Energy Storage Capacity To Stabilize Photovoltaic Output Fluctuations

The energy storage system is often used to smooth the power fluctuation caused by the randomness and intermittency of renewable energy due to its flexible control and rapid response,but the cost of the energy storage system is high.In the case of meeting the output requirements as far as possible,the allocation of appropriate energy storage capacity is the key to promote photovoltaic power generation.At present,many scholars at home and abroad have proposed the configuration of energy storage capacity through time-frequency analysis method.Common time-frequency analysis method includes empirical mode decomposition and wavelet transform.The wavelet transform needs to manually determine the wavelet basis function,and the basis function is non-adaptive that the wavelet basis may be the best globally,but it maybe not in some local areas.Compared with wavelet transform,Empirical Mode Decomposition(EMD)does not need to analyze and study the data to be decomposed in advance.Instead,it hierarchies the data according to some inherent pattern based on the characteristics of the data itself.However,in the existing literature,these methods of configuring energy storage capacity according to the EMD decomposition method do not consider the power fluctuations in different time periods,and corresponding types of reserve capacity can be used to smooth;Secondly,although these methods have been improved by considering the modal aliasing problem existing in EMD,the improved methods also have other problems,such as pseudo-modal problems;finally,these methods do not consider whether the components decomposed by the time-frequency analysis method are all effective components.This paper has done further research on the above problems,and the specific work is as follows:(1)An energy storage capacity configuration based on EMD combined with FFT(EMD-FFT)is proposed,and the control power of photovoltaic power station is decomposed by EMD,and the high and low frequency components are divided by the cumulative mean of standardized modulus(MSAM).Refactoring the divided component,and time-frequency transformation with the FFT.In the frequency domain to fine division of control power by the spectrum analysis,through configuring different type of response time corresponding fluctuation reserve capacity to configuring the required energy storage capacity.Eventually it makes more economical that on the basis of satisfying the photovoltaic output.(2)An energy storage configuration capacity based on improved complete ensemble empirical mode decomposition with adaptive noise(ICEEMDAN)combined with FFT(ICEEMDAN-FFT)is proposed,although EMD can be decomposed according to the photovoltaic control power data itself,EMD decomposition will have serious mode aliasing problem.In view of the mode aliasing problem of EMD,EEMD can improve the mode aliasing problem by adding noise.However,if the number of white noise processing times is not large enough,the residual white noise will cause large errors in the reconstructed signal.CEEMDAN can improve the mode aliasing problem of EMD and reduce the amount of computation,but there are many pseudo-modal components.Therefore,this paper introduces ICEEMDAN,decomposes the control power of photovoltaic power station through ICEEMDAN,and divides high and low frequency components through MSAM.Refactoring the divided component,and time-frequency transformation with the FFT.In the frequency domain to fine division of control power by the spectrum analysis,through configuring different type of response time corresponding fluctuation reserve capacity to configuring the required energy storage capacity.This method can improve the modal aliasing problem of EMD and reduce the reconstruction error.Configuring a more reasonable energy storage capacity which make it more economical that on the basis of satisfying the photovoltaic output.(3)An energy storage capacity configuration based on ICEEMDAN combined with improved variational modal decomposition(IVMD)combined with FFT(ICEEMDAN-IVMD-FFT)is proposed,although ICEEMDAN can improve the modal aliasing problem and reduce the amount of computation and pseudo-modal components,its decomposition nature is recursive decomposition like EMD decomposition,modal aliasing is inevitable.In addition,there will still be pseudo-modal components.If there are false components in the pseudo-modal,there will be a large error in the subsequent signal reconstruction which will lead to a large energy storage capacity in the final configuration.Therefore,this chapter proposes to first decompose the photovoltaic control power by ICEEMDAN,and then calculate the sample entropy and pearson correlation coefficient of each IMF component,and further decompose the IMF components with high complexity and sample entropy greater than 1 by IVMD,and retain the effective components whose the pearson correlation coefficient is more than 0.1.Secondly,the number of IVMD decomposition layers K is determined by the double standard of energy difference and sample entropy,and the pearson correlation coefficient is calculated for the K components decomposed by IVMD,and the effective components whose pearson correlation coefficient is more than 0.1 are retained.Dividing high and low frequency components by the MSAM and reconstructed,time-frequency conversion is carried out through FFT,and the final energy storage capacity is configured through spectrum analysis.