Research On Optimal Scheduling Of Building Load Demand Response Based On The Estimation Of Distributed Algorithm

With the increase of power consumption of the whole society and the domestic power consumption of urban and rural residents,energy consumption is the key problem for China and the whole country at present.In the future,energy consumption will continue to increase,and power consumption will also increase year by year.It will result in greater changes in power consumption at peak and low power consumption times on the power grid side,and the infrastructure projects of the power grid will become more onerous.In order to slow down this trend,some strategies were adopted to reduce the fluctuation of power grid at peak and low times,and to improve the consumption rate of renewable energy,such as the building load optimal scheduling and changing the use scheme of renewable energy.At present,the demand response(DR)strategy is mainly adopted.For example,Time-of-use tariffs(TOU)and real-time pricing(RTP)are adopted to change consumers’ electricity consumption habits and reduce electricity consumption.TOU and RTP are used to motivate residents to increase consumption during low prices and reduce consumption during peak prices.These measures are used to further improve the utilization of primary energy consumption.Taking the building load as the research object,the TOU demand response strategy is used to solve the problem of rapid optimal scheduling and preference selection of building residents.The works of the paper are organized as follows:(1)The relevant theories of building load optimal scheduling and the research status of estimation of distributed algorithm(EDA)at home and abroad is described in detail,and then the deterministic boundary process algorithm is used to optimize the scheduling of single resident load.According to the growth rate of domestic power consumption of the whole society and urban and rural residents in China,the significance of building load optimal scheduling is analyzed,and the relevant theories of building load optimal scheduling is described,including DR,household energy management system,TOU and EDA.Simulation of single household load optimal scheduling: firstly,the categories of household loads are divided.Secondly,the matrix mathematical model of various loads and the objective function of double-layer optimal scheduling are established,and the deterministic boundary process algorithm is used to solve it.(2)An improved multi-objective estimation of distributed algorithm(MOEDA)based on the influence factor of TOU is proposed.Firstly,the advantages and disadvantages of building load matrix and vector building load coding are analyzed.In order to simplify the load coding space and reduce the computational complexity of the algorithm,a composite building load coding is proposed,which is applied to build the building load mathematical model and time distribution model.Under the condition of considering the demand response of power grid side and residents,the building residents’ optimal scheduling objectives: cost and power consumption deviation are constructed.In order to ensure the rapidity of building optimal scheduling and reasonable distribution of temporary optimal scheduling scheme,and transfer the power consumption of high electricity price and peak time to low electricity price and valley time,the characteristics of building load and TOU are analyzed.The impact factor of TOU is constructed to improve the probability model of MOEDA and to accelerate the convergence speed of MOEDA in the process of building load optimization.In the case of building resident load optimal scheduling,the effect of MOEDA optimal scheduling before and after the improvement is compared.It is found that the improved MOEDA has better effect and can allocate an optimal scheduling scheme.(3)A MOEDA based on adaptive compound probability model is proposed.In view of the demand of building residents for diversified scheduling schemes and the lack of diversity of building load scheduling schemes optimized by MOEDA,the population size change as well as its advantages and disadvantages of probability model constructed by global Pareto frontier and current generation Pareto frontier are analyzed in the iterative process.Two adaptive parameters varying with the number of iterations are proposed,and a fast-growing and stable probability model of Pareto front is established.A fast-growing and stable probability model of Pareto frontier is established.MOEDA with adaptive probability model is proposed to ensure the diversity of building load optimal scheduling schemes.Building residents can choose the best load optimal scheduling scheme.Building residents set the load scheduling time range according to their preferences,allocate the target weight by entropy weight method and objective programming method,and select the best optimal scheduling scheme.