基于声纹识别技术的变压器异常状态自动诊断方法

Automatic diagnosis method of transformer abnormal state based on voiceprint recognition technology

为使监测主机在变压器设备发生异常运行时能够发出预警,实现对变压器异常状态的精准诊断,设计基于声纹识别技术的变压器异常状态自动诊断方法。按照声纹识别技术应用原理,提取变压器异常运转情况下的时域特征、频域特征与梅尔倒谱特征,根据变压器指标权重的赋值结果,计算频率复杂度参量的取值范围,联合相关变量指标,求解频谱能量分布指数表达式,完成基于声纹识别技术的变压器异常状态自动诊断方法的设计。对比实验结果表明,在上述诊断方法作用下,变压器出现异常运行状态时,电压有效值始终不会超过预设警戒值,二者差值也始终小于5.0V,符合及时预警并精准诊断的实际应用需求。

In order to enable the monitoring host to give early warning when abnormal operation of transformer equipment occurs and realize accurate diagnosis of transformer abnormal state,an automatic diagnosis method of transformer abnormal state based on voice print recognition technology is designed.According to the application principle of voice print recognition technology,time domain characteristics,frequency domain characteristics and Meir cepstrum characteristics under abnormal operation of transformer are extracted.According to the assignment result of transformer index weight,the value range of frequency complexity parameters is calculated,and the expression of spectral energy distribution index is solved by combining related variable indexes.The automatic diagnosis method of transformer abnormal state based on voice print recognition technology is designed.The comparative experimental results show that under the above diagnostic method,when the transformer has abnormal operating state,the voltage RMS value will never exceed the preset alarm value,and the difference between the two is always less than 5.0V,which meets the practical application requirements of timely warning and accurate diagnosis.