摘要
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。
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 improved 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 conditions.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.
出处
《无线电工程》
2015年第2期42-45,共4页
Radio Engineering
基金
国家自然科学基金资助项目(51277181)
关键词
电磁脉冲
TVS
系统辨识
ELMAN神经网络
electromagnetic pulse
transient voltage suppressor （TVS）
system identification
Elman neural network
