DQN optimization method for maximum power point tracking control of three-core cable energy harvesting devices

Addressing the challenge of sudden load response in three-core cable energy harvesting devices, this paper proposes a maximum power point tracking (MPPT) control method. Current and magnetic field equations are derived based on Ampere's circuit law and other principles. A dynamic current compensation coefficient is introduced, combined with electromagnetic coupling models and equivalent impedance calculations to establish transfer functions. A DQN model for maximum power point tracking control is constructed, featuring an equivalent circuit comprising a speed-to-power conversion unit and an MPPT control algorithm unit, with an optimization objective function derived. An adaptive step size strategy is employed to train the DQN model. Based on the trained model, parameters are adjusted, capacitors are configured, an optimization strategy is designed, and constraint equations are established to achieve DQN optimization for maximum power point tracking control. Experimental results demonstrate that after applying the DQN optimization method, approximately 70% of cable load current samples are concentrated within the 150A-220A range, with reduced fluctuation. During abrupt load changes, the fluctuation of the power tracking characteristic was controlled within 2100W-2500W, verifying that this method can effectively improve the working efficiency of the energy harvesting device.