Application of Neural Network for DC Motor using Controlled Three Phase Converter Drive in Simscape Electrical

The objective of this research is to study the implementation of Artificial Neural Networks (ANNs) for developing an alternative method of motor speed control compared with traditional Proportional-Integral-Derivative (PID) controllers, which may require desired accuracy and time-varying industrial automation/robotics. This work seeks to enhance the performance of motor speed control using ANNs which can learn and adapt to more complex nonlinear interactions. The study proposes a comparative analysis of controllers such as PID and ANN-based under-motor speed control systems, for detailed data acquisition over diversified operating conditions to evaluate their usefulness. To address the limitations of being able to implement ANNs in real-time, such as complexity in computation and training time, we conducted this study by improving our architecture along with more efficient learning methods by tuning or customizing them over different hyper-parameters. Therefore, the final study report will enable a better understanding of the pros and cons of ANN utilized in motor control applications that can have an impact on improving speed regulation for motors besides more industrial automation or robots.