Research On Key Technologies For On-line Measurement Of Dielectric Loss Angle

With the improvement of social productivity and the progress of science and technology,power system has a direct impact on all aspects of social production and life,and power system security has become increasingly important.As an important part of power system,insulation performance of high voltage electrical equipment is an important factor affecting the safe and stable operation of power system.Insulation accidents of high voltage electrical equipment will not only lead to equipment damage,but also lead to other safety accidents.It is very necessary to determine the insulation properties of high voltage electrical equipment quickly and accurately.Dielectric loss angle is an important index to measure the insulation characteristics of electrical equipment,and is of great significance for monitoring the operation of electrical equipment.At present,because of the poor stability and repeatability,large measurement error and poor anti-interference ability,the measuring method of dielectric loss angle can not completely replace the traditional regular interruption maintenance.Therefore,it is of great engineering and practical value to study the dielectric loss angle measurement method with high accuracy and reliability.Fast Fourier transform(FFT)is widely used in the measurement of dielectric loss angle because of its simple and easy implementation in embedded processor.But the spectrum leakage and fence effect influence the measurement accuracy of dielectric loss angle.The time variability and complexity of the actual signals of the power grid have more influence on the accuracy of the method of measuring the dielectric loss angle which based on FFT algorithm.In order to overcome the shortcomings of the method of harmonic analysis in dielectric loss angle measurement,the methods based on windowed interpolation FFT,the measurement when the fundamental frequency changes of power grid and under the complex distorted power grid signals are studied in this dissertation.Firstly,the spectrum leakage and fence effect of spectrum analysis method based on FFT are analyzed.The method of measuring the dielectric loss angle based on the Blackman self-convolution window is proposed.By constructing the Blackman self-convolution window,the sampling signal is weighted and the spectrum leakage is effectively suppressed.The deviation of the fundamental frequency measurement caused by the fence effect is corrected by the discrete spectrum interpolation,and the fitting algorithm of the least square polynomial fitting is used.The corrected spectrum is used to improve the accuracy of dielectric loss angle measurement.The simulation results show that the method based on the Blackman self-convolution window can effectively improve the measurement accuracy of the dielectric loss angle and have good anti-interference ability.Secondly,a dielectric loss angle measurement method based on time domain series reconstruction is proposed in this dissertation.In order to solve the problem that the actual fundamental wave frequency is difficult to keep in the nominal frequency,which causes the problem of asynchronous sampling.By shifting frequency and filtering to the actual sampling signal,the estimated value of the actual fundamental wave frequency is calculated,and the sampling signal is reconstructed by using the cubic spline interpolation method with the estimation value of the fundamental wave frequency.The reconstructed quasi-synchronous sampling signal is used to measure dielectric loss angle,and high accuracy result is obtained.Through the simulation experiments which including fundamental frequency fluctuation,true value change of dielectric loss angle,change of sampling frequency,change of ADC quantizing bits and white noise interference,the effectiveness of the proposed method of time domain series reconstruction of dielectric loss angle is verified.Next,the influence of complex distorted signal on the accuracy of dielectric loss angle measurement is studied.On the basis of S transform,a time-frequency analysis method based on K-S decomposition is proposed to solve the shortage of the existing time-frequency analysis methods in energy aggregation and time frequency parameter identification.The amplitude-frequency characteristics of Gaussian window and Kaiser window are analyzed,and a Kaiser optimization window with adaptive frequency adjustment is established.The frequency shift and spectrum multiplication are carried out by using the adaptive Kaiser optimization window after the fast Fourier transform of the sampled signal,and then the complex time frequency matrix is obtained by the fast Fourier inverse transform.The timefrequency analysis of complex distorted signals is realized.Combined with the measurement of the dielectric loss angle under the complex distortion signal,the influence of the different distortion signals on the measurement error of the dielectric loss angle is analyzed,which provides an improved basis for the measurement method of the dielectric loss angle under the complex distortion signal.Finally,the dielectric loss angle measuring instrument is designed on the basis of the study of the dielectric loss angle measurement method,and the hardware design scheme of the dielectric loss angle measurement with ADC+DSP+ARM is proposed.In order to optimize the design of IIR digital filter,a state evaluation adaptive differential evolution algorithm is proposed,which adaptively adjusts the control parameters according to the evolution state and distribution state of the population in the evolutionary process,and is used to optimize the parameter of the low pass filter of the dielectric loss angle measuring instrument.The software modular design of the dielectric loss angle measuring instrument is introduced,and the accuracy and stability of the designed dielectric loss angle measuring instrument are verified by the reality test.