摘要
电机的振声信号包含了丰富的状态信息,利用电机的声音信号进行故障检测,并提出了综合小波包能量相对熵的概念。对采集的声音信号进行小波包分解,利用重构系数计算故障信号与正常信号小波能量包相对熵,并得到综合小波包能量相对熵,确定电机是否有故障;并通过比较各频带的综合小波包能量相对熵大小判断故障所处的频带位置,从而确定电机为何种故障。电机振声信号诊断实例验证了该方法的有效性。
Vibrating sounds of motors, which contain abundant state information. In this paper we use vibrating sounds for fault detecting, and put forward a conception of relative entropy of integrated wavelet packet energy. It makes wavelet packet decomposition, uses decomposition coefficients to calculate the wavelet packet energy relative entropy of normal signals and faulty signals, and then obtains the relative entropy of integrated wavelet packet energy, which determines whether the motor is faulty. Comparing the results in each frequency to determine the fault location and what kind of fault. The example of audio fault detection has verified the effectiveness of the method.
出处
《煤矿机械》
北大核心
2014年第3期231-233,共3页
Coal Mine Machinery
基金
粤港关键领域重点突破项目(2010Z5104)
关键词
电机噪声
故障检测
综合小波包能量相对熵
motor noise
fault detection
relative entropy of integrated wavelet packet energy
