Compressed Sampling And Reconstruction Method Of Power Quality Disturbance Signal And Its Application Study To Paper Industry

Power quality disturbance is directly related to the power quality and safety of industrial production.It is important theoretical significance and broad application prospects for power quality disturbance monitoring.However,long-term,multi-point,online,and high-frequency sampling will inevitably generate the new problems.Such as large amount of sampling,high storage cost and low transmission efficiency.Compressed Sensing(CS)technology is an information processing tool to solve massive data.Nevertheless,the current CS technology is still in theory and cannot meet the engineering needs.In order to overcome the application problems,this paper carries out the basic theoretical and experimental research on CS technology,and apply the research results to the power quality monitoring of paper industry.The research contents and contributions of this paper can be summarized into the following four aspects.(1)Study on Structure and Performance Analysis of Measurement Matrix Based on the RD-AICTo address the problem of compressed sampling and reliability reconstruction for power quality disturbance continuous signals,a compressed sensing scheme is designed based on the Random Demodulation Analogy-to-Information Converter(RD-AIC).Firstly,a measurement matrix is constructed for studying the irrelevance and sampling phase deviation of the RD-AIC.Then the influence law is explored between the parameter of the RD-AIC and the irrelevance and sampling phase deviation.Thirdly,the RD-AIC parameter setting method is obtained to improve the reconstruction accuracy and anti-noise robustness.These research work and experimental results show that the RD-AIC can solve the above problem and improve the sampling recovery accuracy and anti-noise robustness.(2)Analysis on Sparsity of Power Quality Disturbance Signals and Research on Improved Reconstruction AlgorithmTo address the issue of the real-time performance for power quality disturbance signal reconstruction under the DFT dictionary,a reconstruction scheme is designed based on the Sparsity Adaptive Matching Pursuit algorithm(SAMP).Firstly,the amplitude spectrums are derived to power quality disturbance signals.Secondly,the scientific basis of sparse representation is obtained for power quality disturbance signal in DFT dictionary.Then,an improved SAMP algorithm is proposed based on the spectral energy difference.To address the issue of the real-time performance of the instantaneous pulse disturbance signal reconstruction under the concatenated dictionary,a reconstruction scheme is designed based on the Orthogonal Matching Pursuit algorithm(OMP).An over-complete concatenated dictionary is constructed and its effectiveness and anti-noise robustness is tested.Then,an improved OMP algorithm is proposed based on the partial concatenated dictionary.The above studies will not only reduce the difficulty of dictionary construction,but also achieve the purpose the real-time performance of power quality disturbance signal reconstruction.(3)Study on Reconstruction Algorithm of Power Quality Disturbance Signal Based on TCN theoryTo address the issue that traditional reconstruction algorithms is limited by the sparsity of power quality disturbance signals,a reconstruction scheme is proposed on the basis of the Temporal Convolutional Network theory(TCN).Firstly,a sparseness-free reconstruction model is proposed based on the TCN theory,and a new TCN structure is designed with reconstruction function.Secondly,a power quality disturbance data set and a deep learning framework are constructed,and the feasibility and effectiveness are carried out by experiments.Through the above theoretical and experimental research work,a fast reconstruction method is explored without sparse representation.The new method will not only break the traditional reconstruction algorithm,but also achieves the purpose of rapid batch reconfiguration.(4)Application study of CS technology in power quality disturbance monitoring of paper industryThe paper industry is a typical representative of the energy-intensive process industry.As the application background,an application study for CS technology is carried out.Firstly,the interactive relationship is researched among the main production process,advanced production technology,important power loads and power quality disturbances,and then a conclusion is given that harmonic and voltage sag are the typical power quality disturbances in the paper industry.Secondly,a compressed sampling and reconstruction experimental devices is developed,and an optimization monitoring scheme is proposed for paper industry users.Lastly,the results of harmonic and voltage sag tests show the effectiveness and superiority of CS technology in application.The above researches will lay the foundation for the application and promotion of CS technology in the monitoring of power quality disturbance signals in the paper industry.In conclusion,aiming at the compressed sampling and reconstruction of power quality disturbance signals and its application study to the paper industry,this paper carries out theoretical researches and experimental verifications mainly on the three aspects:the structure and performance analysis of the measurement matrix based on the RD-AIC,the sparse characteristics based on the DFT and concatenated dictionary and the improved reconstruction algorithm,and the reconstruction algorithm based on the TCN.And the part of research results is applied into power quality disturbance monitoring experimental devices and the optimization scheme.The research results indicate that the method proposed in this paper can effectively improve the performance of the CS technology,meet the demands of engineering applications for reliability and real-time performance,and promote the popularization and application of the CS technology in paper industry power quality disturbance monitoring.