Research On The Calculation Method Of Photovoltaic DC Power Based On Modulation Broadband Mode Decomposition

Photovoltaic power generation technology has developed rapidly in my country due to its cleanliness and convenience compared with other traditional power generation technologies(such as hydropower,thermal power,etc.).In my country,the calculation methods of photovoltaic alternating current(AC)signal and further power quality analysis methods are relatively mature,but there are few researches on photovoltaic direct current.The main reason is that the existing time-frequency analysis methods will produce errors caused by the Gibbs effect when decomposing square wave signals,and the traditional electric energy calculation methods are not suitable for direct current(DC)signal electric energy calculation.In order to solve the problem,a Broadband Mode Decomposition(BMD)method is proposed.The main idea of the algorithm is to search in the constructed over-complete dictionary library containing DC signal features.However,when applied to a wideband signal with a relatively small bandwidth,the wideband signal may be regarded as several narrowband components,resulting in errors in the calculation results.To solve this problem,this paper proposes a Modulated Broadband Mode Decomposition(MBMD)algorithm.By multiplying a high-frequency single-frequency component signal,the relative bandwidth of the effective wideband signal is much less than 1,and the wideband signal is treated as an effective wideband signal to obtain a more accurate signal decomposition result.In this paper,the compound Simpson Integral algorithm is used to calculate the decomposed signal results,which improves the accuracy of the calculation and reduces the calculation error.On the basis of accurate power calculation,further power quality analysis is carried out.In this paper,by using the composite multi-scale fuzzy entropy algorithm(Composite Multi-scale Fuzzy Entropy,CMFE)to calculate the signal eigenvalues,and combined with the error back propagation(Error Back Propagation Neural network,BPNN)model in several common neural network algorithm models,the disturbance identification and classification of the signals in the photovoltaic DC system are carried out.The main contents of the research are as follows:(1)The design of the photovoltaic signal simulation model and the construction of the experimental platform for photovoltaic energy metering device.According to the composition of general photovoltaic system,an experimental platform for small photovoltaic power metering device including DC load,AC load and signal acquisition module was built in the laboratory.According to the analysis of multiple experimental signals,a model with photovoltaic signal characteristics is constructed,simulation analysis is performed,and experimental signals are analyzed and processed.(2)Signal decomposition method.In this paper,the MBMD algorithm is proposed.Several representative other signal decomposition algorithms are selected,corresponding mathematical models are constructed,and the comparison with other signal decomposition algorithms is achieved through simulation and experimental signals,it is verified that the algorithm can effectively extract the square wave(broadband)features in the DC signal.(3)Accurate calculation of photovoltaic DC energy.The product-sum algorithm used in the traditional point electric energy calculation and the complex Simpson integration algorithm used in this paper are briefly introduced,and the shortcomings of the traditional algorithm are analyzed.Comparing the complex Simpson integration algorithm with the traditional algorithm through the simulation signal and the collected experimental signal,it is verified that the algorithm can more accurately calculate the photovoltaic DC energy.(4)Photovoltaic power quality intelligent identification.The BPNN algorithm and CMFE algorithm are introduced,combined with the previous signal decomposition algorithm to denoise the signal,and the CMFE algorithm is combined with the BPNN network algorithm to identify and classify various disturbance signals after denoising.Demonstrated by the collected experimental signals,the method realizes the accurate classification and identification of the disturbance signal in the photovoltaic DC signal.