【摘要】 针对瓦斯传感器数据传输过程受到强电磁信号干扰而产生误报警的问题，在实验模拟“大数”干扰的条件下，提出了基于遗传算法（GA）优化BP神经网络的煤矿瓦斯监控系统大数干扰信号辨识方法。首先根据大数干扰信号幅值变换较快的特点，确定BP神经网络的输入与输出，然后利用贝叶斯正则化的方法提高网络的泛化能力，结合遗传算法的全局搜索能力，优化网络的权值和阈值，最后建立了GA优化BP神经网络的大数干扰信号辨识模型。实验结果表明，遗传算法能够有效降低BP神经网络的训练误差，GA优化BP神经网络辨识模型和未被优化的BP神经网络相比，优化后的网络测试信号相对误差由10.003降低到6.096，且辨识后不会影响瓦斯突出信号的正常输出，能够解决煤矿井下因大数干扰信号造成传感器误报警的问题。更多还原

【Abstract】 Aiming at the problem of false alarm caused by strong electromagnetic signal interference in the data transmission process of gas sensor,under the condition of experimental simulation of " large number" interference,a large number interference signal identification method of coal mine gas monitoring system based on genetic algorithm(GA)optimized BP neural network is proposed. Firstly,according to the characteristics of fast amplitude transformation of large number interference signals,the input and output of BP neural network are determined,and then the Bayesian regularization method is used to improve the generalization ability of the network,combined with the global search ability of genetic algorithms,the weight and threshold of the network are optimized,and finally the large number interference signal identification model of GA optimized BP neural network is established. The experimental results show that the genetic algorithm can effectively reduce the training error of BP neural network,and compared with the unoptimized BP neural network,the relative error of the optimized network test signal is reduced from 10. 003 to 6. 096,and the normal output of the gas outburst signal will not be affected after identification,which can solve the problem of false alarm of the sensor caused by large number of interference signals in coal mines.更多还原