The temperature distributions of clustered power cables are different from those of single power cables or single-circuit cable transmission lines due to skin effects and proximity effects of eddy current and Joule losses. Nevertheless, the temperature is one of the main parameters in determining the current rating of power cables. In order to derive the distribution pattern of the temperature field, finite element models and solution methodologies of coupled eddy current-temperature field for buried power cables in clustered layout are proposed, and the effects of the spacing between phases and layers, and the phase-conductor arrangement on the temperature field are studied. The numerical results show that under the conventional arrangement of phase conductors, the maximum temperature of bundle section decreases with increasing spacing between layers, while the relationship between the maximum temperature and the spacing of phase conductors is dependent on the grounding mode of the metal sheaths. In the case of single-end grounding mode, the maximum temperature of bundle section decreases with increasing the spacing between phase conductors, while in the case of both-ends grounding mode, the maximum temperature increases with increasing spacing between phase conductors. The distribution of phase conductors can significantly influence the maximum temperature of the bundle section under the both-ends grounding mode. The more even the distribution of phase conductors is, the lower the maximum temperature is. The numerical results and conclusions as reported in this paper thus provide guidelines for engineering applications.

Antennas are fundamental components of modern communication systems. Antenna optimizations are imperative to achieve a high-performance communication solution. With antenna evolving towards multifunction and systematization, antenna designs are often involved with a large scale of optimization variables and high-dimensional objectives. A decomposition algorithm incorporated with reinforcement learning is proposed to solve high-dimensional many-objective antenna optimization problems. To enhance the evolutionary pressure in high-dimensional objective comparison, the aggregation tree method is adopted to remove the non-conflict or weak-conflict objectives. To search the variable space efficiently, optimization variables are divided into several groups, which are checked and updated based on Q-learning method. A filtering antenna is designed and optimized to verify the effectiveness and efficiency of the proposed algorithm. The numerical and experimental results demonstrate that the proposed decomposition method can obtain qualified and satisfactory design solutions for complex antenna designs.

Accurate soil thermal resistivity is crucial for real-time cable ampacity determination to maximize cable utilization. However, the determination of soil thermal resistivity involves solving a computationally expensive multi-physical field inverse problem where a high-fidelity model (HFM) is used for performance evaluations. Surrogate-assisted evolutionary algorithms (SAEAs) are computationally efficient for such problems; model management strategies (MMSs) are key to SAEAs. Nevertheless, most MMSs struggle to balance the computational cost and the search accuracy due to their reliance on fitness value errors. In fact, maintaining a similar function landscape between the surrogate and the HFM is more essential than achieving precise fitness values on the surrogate. Consequently, a rank-based hybrid MMS-driven two-stage SAEA is proposed. Stage 1 focuses on identifying promising regions. To ensure the similarity between the surrogate and HFM function landscape and thus guide the evolution accurately, a global MMS is proposed. Specifically, a new function landscape similarity metric is proposed to adaptively adjust the surrogate update frequency. A new rank-error-based individual selection strategy selects key individuals for exact evaluations to refine the surrogate similarity. Stage 2 performs a refined local search within the identified promising region, utilizing a local MMS to re-evaluate the optimum of a local surrogate built around the best solution searched in Stage 1. Optimization results confirm the proposed method's superiority on test functions and a prototype cable.

Efficient and accurate calculation of power frequency electric fields in transmission lines in the design phase is of great significance for enhancing the safety of power systems. The Charge Simulation Method (CSM) is a widely used approach in this direction. However, existing CSM-based approaches suffer from oversimplified transmission line models, pronounced end effects, and limited accuracies. To address these limitations, this paper proposes an improved model for three-dimensional electric field calculation in transmission lines based on CSM, utilizing a Competitive Swarm Optimizer (CSO) algorithm to enhance the model accuracy. Firstly, the proposed model discretizes the catenary into segmented line charges as the fundamental charge units to improve the modeling fidelity. Secondly, based on the high-dimensional characteristics of the model, a Large-Scale Global Optimization (LSGO) problem is further formulated, and the CSO algorithm is used to solve the optimal arrangement of the image line charges. Finally, the total spatial electric field is synthesized from the charge distributions. Experimental results indicate that the proposed method improves the computational accuracy when applying CSM to power-frequency electric field analysis in transmission lines.

