Support vector machine in application of modeling of electromagnetic pulse field coupling on electronic devices

The vulnerability of electronic devices to EMP (Electromagnetic Pulse) has been increased, and the assessment of EMP effects is the key to solving such problems. Furthermore, the simulation of EMP effects on electronic devices is an important means of assessing their effects. Targeting the universally existing complex computing process problems in mechanism modeling algorithm ubiquity, it have been researched that SVM (Support Vector Machine) in application of modeling algorithm of EMP field coupling on electronic devices. First of all, through the CST (Computer Simulation Technology) microwave studio the simulation data of the small monopole antenna in the 4ns rectangular pulse and Gaussian pulse field radiation are obtained, and SVR (Support Vector Regression) modeling of transfer function of energy coupling is used. Then through the established model the coupling voltage and current of the small monopole antenna in the types of fields are simulated and predicted. By comparing the two important parameters of regression capacity of characterized model - the GOF(Goodness of Fit) and the MSE(Mean Square Error), the result shows, compared with the ARX system identification, SVR with CST simulation data can obtain a higher GOF and proves to be an effective means of engineering simulation modeling with a smaller MSE and simplified computing process.