Research On Harmonic Detection Algorithm In Power System

With the widespread use of non-linear power electronic devices in the power grid,the problem of harmonic pollution in power systems is becoming increasingly serious,which affects all aspects of production and life.Being able to accurately detect the key parameters of each harmonic and current in the power grid is a prerequisite for the effective management of power system harmonics.This thesis takes the most widely used harmonic detection algorithm based on FFT as the main body of research,and conducts research from two aspects of improving the accuracy of the algorithm and compressing the sample size to reduce the storage space.Harmonic detection algorithm based on FFT has spectrum leakage and fence effect due to asynchronous sampling.In order to solve the above problems,a window function is introduced to perform truncation processing on the original signal,and the detection result is corrected by an interpolation algorithm.The performance of the side lobe of a window function directly determines the detection accuracy of the algorithm.The 6-term combined cosine window is the best choice,because it has the smallest peak side lobe level and the largest side lobe attenuation rate.Therefore,a double-spectrum-line interpolation modified harmonic detection algorithm with6-term combined cosine window is proposed.The experimental results show that the algorithm has higher harmonic detection accuracy than the double-spectrum-line interpolation algorithm with other windows,but the detection results of even harmonic components are significantly lower than those of odd harmonic components.Aiming at the problem that the double-spectrum-line interpolation correction algorithm with 6-term combined cosine window has low detection accuracy of even harmonic components,the concept of convolution is introduced to construct a p-order6-term cosine self-convolution window,which further improves the side lobe performance of the window function,and reduces the effect of long-range spectral line leakage on the detection results of weak harmonic signal.A four-spectrum-line interpolation correction algorithm with second-order 6-term cosine self-convolution window is proposed.The experimental results show that the detection accuracy of the algorithm for even weak harmonic signals is significantly improved,which is equivalent to the detection accuracy of odd harmonics,and the algorithm has better robustness.Compressed sensing technology breaks through the Nyquist sampling framework and combines the compression and sampling processes into one,which greatly reduces the amount of sampling and the waste of computer storage space.The theoretical derivation proves that the compressed sensing theory can be applied to the field of harmonic detection in power systems.Therefore,a four-spectrum-line interpolation correction algorithm with 2-order 6-term cosine self-convolution window based on compressed sensing is proposed.The interpolation correction algorithm is introduced in the signal reconstruction process,and the accuracy detection of each harmonic component is achieved through a smaller sampling amount.The experimental results show that the detection accuracy of the combined algorithm is reduced to a certain extent when the sampling volume is 1/4 of the traditional algorithm,but it is still much higher than the national standard for the detection accuracy of power quality testing instrument.This thesis includes 42 figures,19 tables and 100 references.