Research On Time-frequency Analysis Methods For Dynamic Power Quality Disturbances Based On Window Function Optimization

In recent years,renewable energy generation has increased greatly,power electronic equipment and nonlinear and impact loads have also increased,and the influence of power quality disturbance is becoming more complex.Accurate and rapid analysis of power grid signal parameters and detection of dynamic power quality disturbances are important bases for the diagnosis,prevention,treatment and maintenance of power quality related problems.The deep research on dynamic power quality disturbance detection and the construction of new algorithms with high precision,easy implementation and strong anti-noise performance are of great theoretical and engineering value to improve power quality and promote the safe and economic operation of power grid.In view of the demand for dynamic power quality disturbance analysis,the typical time-frequency analysis method of power quality disturbance in recent years is analyzed,and two new time-frequency analysis methods are proposed to improve the base wave amplitude and phase detection accuracy of power quality disturbance signal,improve the integral transformation time-frequency energy aggregation performance,low complexity,and robustness.The main research contents of this dissertation are as follows:(1)The limitations of the current mainstream linear time-frequency analysis methods S transform and generalized S transform in dynamic power quality disturbance detection are analyzed.This dissertation introduces the improvement idea from the original Fourier transform to S transform,which has attracted much attention,and points out the problems existing in the derivative algorithms of Fourier transform.The important role of window function in the linear time-frequency analysis method is analyzed,and the principles and limitations of S transform to obtain the local characteristics of time and frequency by using Gaussian window related to time and frequency are introduced.By taking some typical modified S transform algorithms as examples,this dissertation analyzes the existing problems of the generalized S transform algorithm and points out the key points for the research content.(2)An improved S transform algorithm based on optimized Gaussian window function is constructed to improve the detection accuracy of power quality disturbances,which is represented by the amplitude fluctuation of the fundamental wave of power grid signals,such as voltage sag,swell,interruption,flicker.The S transform and the typical modified algorithms Optimized S Transform(OST)and Modified S Transform MST in recent years are studied.On the base of theoretical analysis and experimental verification,the limitations of the S transform and its modified algorithms are analyzed,and the characteristics of the Gaussian window function used in integral transformation change with the change of detection frequency are explored.The adjustment mechanism of the Gaussian window function for the time-frequency integral transformation is studied,and then the two-parameter control Gaussian window function is designed,which is beneficial to the detection of power quality disturbance.Using the designed optimized Gaussian window function,an improved S transform(IST)algorithm is constructed for power quality disturbance detection.The definition of IST and its inverse,the principle of adaptive adjustment of variable resolution,the form of Fourier transform of IST,and the result form of IST time-frequency analysis of signal are given.The discrete expressions of IST algorithm and its form expressed by Fourier transform are determined.According to the discrete expression of the Fourier transform of IST,the fast Fourier transform and its inverse are used to simplify the IST,and the redundant information such as the non-sensitive frequency points are removed,which greatly reduces the computational load and improves the computational efficiency.The control parameters of IST are determined according to the signal characteristics of power quality disturbances,which forms the IST method which has the prospect of engineering application.The formed IST method has a promising prospect in engineering application.(3)To verify the performance improvement of window function optimization method for time-frequency analysis algorithm,the IST algorithm is used to conduct time-frequency analysis experiment research on dynamic power quality disturbance signal.To verify the performance of the proposed IST algorithm,the dynamic power quality disturbance signal is analyzed using IST.The normal signal,voltage swell,sag,interrupt,flicker,harmonic,interharmonic,oscillation transient,DC-level,phase jump,voltage sag and time-varying harmonic complex,flicker and oscillation transient complex,and frequency variation and time-varying interharmonic complex signals are used to validate the time-frequency analysis performance of the IST,OST,and MST.The deviation between the actual fundamental wave and the detection result by different algorithms is described by three different error measures: mean square error,mean absolute error and root mean square error.Simulation results show that the proposed IST method has better detection accuracy than the existing typical S transform algorithm.