Studies On Combined Comprehensive Evaluation Methods Of Power Quality

With the continuous increase of non-linear load of power systems and the growth of power consumers,the power quality problems have been more and more serious.On the other side,the rapid development of the national economy leads to higher standard of power quality for users.The increasing contradiction between the increasing demand for power quality and power quality problems is urgent to be solved by the academics and power supply companies.Detecting and evaluating power quality scientifically are not only beneficial for making more effective control of power quality,but also favor for building a fair deal-market of power quality which fixes prices according to the quality of products.Based on the analysis of index character of power quality evaluation and detection methods,the thesis focuses on the detection method of harmonic and inter harmonic in order to enhance the detection accuracy of these indexes.At the same time,a new index weight determining method and a model of combined evaluation are proposed.Then the power quality can be classified.The main research content is as follows:Firstly,the thesis constructs a detection model based on elastic net and sparse representation to inspect the harmonic and inter harmonic component.The method proposed in the thesis solve the problems of unavailability of inter harmonics detection based on FFT and the sensitivity to noise based on traditional sparse representation detection method.The detection precision of frequency of each harmonic and inter harmonic component is 100%.The method could detect the amplitude in error range of?5%and the phase in error range of?5^?.Comparing with traditional sparse representation detection method based on l₁-norm,the method has better noise immunity.Moreover,the method can reduce detection errors greatly caused by noise,especially amplitude errors of harmonic and inter harmonic component.Secondly,the thesis researches on methods of index weight determining in power quality comprehensive evaluation.A new method called cross entropy method is proposed aiming at the shortcoming of the existing objective weight methods.Comparing with three traditional objective weight methods,it not only does much better at making the difference of indicator weights,but it is more stable when using different sources of data or existing a hopping point in the data.The thesis proposes to apply a combination weight method to confirm the weight in combined evaluation of power quality in order to overcome the one-sidedness of subjective and objective weight method and realize the complementary advantages of two weight methods.Finally,the thesis constructs a comprehensive evaluation model of power quality based on the idea of combining evaluation.The evaluation results of the method are obtained by using power mean model to combine the results of three single comprehensive evaluation methods.The consistency checking method including ante and backing tests in this evaluation process is Spearman correlation coefficient method.In this way,the thesis realizes a more comprehensive and reasonable evaluation of power quality and guarantees the credibility of the combined evaluation model.Case analysis shows the feasibility of the evaluation method proposed in the thesis,and the grading and sequencing of power quality of each monitoring site has been realized.