Short-term Photovoltaic Power Generation Prediction Based On EMD-RVM

With the increasing shortage of energy and environmental pollution,photovoltaic power generation has developed rapidly as a clean renewable energy source.However,due to the interference of external uncertainties,it has the disadvantages of volatility and randomness.Photovoltaic grid connection is bound to pose a threat to the safe and stable operation of the power system.Therefore,improving the prediction accuracy of photovoltaic power generation is of great significance to the safe and stable operation of the system,the maintenance of power quality and the effective use of solar energy.In order to improve the accuracy of PV output power prediction,this paper has done the following work on the basis of many related prediction literatures at home and abroad.Firstly,through the characteristics of photovoltaic output power,it is known that it has the characteristics of randomness,periodicity,discontinuity and non-stationarity.It also knows the degree of interference of various environmental factors on photovoltaic output power,Among them,the weather conditions have the biggest interference to the photovoltaic output power.Therefore,the paper divides the weather conditions into four categories:sunny,cloudy,cloudy and rainy and snowy,and constructs corresponding prediction models.Then,based on the support vector machine and the relevance vector machine,two short-term prediction models of single algorithm are constructed respectively.The real power generation data of a photovoltaic power station is used to predict the PV output power under four weather conditions,and the prediction of the two methods is adopted.The result analysis shows that the model prediction based on the relevance vector machine algorithm is better,but there is still a large error between the predicted value and the true value.Considering the non-stationary and nonlinear characteristics of the PV output power numerical sequence and the limitations of the single algorithm prediction model,this paper proposes a combined prediction method based on empirical mode decomposition(EMD)and relevance vector machine(RVM)to predict the PV output power.Firstly,the historical data is divided into four categories according to the weather conditions.The Euclidean distance method is used to screen out the data similar to the day to be predicted,and then the EMD is used to decompose the PV ou tput power numerical sequence to obtain a certain number of relatively stable eigenvectors with different frequencies.The intrinsic modulus function(IMF)component is used to construct the RVM prediction model for all IMF components.Finally,using the equal weight summation method,the predicted values of the components are added to obtain the PV output power prediction value.In this paper,we use the real data of a photovoltaic power station in China as a sample,build a prediction model on MATLAB software and carry out an example simulation to achieve a prediction of the PV output power every 15 minutes between 6:00-19:00.It is known that compared with the above two single prediction algorithms,the error and error fluctuation based on the EMD-RVM combined prediction method are smaller and the prediction accuracy is higher,and the PV output power can be effectively predicted under various weather conditions.