Modeling and Application of Electromagnetic Field Finite Element Simulation Software MBSE for Electrical Equipment Integrated with AI
Jin Liang1,2, Jia Yufang1, Liu Lu1, Guo Shaonan1, Ma Tianci1
1. State Key Laboratory of Intelligence Power Distribution and System Equipment Hebei University of Technology Tianjin 300401 China; 2. Key Laboratory of Electromagnetic Field and Electrical Apparatus Reliability of Hebei Province Hebei University of Technology Tianjin 300401 China
摘要 电磁场有限元仿真软件在电磁设计和分析中至关重要。然而商业软件价格高昂且功能冗余,同时随着人工智能（AI）的发展,深度学习与传统数值模拟的结合提升了仿真精度和效率。为此,亟须开发自主可控、轻量化、面向企业定制的智能仿真软件。该文首先提出了双向耦合的设计方法,使仿真软件能适应复杂需求。然后,基于模型系统工程（MBSE）方法,建立四个层级的电磁场有限元仿真软件的架构设计流程,并完成软件的建模,实现了系统结构和行为的定义和可视化。最后,基于Python开发了集成AI的软件原型IFEM,并通过两个实际案例验证了其功能和准确性。研究成果提高了软件开发的质量、效率和知识传承能力,为数值模拟工业软件在AI for Science背景下的AI集成提供了有效思路。
Abstract：Electromagnetic field finite element simulation software is crucial for electromagnetic design and analysis, particularly in the performance evaluation of complex electrical equipment, which often relies on costly foreign software with high barriers to entry. As the demand for electromagnetic field simulations continues to increase, existing commercial software fails to meet the specific customization needs of various engineering projects. It creates an urgent need to develop autonomous, lightweight, and intelligent simulation tools. This paper introduces the development of a prototype intelligent finite element simulation software, IFEM, which integrates artificial intelligence (AI) technology using a model-based systems engineering (MBSE) approach to create a visual software model. Firstly, this paper proposes a bidirectional coupling software design method consisting of two key aspects: the bidirectional coupling between software developers and users as well as between numerical simulation and AI. The bidirectional coupling between developers and users aims to enable efficient information exchange, ensuring flexibility and adaptability in the software design process. Meanwhile, the bidirectional coupling between numerical simulation and AI retains the accuracy of numerical simulation while leveraging the efficiency of AI. A four-layer architectural design process for the electromagnetic finite element simulation software is constructed, including operational analysis, system analysis, logical architecture, and software architecture. A graphical model represents the complex software structure, significantly enhancing transparency and improving collaboration among developers. During the development of the IFEM prototype, the bidirectional coupling of numerical simulation and AI is implemented using Python programming language and the PyCharm development environment, unifying the development and debugging of finite element and AI modules. It has the advantages of reduced programming workload, rich and user-friendly algorithm libraries, easy prototyping, and seamless AI integration. Finally, a solver based on the node finite element method and Physics-Informed Neural Networks (PINN) is used to calculate the potential distribution between two-dimensional electrodes. Post-processing generates the electric field strength discrete values, field distribution maps, and contour lines. A comparison with COMSOL results shows that the mean absolute percentage errors (MAPE) for the potential and electric field strength are2.735×10-11 and 4.185×10-11, respectively. Additionally, an improved LinkNet model, integrated with the node finite element method, is used to optimize the magnetic field distribution of a solenoid. Compared to COMSOL, the MAPE for the 2D constant magnetic field vector potential and magnetic flux density magnitudes are0.054% and 2.889%, respectively. By replacing the finite element model with the improved LinkNet model, the Pareto front for solenoid optimization under two objectives is obtained. After training, the improved LinkNet model generates 250 samples in 7.5 seconds, whereas the finite element method takes 5 minutes and 43 seconds to generate the same number of samples. Through case validation, IFEM demonstrates excellent computational accuracy and efficiency in solving problems such as 2D electrostatic fields and solenoid magnetic field optimization. The software also shows strong integration and scalability with AI algorithms. By adopting MBSE-based modeling, the transparency of the software design process and communication efficiency within the team are greatly improved, allowing for effective management and resolution of potential issues early in development. This approach has significantly enhanced the development efficiency of the electromagnetic field finite element simulation software. This paper provides effective methodologies for AI integration within specialized software development. Future research will focus on the software’s scalability and applicability enhancement, exploring complex engineering scenarios’ electromagnetic field simulation and optimization.
金亮, 贾雨方, 刘璐, 郭劭男, 马天赐. 集成AI的电工装备电磁场有限元仿真软件MBSE建模与应用[J]. 电工技术学报, 2025, 40(16): 5186-5203.
Jin Liang, Jia Yufang, Liu Lu, Guo Shaonan, Ma Tianci. Modeling and Application of Electromagnetic Field Finite Element Simulation Software MBSE for Electrical Equipment Integrated with AI. Transactions of China Electrotechnical Society, 2025, 40(16): 5186-5203.
