Adaptive Neural Network-Based PID Controller Design for Velocity Control of an Internal Combustion Engine Using Back Propagation Technique

Abstract

Precise velocity control in internal combustion engines (ICEs) is essential for optimizing performance, improving fuel efficiency, and minimizing emissions. However, the nonlinear dynamics and inherent uncertainties within ICE systems present substantial challenges for traditional control methods. In this paper, we propose an adaptive control strategy that integrates a Proportional-Integral-Derivative (PID) controller with Back Propagation Neural Networks (BPNNs) to effectively address these complexities. The proposed controller architecture consists of two key components: a BPNN to estimate modelling uncertainties, such as unknown friction and external disturbances affecting the ICE, and a primary PID controller responsible for velocity regulation. The BPNN functions as a dynamic estimator, continuously learning and adapting to changes in system dynamics, thereby enhancing the robustness and adaptability of the control system. By accurately capturing the nonlinearities and uncertainties inherent in ICEs, the BPNN contributes to improved control performance and system stability. To validate the effectiveness of the proposed approach, extensive numerical simulations are performed using MATLAB Simulink. These simulations encompass a range of operating conditions and scenarios to thoroughly evaluate the controller's performance. Additionally, the proposed method is compared to conventional PID control techniques found in the literature, with a focus on robustness, tracking accuracy, and disturbance rejection. The results indicate that the adaptive PID controller, incorporating BPNNs, outperforms traditional PID methods, delivering superior velocity regulation and disturbance rejection. Furthermore, the proposed approach demonstrates significant potential for real-world applications in ICE systems, providing enhanced control performance and efficiency. This study advances the field of control engineering by introducing an innovative adaptive control strategy tailored specifically for velocity control in internal combustion engines. Leveraging the capabilities of BPNNs to address uncertainties, this approach contributes to improved system performance and offers a promising direction for future advancements in engine control technologies.