Low solution accuracy and efficiency are two bottleneck problems in the existing models and methodologies for spatial distance calculations to verify the minimal electrical clearance of overhead transmission lines if a dynamic windage yaw is considered. To address these two issues, the accurate numerical models and the corresponding efficient solution methodologies tailored for different scenarios are proposed. First, a conductor windage yaw surface model incorporating a horizontal specific load coefficient is established, transforming the wire-to-wire minimal distance determination into a multi-dimensional nonlinear constrained optimization problem. An improved gradient-guided crossover genetic algorithm (GGA) is subsequently developed to solve this optimization problem. By integrating the gradient information to guide the crossover operator and combining an adaptive mutation with a dimension mutation strategy, the solution efficiency is enhanced. For the wire-to-tower minimal distance determination, a simplified tower model and a hybrid optimization methodology combining an oriented octree with the GGA are proposed. Numerical results on typical case studies show that, for a wire-to-wire minimal distance calculation, the GGA outperforms both the basic genetic algorithm and particle swarm optimization in terms of both convergence speed and solution accuracy. For a wire-to-tower minimal distance calculation, the oriented octree improves the spatial utilization, and the proposed hybrid methodology substantially improves the computational performance.

A small high-intensity neutron generator imposes constraints on the beam extraction and acceleration system, which can deliver a deuterium/tritium (D/T) mixed ion beam with an energy of more than 200 keV and a current of more than 126 mA to the target within a limited space of 500 mm in length and 60 mm in radius. The beam at the target should have a spot radius of less than 40 mm and a peak current density of less than 50 A/m2. In this work, beam transport simulations were conducted using the IBSIMU code. A preliminary design was obtained by iteratively optimizing the electrode geometry, which enables the transportation of a 200 keV, 126 mA D-T beam with a spot radius of 37 mm and a peak current density of 49.2 A/m2 at the target. Based on this design, the effects of key geometric parameters, including the extraction gap, extraction aperture radius, acceleration gap, and acceleration aperture radius, on the beam spot radius and peak current density, were systematically analyzed. The results indicate that the variations in the extraction and acceleration gaps significantly affect the beam focusing condition, thus exerting a strong influence on the beam spot size and beam distribution. Under-focused transport conditions are more favorable for meeting the design requirements of the neutron generator. Variations in the extraction aperture radius and acceleration aperture radius do not modify the beam focusing condition and only marginally affect the beam spot and density, thereby allowing fine adjustments to be made according to practical requirements.

Addressing the issue of transformer DC bias induced by metro stray current and the limited generalization capability of conventional diagnostic approaches, this paper introduces a multi-parameter integrated diagnostic methodology that amalgamates neutral point current, vibration, and voiceprint characteristics. The equivalent DC is identified as the pivotal metric, and a feature screening framework is constructed, which incorporates classified Min-Max normalization and coefficient of variation weighting. Furthermore, a nonlinear exponential mapping model is established to quantify the comprehensive impact of bias. Experimental findings substantiate the efficacy and robust generalization of the proposed methodology.

This article presents a novel, low-cost metasurface (MS)-based energy harvester for high-selectivity radio frequency energy harvesting (RFEH). Each cell of the proposed MS integrates four interconnected spoof localized surface plasmon (SLSP) ring resonators in an air-gapped asymmetric topology, using only a single port. A compact, dual-band, low-input-power rectifier is also developed to construct an RFEH microsystem. The main contributions of this article include: 1) a high-selectivity energy harvesting across multiple widebands under two controlled linear-polarization (LP) modes, enabling flexible operations in complex radio frequency (RF) environments; 2) a single-port-per-cell configuration, reducing the integration complexity; and 3) an SLSP-based folded multilayer structure with a low-cost, air-gapped topology, enhancing both the compactness and affordability. Systematic measurement results on a fabricated 4x4 MS array with the rectifier reveal: 1) high harvesting efficiency and strong selectivity of the MS cell at 1.78-2.40/2.61-2.95 GHz (Mode 1) and 3.49-3.96 GHz (Mode 2); 2) maximum load absorption efficiencies of 81.70% (Mode 1) and 86.33% (Mode 2), with corresponding fractional bandwidths (FBWs) of 49.5% and 12.6%; 3) stable performance of the MS cell under oblique incidence angles of 0 degrees-30 degrees; and 4) wireless energy harvesting (WEH) efficiencies of 55.97% (Mode 1) and 65.47% (Mode 2) for the integrated system at an incident power density of 0.87 mu W/mm(2). The experimental results show good agreement with the simulated ones, validating the feasibility and superior performance of the proposed MS harvester.