[1] 程书灿, 赵彦普, 张军飞, 等. 电力设备多物理场仿真技术及软件发展现状[J]. 电力系统自动化, 2022, 46(10): 121-137. Cheng Shucan, Zhao Yanpu, Zhang Junfei, et al.State of the art of multiphysics simulation technology and software development for power equipment[J]. Auto- mation of Electric Power Systems, 2022, 46(10): 121-137. [2] 云道智造. 伏图(Simdroid)通用多物理场仿真PaaS平台[EB/OL]. https://www.ibe.cn/products/simdroid. [3] 英特仿真. 多物理场仿真及优化平台INTESIM- MultiSim[EB/OL]. http://www.intesim.cn/Jmksh. [4] 韩志仁, 罗雪磊. 基于ANSYS Workbench的型材拉弯有限元仿真模块开发[J]. 航空制造技术, 2020, 63(13): 64-68. Han Zhiren, Luo Xuelei.Development of finite element simulate module of ANSYS Workbench on profile stretch-bending[J]. Aeronautical Manufacturing Technology, 2020, 63(13): 64-68. [5] Hinton G E, Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507. [6] Miao Qinghai, Zheng Wenbo, Lü Yisheng, et al.DAO to HANOI via DeSci: AI paradigm shifts from AlphaGo to ChatGPT[J]. IEEE/CAA Journal of Automatica Sinica, 2023, 10(4): 877-897. [7] Yu Hui, Liang Wei, Fan Lili, et al.Sora for social vision with parallel intelligence: social interaction in intelligent vehicles[J]. IEEE Transactions on Intelligent Vehicles, 2024, 9(3): 4240-4243. [8] Zhang Weiwei, Li Xintao, Ye Zhengyin, et al.Mechanism of frequency lock-in in vortex-induced vibrations at low Reynolds numbers[J]. Journal of Fluid Mechanics, 2015, 783: 72-102. [9] Gao Chuanqiang, Zhang Weiwei, Li Xintao, et al.Mechanism of frequency lock-in in transonic buffeting flow[J]. Journal of Fluid Mechanics, 2017, 818: 528-561. [10] 金亮, 王飞, 杨庆新, 等. 永磁同步电机性能分析的典型深度学习模型与训练方法[J]. 电工技术学报, 2018, 33(增刊1): 41-48. Jin Liang, Wang Fei, Yang Qingxin, et al.Typical deep learning model and training method for per- formance analysis of permanent magnet synchronous motor[J]. Transactions of China Electrotechnical Society, 2018, 33(S1): 41-48. [11] 金亮, 尹振豪, 刘璐, 等. 基于残差U-Net和自注意力Transformer编码器的磁场预测方法[J]. 电工技术学报, 2024, 39(10): 2937-2952. Jin Liang, Yin Zhenhao, Liu Lu, et al.Magnetic field prediction method based on residual U-Net and self-attention Transformer encoder[J]. Transactions of China Electrotechnical Society, 2024, 39(10): 2937-2952. [12] 魏蓉, 陈锦培, 仲林林. 基于深度算子网络的电磁轨道发射速度趋肤效应的快速计算方法[J]. 电工技术学报, 2025, 40(5): 1344-1354. Wei Rong, Chen Jinpei, Zhong Linlin.A fast computational method for velocity skin effect of electromagnetic rail launch based on deep operator networks[J]. Transactions of China Electrotechnical Society, 2025, 40(5): 1344-1354. [13] Jin Liang, Liu Yuankai, Yang Qingxin, et al.Prediction method of motor magnetic field based on improved Linknet model[J]. COMPEL-the Inter- national Journal for Computation and Mathematics in Electrical and Electronic Engineering, 2023, 42(1): 90-100. [14] Khan A, Ghorbanian V, Lowther D.Deep learning for magnetic field estimation[J]. IEEE Transactions on Magnetics, 2019, 55(6): 7202304. [15] Khan A, Mohammadi M H, Ghorbanian V, et al.Efficiency map prediction of motor drives using deep learning[J]. IEEE Transactions on Magnetics, 2020, 56(3): 7511504. [16] 李青, 张新燕, 马天娇, 等. 基于SSA-CNN-BiGRU- Attention的超短期风电功率预测模型[J]. 电机与控制应用, 2023, 50(5): 61-71. Li Qing, Zhang Xinyan, Ma Tianjiao, et al.Ultra- short term forecasting model of wind power based on SSA-CNN-BiGRU-attention[J]. Electric Machines & Control Application, 2023, 50(5): 61-71. [17] 张宇娇, 孙宏达, 赵志涛, 等. 基于物理信息神经网络的电磁场计算方法[J]. 电工技术学报, 2024, 39(17): 5251-5261. Zhang Yujiao, Sun Hongda, Zhao Zhitao, et al.Electromagnetic field calculation method based on physical informed neural network[J]. Transactions of China Electrotechnical Society, 2024, 39(17): 5251-5261. [18] 张宇娇, 赵志涛, 徐斌, 等. 基于U-net卷积神经网络的电磁场快速计算方法[J]. 电工技术学报, 2024, 39(9): 2730-2742. Zhang Yujiao, Zhao Zhitao, Xu Bin, et al.Fast calculation method of electromagnetic field based on U-net convolutional neural network[J]. Transactions of China Electrotechnical Society, 2024, 39(9): 2730-2742. [19] 王怀民, 吴文峻, 毛新军, 等. 