基于小波包分析的医疗器械电路故障检测方法研究

Research on the Detection Method of Medical Device Circuit Fault Based on Wavelet Packet Analysis

由于对医疗器械电路故障检测时没有分析故障信号的时频特性,使得故障检测结果不准确,检测时间长、漏检率高,为此提出基于小波包分析的医疗器械电路故障检测方法。采用小波包分析法分析故障信号的时频特性,并对故障信号逐层分解,获取信号频带能量,将各个频段的能量设定为故障特征向量的元素,提取故障信号特征,将提取的特征作为训练样本对BP神经网络进行训练,采用测试样本对BP网络进行测试,获取不同电路故障类型,最终实现医疗器械电路故障检测。实验结果表明,所提方法可以获取高精度的医疗器械电路故障检测结果,检测时间短,漏检率低。

Because the time-frequency characteristics of fault signal are not analyzed in the fault detection of medical device circuit,the fault detection results are inaccurate,the detection time is long and the missed detection rate is high.A medical device circuit fault detection method based on wavelet packet analysis is proposed.The wavelet packet analysis method is used to analyze the time-frequency characteristics of the fault signal,decompose the fault signal layer by layer to obtain the signal frequency band energy,set the energy of each frequency band as the element of the fault feature vector,extract the fault signal characteristics,take the extracted characteristics as training samples to train the BP neural network,and test the BP network with test samples to obtain different circuit fault types.Finally,the circuit fault detection of medical devices is realized.The experimental results show that the proposed method can obtain high-precision medical device circuit fault detection results,with short detection time and low missed detection rate.