Abstract:
By 2030, many governments have shifted their focus to electrical vehicles (EVs). Consequently, this project focuses on transforming a KIA Pride 2000 model to an EV by optimizing the electronic control unit. The objective is to develop an efficient and reliable propulsion system for the EV. To achieve precise speed control of the induction motor, an algorithm known as Pulse Width Modulation with Selective Harmonic Elimination (SHE PWM) is used. However, due to the time-consuming nature of the numerical techniques required for calculating switching angles, the SHE PWM algorithm is impractical for real-time applications. To overcome this challenge, two approaches are compared, one using Artificial Neural Networks (ANN) and the other on Polynomial Interpolation (PI), both in combination with the SHE PWM algorithm. This thesis describes, and implements both algorithms into a microcontroller to evaluate the accuracy and speed of both methods. The results demonstrate the superiority of the ANN approach. To validate the algorithm in a real-time application, an FPGA implementation is presented and discussed. The application is tested on a variable speed induction motor test bench. The obtained results indicate that the ANNSHE PWM algorithm efficiently controls the fundamental voltage, eliminating the desired harmonics in real-time across the entire range of speed variations.
