DESIGN OF PARALLEL STRUCTURE CONTROL SYSTEM WITH PID CONTROLLER AND REINFORCEMENT LEARNING FOR ELECTRIC MOTOR SPEED CONTROL
Electric vehicles have a controller to manage vehicle operations. To increase vehicle efficiency, it is necessary to control vehicle speed and torque generated by the motor. This speed control must makes the electric vehicle running according to the specified speed (tracking set-point) and minimize...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/68456 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Electric vehicles have a controller to manage vehicle operations. To increase vehicle efficiency, it is necessary to control vehicle speed and torque generated by the motor. This speed control must makes the electric vehicle running according to the specified speed (tracking set-point) and minimize the influence of disturbances that can disrupt the stability of the vehicle speed (disturbance rejection). However, in general, electric vehicle speed controllers can only track set-points. In this study, a parallel control structure controller will be designed using PID controller and reinforcement learning to improve the motor system performance in tracking set-points and minimizing the effect of interference on the system. The controller will be implemented in a three phase AC motor.
The PID controller synthesized with RL is able to make the system stable and reached an RMSE value of 23.508 rpm, better than the Ziegler-Nichols synthesis result of 35.972. In the controller with reinforcement learning parallel control structure for any given disturbance type, the RMSE value is smaller than the PID controller without parallel structure. With input disturbance, the RMSE of the controller with reinforcement learning parallel control structure is 14.897 rpm, while the usual PID controller without parallel structure gets 18.266 rpm. With output disturbance, the RMSE of the controller with reinforcement learning parallel control structure is 19.324 rpm and 13.106 rpm, while the usual PID controller without parallel structure gets 20.872 rpm and 16.641 rpm. With set-point change disturbance, the RMSE of the controller with reinforcement learning parallel control structure is 10.282 rpm, while the usual PID controller without parallel structure gets 11.844 rpm.
|
|---|