AI-driven, Model-Free Current Control: A Deep Symbolic Approach for Optimal Induction Machine Performance

TL;DR

The paper addresses the challenge that traditional PI current control for induction machines relies on precise models and struggles to decouple $d$- and $q$-axis dynamics under disturbances. It introduces a model-free current-control approach based on Deep Symbolic Regression (DSR) that outputs explicit $V_{dq}^*$ laws from data, with $v_d^*$ and $v_q^*$ expressed as functions of state inputs (e.g., $x_1=\\Delta i_{ds}$, $x_2=\ abla i_{qs}$, $x_3=\int \Delta i_{ds}$, $x_4=\int \Delta i_{qs}$) such as $v_d^* = 13 x_1 - \sin(x_1 - x_4)$ and $v_q^* = 12 x_2 + x_3 + 2 x_4 + (x_1^2 + x_1 - x_2 - x_4) \sin(x_1) + \sin(x_1(-x_1 + 2 x_3) - x_2) + \cos(2 x_2)$. Trained offline using risk-seeking policy gradient to minimize INRMSE, the method enables model-free, interpretable control suitable for real-time deployment. Experimental validation on a 3.7 kW induction machine demonstrates comparable dynamic performance to PI control while reducing harmonic distortion and eliminating motor-parameter tuning, underscoring the practical impact of integrating AI-driven symbolic methods into power electronics.

Abstract

This paper proposed a straightforward and efficient current control solution for induction machines employing deep symbolic regression (DSR). The proposed DSR-based control design offers a simple yet highly effective approach by creating an optimal control model through training and fitting, resulting in an analytical dynamic numerical expression that characterizes the data. Notably, this approach not only produces an understandable model but also demonstrates the capacity to extrapolate and estimate data points outside its training dataset, showcasing its adaptability and resilience. In contrast to conventional state-of-the-art proportional-integral (PI) current controllers, which heavily rely on specific system models, the proposed DSR-based approach stands out for its model independence. Simulation and experimental tests validate its effectiveness, highlighting its superior extrapolation capabilities compared to conventional methods. These findings pave the way for the integration of deep learning methods in power conversion applications, promising improved performance and adaptability in the control of induction machines. The simulation and experimental test results are provided with a 3.7 kw induction machine to verify the efficacy of the proposed control solution.

Figures (10)

• Figure 1: Proposed DSR vector control for an induction machine.
• Figure 2: DSR architecture.
• Figure 3: Experimental Setup.
• Figure 4: Simulation results.
• Figure 5: Experimental result for motor current response at varied speed employing proposed current control solution.