(4)To realize the time-frequency integral transform with high energy concentration,improve the frequency detection ability of power quality disturbance such as dynamic harmonics,interharmonics,and oscillation transient while ensuring the high precision fundamental wave detection ability,and reduce the algorithm complexity,a fast K-S transform algorithm based on Kaiser optimization window is constructed,and the K-S transform theories are perfected,thus forming a time-frequency analysis theory suitable for power quality disturbance detection.The window function and its role in time-frequency integral transformation are analyzed,and the Kaiser window function with nearly maximum main sidelobe energy ratio,strong adaptability and flexibility and its characteristics are introduced.The time-frequency dependent Kaiser window function is designed as integral transform kernel function,and the K-S transform algorithm is constructed.The form of Fourier transform of K-S transform is deduced,and the amplitude and phase results of the K-S transform are determined.The discrete form of K-S transform is given,and the time-frequency analysis of signal is realized according to the form of Fourier transform of K-S transform.The timefrequency analysis performance of the K-S transform is tested by using typical signals including cross component,mutation component,nonlinear frequency modulation,and noise interference.The experimental results show that the K-S transform algorithm has excellent time-frequency energy concentration performance and strong anti-noise performance.According to the requirements of detecting the characteristics of the fundamental signal components and high-frequency disturbance components of power quality disturbance,the time-frequency window shape adjustment function related to frequency is constructed,and the optimized Kaiser window function is designed for power quality disturbance detection,and the spectrum characteristics of the window function are analyzed.The continuous and discrete expressions of K-S transform based on the designed optimized Kaiser window are given.On the base of the form of the discrete K-S transform,the fast Fourier transform and its inverse transform are used to realize the fast K-S transform.Furthermore,the non-characteristic frequency points are eliminated to greatly improve the computational efficiency,to realize the fast K-S transform,and determine the control parameter selection method of K-S transform,which lays a theoretical foundation for the practical engineering application of K-S transformation.(5)To verify the excellent frequency detection ability of the fast K-S transform based on optimized Kaiser window function for power quality disturbances,such as dynamic harmonics,interharmonics and oscillatory transient,simulation experiments are carried out.The K-S transform is tested by using voltage sag,interrupt,swell,flicker,harmonic,interharmonic,oscillation transient,noise interference,voltage sag and oscillation transient complex,flicker and time-varying harmonic complex disturbance signals.In addition,the simulation tool Simulink and power electronic system modeling and simulation module set Sim Power Systems are used to construct voltage sag and interruption events caused by power system circuit faults to test the K-S transform.Experimental results show that the fast K-S transform method proposed in this paper not only has high detection accuracy of fundamental amplitude,but also has excellent high frequency disturbance detection capability,which is suitable for dynamic power quality disturbance signal analysis.(6)To verify the effectiveness and feasibility of K-S transform algorithm for detecting power quality disturbances in complex power grid signals,a dynamic power quality disturbance analysis system based on fast K-S transform is constructed.The standard power source Fluke 6100 B is used to generate power quality disturbance signals.The current type voltage transformer is used to adjust the power grid signal,and the signal acquisition module is formed by combining with the analog data acquisition card to realize the signal conditioning,conversion,acquisition and transmission.The system interface is designed by MATLAB App Designer,and the communication between upper computer and data acquisition module is realized by USB 2.0.The standard signal is used to calibrate the system,and the flicker,sag,interrupt,swell,and voltage sag and harmonic complex disturbance are used to test the system.The test results show that the dynamic power quality disturbance analysis system based on fast K-S transform has excellent detection ability for time,frequency and amplitude of complex power grid signals.The maximum effective voltage value and the actual cut-off frequency that can be analyzed by the system are 715.6150 V and 9k Hz,respectively.For the disturbance signal of the voltage sag compounded with the 7-th and13-th harmonics,the relative error of the voltage sag value is 3.32%,and the duration error is 0.0002 s.The relative errors of amplitude of the 7-th and 13-th harmonics are 0.40% and0.33%,respectively.