Liang Jin

The complex nonlinear behavior of lithium iron phosphate batteries cannot be fully characterized by conventional electrical-thermal parameters. Ultrasonic techniques provide complementary insights through acoustic signature analysis. Accurate state-of-health (SOH) estimation during long-term cycling is addressed through ultrasonic-electrical-thermal data fusion. Unlike conventional SOH estimation methods that rely solely on electrical-thermal signals or purely data-driven models, this study further integrates physically interpretable acoustic-derived aging indicators to enhance structural sensitivity and robustness. Voltage, temperature, time-of-flight (TOF), and signal amplitude (SA) variations are analyzed during charging. Five key aging indicators are identified. An improved Bidirectional Gate Recurrent Unit (IBiGRU) model with enhanced temporal feature representation is developed to extract multi-source degradation characteristics adaptively. Only 30% of the initial cycling data is required for accurate prediction. Mean absolute error (MAE) is maintained below 0.0066 in single-battery tests. Root mean square error (RMSE) remains under 0.0078 in these tests. Cross-validation yields an MAE of 0.0195 and an RMSE of 0.0216. These results validate the effectiveness and robustness of the proposed multi-source data fusion framework for accurate state-of-health (SOH) estimation of lithium iron phosphate batteries.

State-of-charge (SOC) and state-of-health (SOH) of lithium-ion batteries are critical parameters in battery management systems, directly impacting the driving range, performance stability, and safety of electric vehicles. To improve the accuracy and stability of SOC and SOH estimation simultaneously, this paper proposes a joint estimation method with constant-current bias compensation. First, based on a second-order RC equivalent circuit model, a constant-current bias compensation term is introduced into the Kalman filter framework. The estimation accuracy and robustness of SOC are validated under multiple operating conditions and noise levels. Then, a model integrating Transformer and gated recurrent unit is constructed. The fata morgana algorithm (FATA) is adopted for hyperparameter optimization. Ablation studies and multi-model comparative experiments are conducted to verify the model’s accuracy. Finally, capacity correction is performed using SOH results. By combining current bias compensation and precise temporal features extracted from aging data, joint estimation of SOC and SOH is achieved. Results show that after introducing current bias compensation and aging-based capacity correction, the accumulated SOC estimation error is reduced by more than 10%, while SOH estimation achieves a MAPE below 0.90% and an RMSPE below 1.10%. The proposed joint method is thus verified to be accurate and practical.

Traditional finite element methods (FEM) can accurately simulate the coupled electromagnetic–thermal fields in wireless power transfer (WPT) systems, but they suffer from high computational costs and slow response times in multi-condition analysis and structural optimization. Therefore, this paper proposes a generative deep learning model based on the conditional diffusion mechanism, termed CoUSD, which takes five-dimensional operating parameters such as coil turns, transmitting current, receiving current, lateral offset and vertical distance as conditional inputs. A unified structural network (UniStructNet) is constructed by integrating residual convolutional blocks and Transformer mechanisms, enabling efficient generation and accurate reconstruction of electromagnetic-thermal fields from Gaussian noise. In simulation-based validation, the model achieves a mean squared error (MSE) of 0.43×10⁻³, a mean absolute percentage error (MAPE) of 2.5%, and a high-frequency MSE (HF-MSE) below 0.017, with a single inference time of only 1.8 seconds—approximately 100 times faster than FEM. In experimental validation at discrete measurement points, the predicted results yield a coefficient of determination (R²) close to 0.9 with respect to the measured data. These results demonstrate that the CoUSD model has significant advantages in high-frequency detail recovery and multiphysics field coupling modeling under complex operating conditions, making it a promising tool for efficient design optimization and rapid performance evaluation of WPT systems.

The muzzle velocity of electromagnetic rail launchers approaches 1550 m/s, exhibiting typical hypervelocity electrical contact characteristics. During the electromagnetic launching process, extreme conditions, such as high current density, high temperature rise, and strong strain can cause wear on the surfaces of the armature and rail. Electromagnetic launch tests are conducted to study the wear conditions of the rail surface and the relationship between the wear state and contact resistance. After the rail is abraded by hundreds of launching armatures, its surface 2D profile and morphological characteristics are measured and analyzed. Based on fractal theory, the static contact resistance model is developed. Concurrently, the contact resistance at various positions is measured to reveal the evolution of the static contact resistance between the armature and the rail under wear. The research results show that along the direction of the armature launch, the rail surface wear transitions from mechanical wear to electrical wear, the fluctuation range of the 2D profile becomes smoother, and the roughness of the rail surface shows a decreasing trend. When the roughness is greater, the contact resistance is more sensitive to changes in external load.