To mitigate new energy fluctuations,the joint optimization and scheduling of different units as a unified generating system is generally applied.In this study,to address the impact of uncertainties on scheduling plans,research on optimizing scheduling of new energy under uncertain conditions is necessary.In this point of view,this study proposes a distributionally robust optimization method based on Wasserstein distance to tackle the uncertainty in photovoltaic outputs.The proposed methodology first constructs an uncertainty set based on the historical output data of photovoltaic power plants,and converts the original model into a mixed-integer linear model easy to solve by using the duality theory and Karush-Kuhn-Tucher(KKT)conditions.The converted model is then solved by using column-and-constraint generation(CCG)algorithm.Finally,numerical experiments on a comprehensive system comprising multiple units to compare the optimization results of the deterministic model,the robust optimization model and the distributionally robust optimization model are given,demonstrating the effectiveness and the superiority of the proposed distributionally robust optimization model.

It is demanding to maximize the power cable ampacity to make full usage of the cluster. Since the maximization of the cable ampacity in a cable cluster is a hard constrained global optimization problem, and the final solution may be located on the constraint boundary; the existing optimization methodology will incur deficiencies in pinpointing the global optimal solution efficiently due to the underuse of the infeasible solutions. To use infeasible solutions to pinpoint exactly the global optimal solution, a hybrid optimizer based on an improved genetic algorithm and a local search adaptive tabu search (IGA-ATS) is proposed. In the proposed optimizer, the population is divided into two subpopulations: a feasible one to drive the search toward promising region while an infeasible one to stimulate to pinpoint the potential solutions on the constraint boundary. Consequently, two different criterions are designed to criticize individuals in the two subpopulations, and the corresponding selection strategies are proposed. Moreover, a boundary intensity searching strategy is proposed to drive promising infeasible individuals clustering to the potential constraint boundary efficiently. The superiorities of the proposed method are confirmed by the numerical results on a test function and a case study.

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.

In this paper, a multi-objective topology optimization methodology which incorporates a novel hybrid multi-objective optimization (MOO) algorithm with a material-field series expansion (MFSE) method is proposed. The hybrid algorithm is designed by combining a beluga whale optimization (BWO) with an improved NSGA-II where the original BWO is transformed to a MOO version and NSGA-II is enhanced by using opposite-learning initialization and revised crowding-distance strategy. To alleviate the computation burden, the MFSE method is adopted to reduce the number of design variables. The whole proposed methodology is validated by using an electromagnetic actuator prototype. Numerical results demonstrate the effectiveness of the proposed methodology.

Topology design is the concept design of a product, which can create a novel topology of the product that might be “unthinkable” beforehand. Topology Optimization (TO) is thus increasingly recognized as the paradigm of the predominant engineering techniques to provide a quantitative design method for modern engineering design. Nevertheless, the extremely high dimensionality and overwhelmingly heavy computational burden are the main factors restricting the implementation of TOs in engineering practice. To address the aforementioned issues, a TO methodology based on material-field series expansion (MFSE) is proposed. Based on the MFSE, the structural topology is represented by a newly-introduced unbounded material field with a spatial dependency and the TO problem is approximated as a constrained mathematical programming problem with the truncated series expansion coefficients as the design variables, significantly reducing the dimensionality of design variables and inherently avoiding the checkerboard pattern without any additional measure. The numerical example is given to validate the feasibility and effectiveness of the proposed methodology and demonstrate its advantage for dimensionality and computational cost reductions in TO designs.

In avoiding the inconvenient and costly maintenance and replacement of the conventional batteries in large-scale and low-power wireless sensor applications, radio frequency (RF) energy harvesting shows exclusive potential due to its abundant availability and resilience in environmental conditions. This study introduces firstly a novel metasurface (MS) absorber topology employing four sequentially rotated spoof-local-surface-plasmons (SLSPs) resonators to achieve a multi-band energy harvesting and a polarization insensitivity. To realize an efficient and accurate design optimization of the proposed MS absorber, a customized multi-stage collaborative machine learning-assisted optimization methodology, incorporating three different fidelity levels, the nonlinear autoregressive multi-fidelity Gaussian Process (NARGP), and the multi-task Gaussian Process (MTGP), is then proposed. A prototype MS absorber is finally optimized and fabricated. Both the numerical and the experimented results validate the presented works.

