Research On Power Demand Side Management Optimization Under Air Pollution Prevention And Control Background

With the development of industry and economy in China,the consumption of resources is on the increase and the air pollution problems are increasingly apparent.Especially from the winter of 2012,the haze-fog broke out in Yangtze River delta area,Beij ing-Tianj in-Tangshan region,northeast region and other area.The contradiction between air pollution and economic sustainable development is increasingly prominent.The main reason of air pollution are electric power industry electricity-coal combustion,end-user scattered coal combustion,oil-fueled vehicles,among which the electric power industry electricity-coal combustion is one of the most important source of air pollution.According to statistics,electric power industry SO2 emission accounts 34.6%of the national total emission,and NxOy accounts 37.68%,carbon emission accounts more than 50%.Therefore,at current stage,an advanced power demand side management theory should be adopted to assist the implementation and deepening of air pollution prevention and control policy.Power demand side management means by multi-agent collaboration,taking effective economic,technological,administrative and guiding measures to optimize the way of using power and improve terminal power utilization efficiency,which can realize load shifting,energy utilization efficiency improvement,energy saving and environmental protection,etc.At the same time,under the background of rapid development of smart grid,serious air pollution and strong user environmental awareness,demand side management needs constantly improvement and innovation.Therefore,based on current air pollution problems facing our country,this thesis optimizes power demand side management method from five sides:power demand forecasting,time-of-use electricity price,peak dispatching,policy incentives and management optimization proposals.The main research results and innovation are listed below:(1)Carry out research on power demand forecasting under air pollution prevention and control background,and set up power forecasting model based on guassian disturbance firefly algorithm and support vector machine,and load forecasting model based on density K-medoids,bayesian neutral network and generalized autoregressive conditional heteroscedasticity model.For power forecasting,the thesis quantifies the influence of ironmaking production,steelmaking production and cement production on electric power.On this basis,guassian disturbance firefly algorithm is proposed to optimize support vector machine parameters and then adapt guassian disturbance firefly algorithm and support vector machine forecasting power demand level in recent years,which can greatly improve power forecasting accuracy.For load forecasting,air quality index,steel production,iron production and concret production are regarded as air pollution prevention and control quantilification factors and used in daily peak load forecasting.Sample is separated into three clusters by density K-medoids algorithm.Then,using bayesian neutral network forecasts the load level of each clusters and adapting generalized autoregressive conditional heteroscedasticity model to correct forecasting error,which also greatly improve the forecasting accuracy of daily peak load and 24 point load.(2)Carry out research on green time-of-use electricity price mechanism optimization based on power demand forecasting under air pollution prevention and control background,and set up green time-of-use electricity price optimization model based on dynamic newman watts-two stage force-particle swarm optimization algorithm.Based on green time-of-use electricity price mechanism and cost-benefit principle,the thesis analyzes power user tradable green certificate purchasing psychology,and use utility theory to quantify tradable green certificate purchasing behaviors.Combined with users' reflection status under time-of-use electricity price,the thesis establishes green time-of-use electricity price user reflection function,and adapting fuzzy fast K-medoids algorithm divide time quantum.On the basis of user behavior recognition and time quantum division,with the aim at valley-to-peak optimization,this thesis builds green time-of-use electricity price optimization model,and then uses dynamic newman watts-two stage force-particle swarm optimization algorithm to get the best price scheme.The optimization results reveal that:green time-of-use electricity price mechanism can better mobilize user to change power utilization custom,which playing better role in load shifting.In addition,green time-of-use electricity price can enhance load optimal configuration and power generation efficiency,as well as largely reducing coal consumption and pollutant emission.(3)Carry out research on multi-energy complement combined peak dispatching system based on power forecasting and price under air pollution prevention and control background,and set up combined peak dispatching optimization model based on non-dominated sorting genetic algorithm ?.In addition,adapt matter-element extension method and variable weight theory to evaluate the total benefits of combined peak dispatching system.This thesis selects the most suitable wind speed distribution function to fit the random wind sequence probability distribution,which can accurately depict wind power.On the basis,this thesis analyzes the peak dispatching mechanism of regenerative electric heater and electric vehicle.Then,non-dominated sorting genetic algorithm ? is adapted to get the best accessing scheme.After that,this thesis establishes combined peak dispatching system synthetic evaluation indicator system from two internal aspects(technology enhancement and economic development)and two external aspects(social benefit increasing and environmental benefit growth).Later,matter-element extension method and variable weight theory are used to evaluate the system total benefit.The research result reveals that:the combined peak dispatching system proposed in this thesis can presents better performance in balancing load curve,optimizing power,energy conservation and emission reduction,etc.(4)Carry out research on demand side management incentive mechanism optimization under air pollution prevention and control background,and propose demand side management incentive process analysis method based on dynamic evolutionary game model.This thesis combs the operation framework of demand side management incentive mechanism,and then indicates the roles and jobs of government,grid corporation and power user.After that,the game behaviors of government-grid corporation and government-user are analyzed.On the basis of incentive mechanism analysis,dynamic evolutionary game models are built to simulate game behaviors of different objects in the process of demand side management incentive,and finally obtained the best game strategy profile and system dynamic evolutionary path.Based on the game results,the main factors and incentive mechanism influencing strategy selection are obtained.Finally,the policy incentive measures for Grid Corporation and user are set according to target constraint.In the end,the policy incentive measures proposed in this thesis are embedded into gear dynamic analysis model and establish an interconnected and closed-loop excitation system,which realizes the demand side management incentive mechanism optimization.(5)Design demand side management optimization proposals under air pollution prevention and control policy.Based on former researches on power demand forecasting optimization,green time-of-use electricity price optimization,united peaking dispatching system optimization and policy incentive optimization,this thesis analyzes the blank or shortcomings of demand side management in air pollution prevention and control,and then come up with constructive proposals.