Linear propulsion electromagnetic energy equipment utilizes the Lorentz force to apply power to the emitter, achieving launch speeds and efficiencies that far surpass those of traditional chemical propulsion technologies. This capability has significant strategic and practical value in both defense and civilian applications. Analyzing and calculating the armature electromagnetic thrust is essential for studying the dynamic characteristics of linear propulsion electromagnetic energy equipment. However, high-speed scenarios often present challenges in obtaining numerical solutions due to “pseudo-oscillation.” To address this issue, we propose a novel prediction method for the armature electromagnetic thrust in linear propulsion electromagnetic energy equipment using deep operator networks. First, a finite element simulation model of the system was developed to generate stable numerical simulation data under varying excitation currents. This model is then used to analyze the dynamic characteristics of the armature electromagnetic thrust and evaluate the stability of the numerical solutions. Based on the simulation data, we introduce a deep operator network-based prediction method for armature electromagnetic thrust. Finally, the prediction performance of the model is validated using linear propulsion electromagnetic energy equipment as a computational example. The results showed that the prediction error for the electromagnetic thrust on the test set was within 1%. After training, the model can calculate the electromagnetic thrust under different excitation currents in a matter of seconds, providing a fresh perspective for analyzing the dynamic characteristics of linear propulsion electromagnetic energy equipment.

Linear propulsion electromagnetic energy equipment can convert electromagnetic energy to kinetic energy instantaneously and has many advantages, such as high kinetic energy, efficiency, precision, and strong controllability. It surpasses traditional mechanical and chemical energy methods, significantly impacting various fields. Aiming at the numerical simulation method, which has problems with large computation volume and a long time of physical field simulation, the current field prediction method of linear propulsion electromagnetic energy equipment based on a Deep Generation Adversarial Operator Network is proposed. Firstly, deep operator network is combined with conditional generative adversarial network to obtain the knowledge-embedded conditional adversarial operator network (CGAONet) model. Then, Res-Transformer-Unet (RTUnet) is used as a branch network of CGAONet to establish the RTUnet-CGAONet model, and the current field method using a deep adversarial operator network is proposed. Finally, the finite element simulation model of the US public linear propulsion electromagnetic energy equipment calculation example is established to construct the simulation dataset from 1D excitation current value and time data to a 2D current field. The trained RTUnet-CGAONet model predicts the mean absolute percentage error of 2.94% in the 2D current field, and the model is minimally affected by the number of samples in the dataset. The results of this paper can achieve the second-level calculation of the current field under different excitation currents, which provides a new way of thinking for the analysis of dynamic characteristics of linear propulsion electromagnetic energy equipment.

Endowed with high sensitivity, fast response, and excellent flexibility, flexible piezoelectric sensors proffer substantial advantages for establishing efficacious contact with coarse object surfaces. They exhibit remarkable potential for application in wearable devices, motion state recognition, and device vibration monitoring. This paper introduces the development of a novel pliable piezoelectric nanocomposite harnessing modified lead zirconate titanate (PZT) and a flexible photosensitive resin. The PZT-based flexible piezoelectric sensors were fabricated using the light-curing molding technique. By establishing a sensor test platform, the performance parameters of the PZT-based sensors were obtained. Compared to conventional flexible piezoelectric sensors, the output voltage of the PZT-based sensors increased by 1.67 times under the same excitation, resulting in an improved response time and sensitivity. Moreover, the sensors demonstrated an excellent voltage output performance for recognizing human motion states and monitoring electrical equipment vibrations, both in real-time motion state changes and in core vibration under high-frequency operating conditions. These 3D-printed flexible piezoelectric sensors assume paramount significance across diverse domains, including motion science and real-time equipment condition monitoring.

Access full-text academic articles: J-STAGE is an online platform for Japanese academic journals.

TEAM Problem 35, a benchmark problem in the optimization of electrical equipment, is a multi-objective, multi-variable optimization problem involving uniform field optimization. This paper proposes a novel guided optimization algorithm to address this problem and achieve optimal magnetic field regulation. First, the deep learning method is used to predict the magnetic field of the solenoid; second, the generated magnetic field cloud map is screened to extract the magnetic field information of the more optimal solution with uniform magnetic field distribution by calculating the variance of the generated magnetic field map; finally, the magnetic field features are combined with the particle swarm optimization algorithm; specifically, the variance of the calculated magnetic field map is added to the objective function of the optimization algorithm to strengthen the optimization process. This integration improves the convergence speed of the optimization process and ensures the continuity and stability of the obtained Pareto solutions. To optimize the solenoid, this paper considers the uniformity of magnetic field distribution and power loss as the primary optimization objectives. The radius of the coil is chosen as the design variable. The optimization results are compared with those obtained by conventional optimization algorithms, demonstrating superior optimization performance.

The complex topology of dry-type hollow high-voltage reactors leads to significant inaccuracies in calculating inter-layer voltage distribution using traditional equivalent models during lightning impulse testing. In this study, the finite element method was employed to calculate the inductance and capacitance parameters of the reactor using the winding layers as subdivision units. The improved equivalent circuit model incorporates mutual capacitance between non-adjacent layers. Through a comparison of simulation and experimental results, the proposed improved equivalent circuit model achieves accurate simulation of the reactor’s voltage distribution. This research contributes to the insulation design of the reactor.