Optimization algorithms play a critical role in electromagnetic device designs due to the ever-increasing technological and economical competition. Although evolutionary algorithm-based methods have successfully been applied to different design problems, these methods exhibit deficiencies when solving complex problems with multimodal and discontinuous objective functions, which is quite common in electromagnetic device optimization designs. In this paper, a hybrid multi-objective optimization algorithm based on a non-dominated sorting genetic algorithm (NSGA-II) and a multi-objective particle swarm optimization method (MOPSO) is proposed. In order to enhance the convergence and diversity performance of the algorithm, a new population update mechanism of MOPSO is introduced. Moreover, an adaptive operator involving crossover and mutation is presented to achieve a better balance between global and local searches. The performance of the hybrid algorithm is validated using standard test functions and the multi-objective design of a superconducting magnetic energy storage (SMES) device. Numerical results demonstrate the effectiveness and superiority of the proposed method.

To tackle the nonlinear, multimodal, and computationally expensive challenges inherent in the optimal design of RF energy harvesters, an innovative hybrid algorithm combining particle swarm optimization and fireworks algorithm is proposed. In the proposed hybrid algorithm, 1) a stretching-and-ejection mechanism is developed to balance exploration and exploitation searches, 2) a self-adaptive inertia weight is introduced to enhance the convergence rate, 3) an elite- preservation strategy is employed to guarantee a global convergence. Also, a new topology of a wideband and dual-polarized metasurface absorber for RF energy harvesting applications is introduced. The feasibilities of the presented work in RFEH applications and design optimizations are validated using a case study.

A surrogate-assisted evolutionary algorithm based on an artificial neural network (ANN) is proposed for computationally expensive multi-objective optimization problems of electromagnetic devices. The training data of the ANN are selected from the current population according to the Pareto rank and the crowding distance of individuals to learn cumulative knowledge from the searched solutions in the evolution. The reproduction mechanism, alternatively using the trained ANN and the traditional genetic operators, is designed for new offspring. Finally, the performance of the proposed algorithm is tested on a case study. The optimized results show that the proposed algorithm is more competitive than the existing multi-objective evolutionary algorithms.

To solve the computationally heavy inversion of the soil thermal resistivity and the real time ampacity of power cable, a reinforcement learning-based two-modes (Mode 1 and Mode 2) multi-surrogate-assisted evolutionary algorithm is proposed. To address the inefficiencies of existing surrogate-assisted evolutionary algorithms; 1) A new surrogate selection methodology is proposed to enhance the robustness and the accuracy of the predicted results of the surrogates. This method employs different surrogates without any weight to finally compromise the results of the models. Specially, two distinct global and local surrogates are built in Mode 1 and Mode 2, respectively. In each mode, a surrogate will be adaptively chosen by their performance; 2) A hybrid fitness uncertainty estimation methodology combining three distinct indicators is proposed to achieve a more accurate and robust uncertainty estimation, effectively improving the surrogate accuracy; 3) An adaptive evolutionary operator selection strategy (AEOSS) based on reinforcement learning is designed for Mode 1. In the searching process of Mode 1, each individual selects an evolutionary operator based on the AEOSS. In Mode 2, an exploitation-biased operator is used to all individuals to accelerate the convergence. The superiorities of the proposed method are confirmed by the optimization results of test functions and a prototype cable.

The rotor containment sleeve brings serious deficiencies in heat dissipations of a high-speed machine with surface-mounted magnets while preventing the centrifugal force generated by the high-speed rotation of the rotor from fracturing the permanent magnet (PM). To address the aforementioned issues, this paper explores the influence of rotor containment sleeves on the comprehensive performance of a high-speed PM generator. In this respect, the electromagnetic losses, temperature distribution and rotor strength considering pole fillers under different material fabricated sleeves are compared and analyzed by electromagnetic field, coupled three-dimensional fluid-thermal field and stress field analysis. The numerical results show that the PM with a copper-embedded composite sleeve has a higher margin factor under a high-speed rotation at a high temperature, while the rotor with a carbon fiber sleeve performs better in view of eddy current losses and temperature rises.

Metamaterial (MM) is very promising in engineering application since it exhibits extraordinary physical properties that do not exist in nature. Nevertheless, the development of a MM still faces bottleneck problems such as to maximize negative permeability and ensure the robustness of the high permeability at the working frequency in engineering applications. To address the inefficiencies of existing multi-objective robust optimization methodologies in applications to MM designs, an improved multi-objective genetic algorithm and an adaptive response surface model are proposed. The numerical optimization results of a prototype MM unit have demonstrated the feasibility and merits of the proposed methodology.