高斯-牛顿法在模拟特征分析测试参数优化中的应用

Application of Gauss-Newton method in analog signature analysis and testing parameter optimization

普通的模拟特征分析测试方法直接根据采集的VI曲线进行故障判定,无法自动调整测试参数,测试人员需要手动调整测试参数,这样严重影响测试工作效率。以容性器件为例,根据模拟特征分析原理建立故障前后信号差异的目标函数,依据被测器件参数利用改进的高斯-牛顿迭代法对测试信号进行自动灵敏度调整,计算机仿真结果表明,该方法有效,且优化参数对电流信号差异的提升尤为明显。

Regular analog signature analysis and testing methods judge failures directly accordiing to the acquired VI curves,but can not automatically adjust the testing parameters,that is,the testers have to adjust the parameters manually. Therefore,it affects the testing efficiency seriously. In this paper,taking the capacitive device as an example,the objective functions of signal difference before and after the fault were established according to the principle of simulation signature analy-sis,and the sensitivity automatical adjustment for the testing signal was conducted by using the improved Gauss-Newton iterative method based on the parameters of a device under test. The computer simulation results show that the method is effective,and the optimization parameter is obviously benefit to current signal difference′s ascension.