IMPLEMENTING A MACHINE LEARNING BASED PID CONTROL ON AN EDDY CURRENT DYNAMOMETER


Uluocak İ.

ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII, Ankara, Turkey, 2 - 04 December 2022, pp.853-858

  • Publication Type: Conference Paper / Full Text
  • City: Ankara
  • Country: Turkey
  • Page Numbers: pp.853-858
  • Çukurova University Affiliated: Yes

Abstract

The automotive companies have been using the Eddy Current Dynamometers (ECD) for the auxiliary braking system, also for engine testing and various internal combustion engine analyzing operations such as biodiesel, combustion timing and harmful gas emission. Eddy current dynamometer brings versatile advantages such as contactless braking effect and zero carbon footprint. However, Eddy Current Dynamometer braking performance varies unexpectedly divergent rotor speeds and temperatures resulting a nonlinear braking torque. PID control method which is known as the most popular conventional control method in industry is inadequate robust for such nonlinear systems. On the other hand, Machine learning methods are used as artificial intelligence-based techniques that can increase the performance of conventional control systems in particular by means of improving robustness of the designed control systems. Therefore, the proposed control system proves changeable Kp (Proportional gain), Ki (Integral gain) and Kd (Derivative gain) parameters in real-time to adapt and presents a good capacity to adapt nonlinearities and bring robustness using soft computing methods. The testing and training dataset is extracted from experimental studies and its robustness has also been verified with R2 and MAPE. The presented technique is observed to have a good performance in terms of response time (t) and accuracy of desired speed value (V) under different parameters such as non-linear dynamics (V, T) of the system elements and the varying load effects.