Deep Reinforcement Learning-Aided Frequency Control of LCC-S Resonant Converters for Wireless Power Transfer Systems

TL;DR

This work addresses robust output voltage regulation for LCC-S resonant converters in wireless power transfer by integrating a Direct Piecewise Affine (DPWA) switched-affine model with a DRL-based PI tuning using the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm. The converter is represented by six DPWA subsystems with switching controlled by $s=\mathrm{sign}(\mathbf{Kx}+\mathbf{m})$, and a TD3 agent tunes the PI gains to minimize the voltage error, using state $\mathbf{x}=[\int e(t)\,dt,\ e(t)]^T$ and reward $R=-(V_{ref}-V_{out})^2$. Experiments and simulations show the TD3-tuned PI controller achieves reduced voltage ripple, improved stability, and resilience to input and load disturbances compared with conventional tuning and DDPG baselines. The work demonstrates a scalable, data-driven control framework for power electronics that can adapt to varying operating conditions in wireless power transfer systems.

Abstract

This paper presents a novel deep reinforcement learning (DRL)-based control strategy for achieving precise and robust output voltage regulation in LCC-S resonant converters, specifically designed for wireless power transfer applications. Unlike conventional methods that rely on manually tuned PI controllers or heuristic tuning approaches, our method leverages the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm to systematically optimize PI controller parameters. The complex converter dynamics are captured using the Direct Piecewise Affine (DPWA) modeling technique, providing a structured approach to handling its nonlinearities. This integration not only eliminates the need for manual tuning, but also enhances control adaptability under varying operating conditions. The simulation and experimental results confirm that the proposed DRL-based tuning approach significantly outperforms traditional methods in terms of stability, robustness, and response time. This work demonstrates the potential of DRL in power electronic control, offering a scalable and data-driven alternative to conventional controller design approaches.

Figures (11)

• Figure 1: A basic visual diagram of DPWA systems with six polyhedral regions.
• Figure 2: Block diagram of a resonant converter using DPWA technique molla2014hybrid_switched_5.
• Figure 3: LCC-S resonant converter.
• Figure 4: The variation of the voltage $V_s$ under frequency control method.
• Figure 5: The six modes of LCC-S converter with following conditions as: (1) $s=1$, $x_5>x_4$, (2) $s = 1$, $|v_{pr}|<x_7$, (3) $s = 1$, $x_5<x_4$, (4) $s=-1$, $x_5<x_4$, (5) $s=-1$, $|v_{pr}|<x_7$, (6) $s=-1$, $x_5>x_4$ .