6 August 2024 in IEEE Transactions on Power Electronics
Rotation misalignments cannot be avoided in an underwater wireless power transfer (UWPT) system of autonomous underwater vehicles (AUVs). However, the existing designs of magnetic coupled structures (MCS) in the AUVs’ wireless charging system can only achieve rotation misalignment tolerance (RMT) in a certain range, which would limit its practical applications because of the dead areas where the coupling coefficients of the MCS have significant variations and may even approach to zero at some rotation angles. In this letter, a novel complementary enhancement MCS with a U-shaped bipolar coil transmitter and an arc-shaped unipolar receiver is first proposed, which can achieve full range of 360° RMT capability, thus keeping stable and highly efficient outputs under any rotation angles (0°–360°). Moreover, the MCS can be reconfigured by optimizing the dimensions and winding density based on simulated results. An experimental UWPT platform with 500 W/85 kHz is built to prove the effectiveness of our designed MCS. The measured results show that output powers can remain stable at 420 W under any rotation ranges, while the system efficiency is all over 87%. Simultaneously, the constant voltage with the maximum fluctuation error of 5.37% can be achieved under load changing from 10 to 50 Ω.
26 May 2025 in IEEE Transactions on Power Electronics
Existing research on efficiency optimization for wireless power transfer (WPT) systems is limited by complicated models and low control accuracy. In this article, a data-driven efficiency optimization control method is proposed for phase-shift controlled WPT systems. This method integrates an experimental data-driven output prediction and efficiency optimization approach, driven by a multilayer perceptron (MLP) and particle swarm optimization (PSO). First, the MLP model establishes a nonlinear mapping between phase shift angles and key performance indicators, including output power and system efficiency. Subsequently, the PSO algorithm determines the optimal phase angle combination that maximizes efficiency for a given reference power. In addition, the PSO-generated control parameters are regressed using lightweight artificial neural networks, enabling real-time deployment onto microcontrollers. To verify the accuracy of the proposed data-driven efficiency optimization control WPT method with MLP and PSO, an experimental 430 W/85 kHz prototype is built. The measured results show that the mean absolute percentage error of the MLP predictive model is less than 3.5% across all selected load conditions. Meanwhile, the efficiency tracking algorithm driven by PSO ensures high-efficiency operation under variation of the reference power, load resistance, and mutual inductance with a maximum efficiency of 93.5%.
1 October 2025 in IEEE Transactions on Power Electronics
Accurate identification of key parameters is crucial for closed-loop control and system stability in underwater wireless power transfer (UWPT) systems. However, the absence of communication and severe harmonic distortion from eddy currents in sea environments present significant challenges to identification results. This article proposes a hybrid mechanism- and data-driven method for accurate joint identification of mutual inductance and load. First, to resolve the linear dependence and parameter coupling inherent in typical LCC–S compensated UWPT systems, a multifrequency excitation strategy is introduced to extract informative features for joint identification. Second, an artificial neural network is pretrained on mechanism-based data to construct the source domain model (SDM). Moreover, transfer learning fine-tunes the SDM with limited experimental data to obtain a target domain model (TDM), enhancing model generalization under varying conditions. Finally, an experimental platform with 1 kW/85 kHz is constructed to validate the proposed method. Experimental results and three different typical underwater conditions are captured and collected, which demonstrates that the proposed mechanism and data-driven joint parameter identification method can achieve quick and accurate identification results with prediction error below 3% and response time within 2.72 ms.
31 December 2025 in IEEE Transactions on Power Electronics
Online parameter identification is essential for underwater wireless power transfer (UWPT) systems to ensure charging stability and high efficiency. However, strong conductivity, harmonic interference, and parameter deviations pose huge challenges for traditional parameter identification methods that rely entirely on circuit models. Data-driven methods are an effective method to solve this problem, while a large amount of on-site data and long computation time are necessary for high accuracy, which limits its versatility. To overcome this limitation, this letter proposes a transfer-learning-driven online joint parameter identification method with a small amount of experimental data and fast computation speed for UWPT systems. First, a source-domain model is pretrained on augmented air-domain data. Then, a target-domain model is obtained by fine-tuning with limited seawater samples, enabling joint identification of mutual inductance and load resistance while reducing underwater testing effort. A 1-kW UWPT prototype was built to validate the effectiveness of the proposed method. The results show that the proposed framework provides fast, accurate, and practical parameter identification under seawater conditions. The framework achieves 0.12 ms estimation latency and errors below 5% with 109 samples.