CAJ | 학술논문

针对传输线脉冲（Transmission Line Pulse， TLP）测试方法实施过程工作量较大、实验结果与实测数据吻合较差的问题，提出一种基于改进型Elman神经网络的电磁脉冲（Electromagnetic Pulse， EMP）响应建模方法。在TLP方法基础上增设机器模型静电放电和人体金属模型静电放电2类电磁脉冲，利用隐含层神经元数目为10的改进型Elman神经网络对NUP2105L型瞬态抑制二极管（ Transient Voltage Suppressor， TVS）的实验数据进行建模，并预测不同脉冲条件下TVS的响应。仿真结果表明，该方法建模精度高、泛化能力强，能够定量判断TVS性能，满足电路快速选件需要。
침대전수선맥충（Transmission Line Pulse， TLP）측시방법실시과정공작량교대、실험결과여실측수거문합교차적문제，제출일충기우개진형Elman신경망락적전자맥충（Electromagnetic Pulse， EMP）향응건모방법。재TLP방법기출상증설궤기모형정전방전화인체금속모형정전방전2류전자맥충，이용은함층신경원수목위10적개진형Elman신경망락대NUP2105L형순태억제이겁관（ Transient Voltage Suppressor， TVS）적실험수거진행건모，병예측불동맥충조건하TVS적향응。방진결과표명，해방법건모정도고、범화능력강，능구정량판단TVS성능，만족전로쾌속선건수요。
In view of large workload and poor consistency between experimental results and actual testing data,when using the transmission line pulse (TLP) testing methods,an electromagnetic pulse (EMP) response modeling method is proposed based on im?proved Elman neural network,which is added with such two types of EMP as machine model electrostatic discharge ( ESD) EMP and human metal model ESD EMP on the basis of TLP method.The improved Elman neural network whose hidden layer has 10 neurons,is used for modeling of NUP2105L transient voltage suppressor ( TVS) ,and then predicting the response of TVS under different EMP con?ditions.The simulation results show that the method has high modeling accuracy,strong generalization capability.It can determine the quantitative performance of TVS and options to quickly meet the needs of circuit.