| Since the end of the 1980 s in China,power industry began to implement reforms to break monopolies,gradually introduced competition mechanisms,experienced trials of regional electricity markets based on a single electricity purchase model,and gradually launched trials and points for large-scale direct electricity purchases.The series of reforms,such as the electricity price policy,have accumulated experience for China's constant push for the establishment of a demand responsive electricity market.The electricity market will gradually shift to real-time price.This process will be accompanied by a sudden drop in costs,and the overall efficiency of the power system will be greatly enhanced.The article will study the DSR based on supply-demand matching and source-load benefits maximization;Carry out multi-objective optimization screening on the basis of analyzing the power user's response potential;Build load forecasting model when taking into account the impact of DSR on power load;Set up double-sided benefits optimization model with the goal of maximizing system benefits;Analyze the relationship between participants of the demand side project,and propose a two-phase optimization scheduling strategy combining the complementary relationship between price response and incentive response.The main contents and conclusions are as follows:(1)Clarify the power market trading mechanism and supply-demand matching issues,and lay theoretical foundation for the study of articles.There are bilateral trading markets and unilateral trading markets from the perspective of supply and demand matching.Firstly,compare and analyze the mentioned two market forms from aspects of transaction mode and equilibrium mechanism.Secondly,analyze the relationship between incentive-based demand response and price-based demand response.Finally,based on the above analysis,it will analyze supply-demand equilibrium in electricity market that is compatible with DSR,mainly including demand elasticity analysis,supply-demand analysis,and supply-demand equilibrium points determination.(2)Apply the improved detection algorithm which is based on minimum distance method in information entropy principle to obtain power data set cluster,and the extracted feature data is corrected throughKohonen network to complete detection of feature data;The user's comprehensive electricity consumption information is clustered and analyzed through the improved k-means algorithm;Screen potential users by optimization goals,such as total demand-response price,the highest overall sensitivity price and both of them.The results show that,the obtained data parameters have better performance by introducing information entropy principle,and solve uncertainty problem caused by improper selection of the decision shrinkage factor and clustering parameters in traditional CURE algorithm,also the calculation time is lower than CURE algorithm with the increase of data points.The Kohoenn network is used to modify feature data,and the original characteristics of input error data are effectivelyobtained,so that detection data completely shows the characteristics of source data and the state of regular matrix.During the detection process,when the detection target is divided into10 clusters,the number of first cluster is 5,and the number of regular feature points is 50,the detection result is ideal.The cluster analysis and the screening of potential users are completed based on the true and accurate feature data,and screening results are in line with actual situation.(3)A load forecasting method "dividing load response resources and superimposing traditional load forecasting" is proposed based on the efficient GNN-W-GA forecasting model.The impact of energy efficiency resources and load resources on load patterns,and the response potential of power users in the region are fully considered in forecasting method.The original load part and response load part are superimposed by respective impact mechanisms and priorities on the load patterns which is quantified.The forecasting method improve prediction accuracy.Moreover,the method can be effectively applied to real-time price response.In simulation analysis,it is found that the correlation between forecast date and similar date,the definition of solution space,and the number of similar days have a greater impact on the prediction accuracy.It is concluded that high correlation and non-linear similar day number equal to dimension of solution vector can improve prediction accuracy.It provides new ideas for load forecasting in power grid operations as demand side gradually participate in electricity market.(4)Taking the cost and effectiveness of power companies as an example,the method for calculating various indicators of DSR costs and benefits is described,the basic process of cost-benefit is explained,and the feasibility of implementing DSR projects in a certain region is demonstrated;Built a load response model based on DSR measures which are "Time-of-use price" and "incentive emergency" to explore the ways of changing electricity demand and demand-side optimal electricity consumption patterns in accordance with the principle of user benefit maximization.An example is used to analyze the impact of changes in TOU price,changes in elasticity matrix,and changes in incentive subsidies to load curve.The analysis concludes that,the introduction of an effective DSR during the peak period make consumers adjust their electricity consumption based on price signals.Since DSR can reduce power consumption during the peak period,stability benefits are generated in electricity market.Through the short-term benefit analysis of DSR,it can be seen that the short-term benefits generated by the introduction of DSR are not related to whether or not the reduction in the price increase;Through the analysis of medium and long-term benefits,it shows that the frequency and magnitude of fluctuations in price can be reduced by the effective DSR,realize the linkage between retail market and wholesale market,play a positive role in the reliable operation of system in power supply cycle.(5)In order to maximize power system benifit,supply-demand benefits optimization model is put forward by fullly demonstrating supply benefit,demand benefit,and systemoperational constraints.According to the characteristics and practical significance of model,the improved PSO is used to solve the model.The results show that,the improved PSO can achieve better performance when dealing with system fluctuation unbalance constraints.Not only does the power imbalance measure decrease,but also the convergence speed increases significantly.The example analysis shows that,system benifit is significantly improved throgh the proposed model,also satisfying electricity demand and system operation target;when demand benifit is reduced,the system benefit is reduced,but the supply revenue does not change significantly.The reason is that,in order to ensure the benefit maximization of system,user power consumption is stable through load response.When demand benefit is increased,the system benefit and supply revenue are also increased,indicating that the focus of model is to adjust demand side response load,in order to improve consumption efficiency,so that the DSR project achieve the desired goal;The changes of large users' electricity benefit have a greater impact on system bebefit,and small users on the contrary.If set subsidy limit within a reasonable range will not only reduce subsidies,but also avoid the use of subsidy amounts for arbitrage by maliciously reducing electricity consumption.(6)Through the two-stage coordinated optimization scheduling,the system benefit is maximized,and solve the supply-demand optimization problem,use GWO to perform optimization calculations.The first stage is DRP priority response scheduling,this stage is mainly used to reduce electricity expenses.The second stage is DRA scheduling,the purpose is to improve system performance,such as reducing system loss,reducing voltage standard deviation,increasing load factor,and smoothing loading curve.At the same time,DRP formulates different incentive compensation strategies based on the characteristics of users,enabling users to actively participate in DSR project and contribute to system benefit.The example analysis shows that:electricity expenditure is significantly reduced compared to no response when performing the first stage of scheduling,effectively achieving the optimization goal,also playing an active role in economic benefit increased and peak load filled;In the second phase of scheduling,DRA allocates different proportion of DRP response load according to different incentive compensation strategies,and further reduces system loss and voltage deviation by improving load factor and using incentive subsidy;When the DRP assigns the highest response ratio according to the prioritization principle to the registered users who do not want to change load plan in scheduling phase 1,the system benefit is the highest;when weights are set 0.5 and 0.5,the system operation performance is optimal;when weights are set 0.3 and 0.7,the power cost is the lowest;Under different load response levels,load curve of Case 3 gradually becomes smoother as the response level increases,and the load curve of Case 2(DRP assigns the highest response ratio to an incentive strategy)reaches a better shape when the response level is 20% to 30%,but as the response level increases,the difference between peaks and valleys is gradually widening.Similarly,system loss,voltage standard deviation,and electricity expense follow thesame pattern.The reason is that,DRP scheduling based on real-time price may generate strong or weak peak rebound during off-peak hours,and peak rebound increases the system pressure instead,so system loss and voltage deviation are increased. |
PDF Full Text Request