Adaptive PI Controller for Speed Control of Electric Drives Based on Model Reference Adaptive Identification

Yuefei Zuo,Shushu Zhu,Yebing Cui,Chuang Liu,Xiaogang Lin

DOI: https://doi.org/10.3390/electronics13061067

IF: 2.9

2024-03-14

Electronics

Abstract:In this paper, to achieve auto-setting of PI controller gains when mechanical parameters are unknown, two adaptive PI controllers for speed control of electric drives are developed based on model reference adaptive identification. The adaptive linear neuron (ADALINE) neural network is used to interpret the proposed adaptive PI controller. The effect of the low-pass filter used for the feedback speed and the Coulomb friction torque on parameter identification is analysed, and a new motion equation using filtered speed is given. Additionally, a parameter identification method based on unipolar speed reference is provided. The two proposed adaptive PI controllers and the conventional PI controller are compared based on the high-precision digital simulation using MATLAB/Simulink (version R2023a). The simulation results show that both of the two proposed adaptive PI controllers are able to identify mechanical parameters, but the adaptive PI-1 controller outperforms the adaptive PI-2 controller due to its better noise attenuation performance.

engineering, electrical & electronic,computer science, information systems,physics, applied

What problem does this paper attempt to address?

The paper primarily focuses on addressing the issue of automatic parameter adjustment for speed control in electric drive systems, especially when mechanical parameters are unknown. Specifically, researchers have developed two types of adaptive PI controllers based on Model Reference Adaptive Identification to achieve precise regulation of electric drive speed control.

In traditional PID control, a first-order degree of freedom PID controller struggles to optimize both tracking performance and disturbance suppression simultaneously, particularly when faced with the complexity and uncertainty of electric drive systems. To tackle this challenge, the paper proposes an adaptive design method that uses two adaptive PI controllers to identify mechanical parameters, such as rotational inertia, viscous friction torque coefficient, and disturbance torque, thereby dynamically adjusting the controller gains.

The research employs an ADALINE neural network to interpret the proposed adaptive PI controllers and analyzes the impact of using a low-pass filter for feedback speed and Coulomb friction torque on parameter identification. Additionally, a parameter identification method based on a unipolar speed reference is provided.

The performance of the proposed adaptive PI controllers was compared with that of traditional PI controllers through high-precision digital simulations in MATLAB/Simulink. Simulation results indicate that both adaptive PI controllers are capable of identifying mechanical parameters, but the adaptive PI-1 controller exhibits superior performance due to its better noise attenuation capabilities.

Overall, the paper aims to solve a key issue in electric drive speed control through adaptive control strategies, namely the automatic adjustment of PI controller gains in the face of unknown or changing mechanical parameters, to enhance the robustness and dynamic response of the control system.