在向着高技术参数的发展过程中，现代电工装备的电磁特性分析与优化成为一个亟待解决的计算难度大、多时空尺度、强耦合和非线性问题。传统方法一般采用简化模型降低对计算资源的需求实现电磁特性分析，以及依靠人经验在优化进行过程中实施必要干预实现优化，造成计算时间、结果精度与可信度、实用性等方面无法令人满意，甚至无法完成电磁特性分析与优化。基于此，本项目一方面，基于云计算实现电工装备电磁特性的精细模拟与快速计算，提高计算结果精度与可信度、减少计算时间；另一方面，考虑到深度学习可以模拟人在思考过程中的科学行为和作用机理，因此开展基于深度学习的电工装备参数的深层次特征提取方法和决策判定的研究，减少对人经验和样本数的依赖，建立电磁特性云计算分析与自学习优化方法，改善可信度、实用性和效率差的问题，为解决我国电工装备容量等级高、制造规模大而设计优化水平低的“不匹配”问题做出实质性贡献。

In the process of developing toward high-tech parameters, the analysis and optimization of electromagnetic characteristics of modern electrical equipment has become a difficult, multi-temporal scale, strong coupling and non-linear problem to be solved urgently. Traditional methods generally use simplified models to reduce the demand for computing resources to achieve electromagnetic characteristics analysis, and implement necessary interventions to achieve optimization based on human experience in the process of optimization, resulting in unsatisfactory calculation time, accuracy and reliability of results, practicability, etc., or even unable to complete electromagnetic characteristics analysis and optimization. Based on this, this project, on the one hand, realizes the fine simulation and fast calculation of electromagnetic characteristics of electrical equipment based on cloud computing, improves the accuracy and reliability of calculation results and reduces calculation time; on the other hand, considering that in-depth learning can simulate the scientific behavior and action mechanism of human beings in the process of thinking, it carries out deep-level feature extraction of electrical equipment parameters based on in-depth learning. The research on methods and decision-making can reduce the dependence on human experience and sample number, establish cloud computing analysis and self-learning optimization method of electromagnetic characteristics, improve the problems of low reliability, practicability and efficiency, and make substantial contributions to solving the "mismatch" problem of high capacity level, large manufacturing scale and low design optimization level of electrical equipment in China.