基于BIM振动信号的变压器机械故障诊断方法研究

Research on Diagnosis Method of Transformer Mechanical Fault Based on Vibration Signal

为了能够提高电力系统运行过程中的可靠性,基于振动信号研究了电力系统变压器中出现的机械故障诊断方法。分析了BIM定位不同测点下的振动信号,利用时域频谱图对比分析了相同相、不同位置和不同相、相同位置的振动信号基频幅值。提出了基于小波包变换的振动信号故障特征提取方法,将信号特征转化为数值特征,为变压器机械故障的智能诊断提供了工具方法。最后,在传统PSO算法中引入了惯性权值,将改进后的IPSO算法与BPNN进行结合,得到了优化后的IPSO-BPNN算法模型。通过分析得知,IPSO-BPNN算法模型的收敛速度更快,具有更少的时间成本、更低的计算资源消耗、更高的准确性。

In order to be able to improve the reliability of the power system during operation,the diagnosis method of mechanical faults in the power system transformer was studied based on the vibration signal.The vibration signals under different measuring points were analyzed,and the fundamental frequency amplitudes of the vibration signals of the same phase,different positions and different phases,the same position were compared and analyzed using the time-domain spectrogram.A vibration signal fault feature extraction method based on wavelet packet transform was proposed,and the signal features was converted into numerical features,which provided a tool for the intelligent diagnosis of transformer mechanical faults.Finally,the inertia weight was introduced into the traditional PSO algorithm,and the improved IPSO algorithm was combined with BPNN to obtain the optimized IPSO-BPNN algorithm model.Through analysis,it was known that the convergence speed of the IPSO-BPNN algorithm model was faster,with less time cost,lower computing resource consumption,and higher accuracy.