A Research On Demand Response Modeling And Algorithm With Microgrid

The imbalance between power supply and load demand often leads to power interruption,power instability and damage to power equipment.As the user’s load demand grows rapidly,ensuring the stability of the power grid becomes a great problem.In order to solve this problem,we must first ensure the balance between power load demand and production.For the power grid,accurately portraying the consumer’s electricity behavior and predicting the consumer’s power demand are the premise to ensure the balance of power supply and demand.It helps to formulate power generation plans more reasonably and effectively.At the same time,the grid can guide users to temporarily change their inherent electricity usage habits by adjusting electricity policies and electricity prices during peak or low peak periods.In order to achieve optimal allocation of resources,to ensure the stability of the power grid system,and further improve its economy and safety.However,with the rapid development of the economy,the demand for electricity is also growing,and the power structure is becoming more and more complex.Centralized power generation and long-distance transmission power are gradually being replaced by micro-grids.More and more microgrids participate in power trading as retailers,but the addition of microgrids also increases the uncertainty of power demand,increases the difficulty of planning and scheduling for power grids,and enables demand response research with microgrids has become a difficult problem.This paper has carried out modeling and algorithm research from consumer’s electricity behavior analysis,load forecasting and demand response.By extracting the user’s power consumption mode,the correlation of the user load on the time series information is found,and the time series characteristics of the user load data are input to the long-term and short-term memory neural network load forecasting model to capture the user’s power consumption law and achieve accurate prediction.By forecasting the demand,the power companies formulate a three-level Stankelberg game strategy based on the the Prediction-of-Use tariff mechanism to guide users to use electricity reasonably.(1)In the aspect of consumer electricity behavior analysis,firstly the K-Means method is used to cluster the daily load curves of non-resident consumers,classify and mine the consumer’s electricity behavior patterns,and then the Spearman correlation coefficient is used to calculate the time correlation of the non-resident consumers’s load data on the time series information,it is found that there are adjacent-time related,day-related and week-related for a specific consumer.(2)In terms of demand forecasting,a non-resident consumer load forecasting model based on Long Short-Term Memory neural network is proposed o predict the load using three different time series features:adjacent-time correlation,day-correlation and week-correlation.The experimental results shows that the proposed method can effectively utilize multiple sequences features and successfully capture the dependence between multiple sequences features and load,it has better prediction accuracy than existing load forecasting methods.(3)In terms of demand response,this paper proposes a three-level Stankelberg game model based on the prediction-of-use tariff mechanism,which solves the multi-objective optimization problem in the three-level game through the backward induction method,and derives optimal price and demand strategy to solve maximum benefits for power grids,retailers and users.At the same time,after joining the POU tariff mechanism,the grid has increased the regulation of user load demand,reduced power generation uncertainty and ensured the smooth operation of the power grid.