基于GA特征选择和BP神经网络的模拟电路故障检测

Analog Circuit Fault Detection Based on BP Neural Network and GA Feature Selection

摘要:

针对模拟电路故障检测中存在测试节点数较多的问题,提出遗传算法与BP神经网络相结合的方法；利用遗传算法的全局、并行寻优能力对模拟电路的系统特征进行优化选择,从而减少BP神经网络输入层节点数；用MATLAB软件对仿真实例数据进行编程实验,直接使用BP神经网络,检测率为66.7%,采用遗传算法与BP神经网络结合的方法,检测率可为100%；结果表明,相对于传统的BP神经网络方法,该方法提高了模拟电路故障检测的平均正确率。

Abstract:

For analog circuit fault detection in the presence of the considerable number of test nodes problems, put forward the genetic algorithm method combined with BP neural network. Using a global, a parallel genetic algorithm optimization of analog circuit system characteristics of the optimal selection, thereby reducing BP neural network input layer nodes. An example of the simulation data with the MATLAB software programming experiment, The direct use of BP neural network, the detection rate is 66.7%. By using the method of genetic algorithm and BP neural networks, the detection rate is 100%.Results show that, compared with the traditional BP neural network method, this method improves the average accuracy of analog circuit fault detection.