Research On Operation Status Identification Technology Of Distributed New Energy

With the increasing demand for energy all over the world and the depletion of traditional non-renewable energy,people pay more and more attention to green and renewable new energy.New energy can be divided into centralized new energy and distributed new energy according to the way of access to the power grid.Centralized new energy means a large number of new energy resources centralized access to the power grid,in addition to new energy equipment,there are almost no other equipment within the access point,so it is easy to identify its operation status.Distributed new energy means a mixture of new energy resources and general load(residential load,industrial load,etc.),so it is difficult to identify the operation status of new energy resources alone.With the increasing proportion of distributed new energy in power grid,in order to ensure the safe,stable and efficient operation of power grid,it is necessary to identify its operation status effectively.There are two main methods to identify distributed new energy.The first method is centralized measurement,this method is to install a sensor on the the high voltage bus connected with new energy,because only one sensor is needed,which can put a higher cost on the sensor to ensure the accuracy of the data.However,the collected data contains both new energy resources and general load information,so we need to find a suitable method to extract the new energy resources.The second method is distributed measurement,which is to install a sensor on each new energy equipment,by calculating the sum of sensor readings,new energy resources can be identified.However,due to the cost,the cost of each sensor is low,so the collected data may have data quality problems,such as abnormal data and missing data.In order to solve the problem that it is difficult to extract new energy resources data in centralized measurement method,this paper proposes a distributed new energy identification method based on data collected from high voltage bus.This method uses the method of data mining to find a feature with obvious distinction between new energy resources and general load data,by analyzing the difference between the new energy resources and the general load data,including the content of the feature in the high voltage bus side data,the new energy resources can be identified.Taking photovoltaic output data as an example,the intrinsic mode function(IMF)component IMF6,which can be used as identification feature,is successfully extracted by using ensemble empirical mode decomposition(EEMD)algorithm.The simulation results based on the measured photovoltaic data in Jiangsu Province show that IMF6 can be used as the identification feature,and when the ratio of distributed photovoltaic data and general load is greater than 1:15,this method can effectively identify the distributed photovoltaic data.Aiming at the problem of missing data in distributed measurement method,this paper proposes a data patching method based on bidirectional prediction.This method makes full use of the data before and after the missing data to construct two one-way prediction models in different directions.Through the weighted average of the prediction results of the two models,the missing data can be repaired.At the same time,in order to improve the accuracy of oneway prediction,we propose a prediction model based on EEMD algorithm,Savitzky-Golay(SG)filter,long short-term memory neural network(LSTM)and autoregressive integrated moving average model(ARIMA).Taking wind speed data as an example,the simulation experiment based on the measured wind speed data in the United States shows that this method can effectively repair the missing data within four points.