilient to instability. Finally, numerical outcomes demonstrate the effect of the presented controller application on improving the dynamic demeanor of the two-mass test system. Based on the outcomes, the effectiveness of the proposed control scheme is highly dependent on system parameters. Due to the inherent parameter uncertainty in the multi-mass system, the use of parameter estimators based on artificial neural networks (ANNs) is suggested for future work.

Keyphrases: Artificial Neural Networks (ANNs), Eignevalues, Multi-mass electromechanical system, torque control

@booklet{EasyChair:9665,
  author    = {Mansoor Zeinali and Sayed Mohammadali Zanjani and Somaye Yaghoubi and Amir Mosavi and Arman Fathollahi},
  title     = {Torque Control in a Two-Mass Resonant System: Simulation and Dynamic Analysis},
  howpublished = {EasyChair Preprint 9665},
  year      = {EasyChair, 2023}}