复杂软件系统的成长性构造与适应性演化[J]. 中国科学: 信息科学, 2014, 44(6): 743-761. Wang Huaimin, Wu Wenjun, Mao Xinjun, et al.Growing construction and adaptive evolution of complex software system[J]. Scientia Sinica: Infor- mationis, 2014, 44(6): 743-761. [20] Henderson K, Salado A.Value and benefits of model- based systems engineering (MBSE): evidence from the literature[J]. Systems Engineering, 2021, 24(1): 51-66. [21] 董梦如, 王国新, 鲁金直, 等. 基于WordCloud技术的MBSE发展态势研究[J]. 系统工程与电子技术, 2024, 46(2): 534-548. Dong Mengru, Wang Guoxin, Lu Jinzhi, et al.Research on the development trend of MBSE based on WordCloud technology[J]. Systems Engineering and Electronics, 2024, 46(2): 534-548. [22] 王文跃, 侯俊杰, 毛寅轩, 等. 面向复杂产品研制的MBSE体系架构及其发展趋势研究[J]. 控制与决策, 2022, 37(12): 3073-3082. Wang Wenyue, Hou Junjie, Mao Yinxuan, et al.MBSE architecture for complex product development and trends[J]. Control and Decision, 2022, 37(12): 3073-3082. [23] 薛芳芳, 王亮亮, 缪炜涛, 等. 基于MBSE的民机飞行管理软件设计[J]. 航空计算技术, 2019, 49(5): 111-116. Xue Fangfang, Wang Liangliang, Miao Weitao, et al.Design of airborne flight management software based on MBSE[J]. Aeronautical Computing Technique, 2019, 49(5): 111-116. [24] 吕冠甫. 基于MBSE的无人机地面站应用系统软件设计与实现[D]. 太原: 中北大学, 2020. Lü Guanfu.Software design and implementation of UAV ground station application system based on MBSE[D]. Taiyuan: North University of China, 2020. [25] Yu Chao, Li Qing, Liu Kui, et al.Industrial design and development software system architecture based on model-based systems engineering and cloud computing[J]. Annual Reviews in Control, 2021, 51: 401-423. [26] 侯淑萍, 杨庆新, 陈海燕, 等. 新型高性能电工材料应用特性建模的模块化设计[J]. 电工技术学报, 2008, 23(10): 1-5. Hou Shuping, Yang Qingxin, Chen Haiyan, et al.Modularity design of modeling for new high per- formance materials applied to electrical engineering[J]. Transactions of China Electrotechnical Society, 2008, 23(10): 1-5. [27] 王雨农, 毕文豪, 张安, 等. 基于DoDAF的民机MBSE研制方法[J]. 系统工程与电子技术, 2021, 43(12): 3579-3585. Wang Yunong, Bi Wenhao, Zhang An, et al.DoDAF- based civil aircraft MBSE development method[J]. Systems Engineering and Electronics, 2021, 43(12): 3579-3585. [28] Izygon M, Wagner H, Okon S, et al.Facilitating R&M in spaceflight systems with MBSE[C]//2016 Annual Reliability and Maintainability Symposium (RAMS), Tucson, AZ, USA, 2016: 1-6. [29] 向超群, 尹雪瑶, 伍珣, 等. 基于物理信息神经网络的牵引变流器直流支撑电容参数辨识方法[J]. 电工技术学报, 2024, 39(15): 4654-4667. Xiang Chaoqun, Yin Xueyao, Wu Xun, et al.Parameter identification of DC-link capacitor in traction converter based on physical information neural network[J]. Transactions of China Electro- technical Society, 2024, 39(15): 4654-4667. [30] Ma Yaoyao, Xu Xiaoyu, Yan Shuai, et al.Extraction of interconnect parasitic capacitance matrix based on deep neural network[J]. Electronics, 2023, 12(6): 1440. [31] Bashatah J, Sherry L.Usability analysis of an MBSE model of standard operating procedures[J]. INCOSE International Symposium, 2022, 32: 38-47. [32] De Saqui-Sannes P, Vingerhoeds R A, Garion C, et al. A taxonomy of MBSE approaches by languages, tools and methods[J]. IEEE Access, 2022, 10: 120936-120950. [33] Biggs G, Juknevicius T, Armonas A, et al.Integrating safety and reliability analysis into MBSE: overview of the new proposed OMG standard[J]. INCOSE International Symposium, 2018, 28(1): 1322-1336. [34] De Pra Y, Fontana F.Programming real-time sound in Python[J]. Applied Sciences, 2020, 10(12): 4214. [35] 金亮, 刘璐, 贾雨方. Python语言下长直导线电场计算代码[EB/OL]. https://gitee.com/fang-xiaojia1123/ifem. [36] Di Barba P, Mognaschi M E, Lowther D A, et al.A benchmark TEAM problem for multi-objective Pareto optimization of electromagnetic devices[J]. IEEE Transactions on Magnetics, 2018, 54(3): 9400604.
2025, Vol. 40 
