Simulation Application of Interconnected Structure of Microsystems Based on Artificial Neural Network Technology

This paper mainly describes the construction process of a class of electromagnetic model simulation APP for interconnected structures of Microsystems. The driving algorithm of simulation application uses artificial neural network algorithm instead of the original electromagnetic numerical algorithm (finite element method, finite difference method, moment method). The algorithm model constructed by artificial neural network can ensure the accuracy and reliability, and the computing speed is significantly improved. The trained algorithm model is encapsulated into a simulation application through customized application design environment, which can also be easily used by microsystem designers without intelligent algorithm basis. At the initial stage of microsystem design, microsystem design engineers apply simulation application to carry out design and simulation, which can simply set geometric parameters of interconnected structures and obtain simulation result curves in a very short time. Application also supports inverse calculation for design objectives, which can greatly save the design time.