Dielectric Loss Angle Detecting Method Of Insulating Materials

Aluminum electrolysis multifunctional unit insulation state monitoring is a important link related to the equipment and person safety.Dielectric loss angle is the key performance parameter of the unit insulating materials.However,spectrum leakage and Hurdle Effect of the harmonic analysis affects the accuracy seriously.For the strong noise interference of multiple-function units,this paper adopts the wavelet delamination threshold carry on signal preprocessing,after the analysis and extraction of interference features.this paper determines the selection methods of wavelet function,decomposition layer number,threshold and threshold function.The simulation results show performance of the proposed algorithm has much improved compared with the classical thresholds,such as fixed threshold and unbiased risk estimation threshold.At the same time,comparison with Butterworth filter,Chebyshev filter and elliptical filter is carried out to verify the effectiveness of the algorithm.For the spectrum leakage and Hurdle Effect,this paper proposes an improved triangular self-convolution window all-phase FFT correction algorithm,and designs a second-order triangular convolution window to replace the traditional rectangular window for intercepting the data.The triangular convolution window has the characteristic of low peak value of side lobes and high attenuation rate,which can effectively suppress spectrum leakage.The phase-invariant of all-phase FFT can be used to calculate the signal phase Angle accurately.According to the simulation analysis,in the case of signal frequency fluctuations,harmonics and weaker noise interference,this paper can calculation phase angle signal accurately.The calculation amount of the triangular convolution window all-phase FFT phase correction algorithm only adds N accumulation operations on the basis of FFT and increases the accuracy by 10%compared with traditional add window interpolation,phase difference correction algorithm and energy gravity method.Finally,the paper designs a dielectric loss Angle detection system based on TMS320F28335 and STM32F103 dual processors,in whichDSP is responsible for algorithm implementation and microcontroller for the human-computer interaction and system control.