Optimal Investment Decision Of Complex Power Grid Based On Power Demand And Investment Capacity

In recent years,with more and more attention to energy resources and environmental protection,sustainable social development is a major challenge in the development of many countries.The electric power industry is one of the main sources of sulfur dioxide,nitrogen oxides and particulate dust pollution.Among them,the sulfur dioxide and nitrogen oxides of the power plant emissions accounted for 40% of the national total emissions.Facing the multiple pressures,including continuous economic development,resource shortages,environmental constraints,the traditional power industry,the traditional electric power industry,which is dominated by thermal power,is in urgent need of structural adjustmen and management optimization.The emergence of new energy sources which are flexible,clean,safe,economical and friendly brings a new era of power system development.Although the development of new energy is helpful to solve the existing problems,for the complex power grid with new energy distributed power supply,under the environment of the new round of power system reform,the accounting and the sale of electricity release competition,which make the income of the regional power grid be greatly affected,profit margins of power grid enterprise decline and investment capacity be greatly constrained.At the same time,the new electricity reform also proposed to strengthen the overall planning of electricity.Therefore,the electric power demand forecasting,investment capacity,power grid planning and investment decision are facing new challenges and problems.It is urgent to study the optimal investment theory based on the power demand and investment ability of complex grid,so as to make the optimal decision for the investment of the complex grid project.The primary research contents and innovations are followed:(1)Set up an intelligent forecasting model of power demand for complex power grid with distributed generation.This model consists of a distributed power generation power forecasting model based on immune-fly optimization support vector machine(IAFOA-SVM)and a combined forecasting model based on BP neural network and polynomial regression.Based on the prediction results of the two models,the distributed power generation is removed from the total power demand level,and the complex grid power demand is obtained.In view of the regional power grid load forecasting,this paper proposed the phase combination model based on BP neural network and polynomial regression,by setting training conditions,using polynomial regression to avoid neural network training into local optimum.For distributed wind power generation,this paper presented the combination forecast model based on immune flies optimization SVM,using immune algorithm and fruit flies algorithm to achieve the complementary advantages,combining and optimizing the support vector machine(SVM)parameters,improving the new energy power generation prediction accuracy.Then,combined with predicted results,this paper further analyzed the degree of reduction of distributed energy to the overall load level.(2)Set up a non dominated sorting genetic algorithm and tabu search(NSGA-TS)based on hierarchical multi-objective optimization model of complex power grid planning with distributed power supply.From the perspective of distributed power grid-connected,combined with complex grid power demand results,the expansion planning model of the new energy grid-connected is established.Using non dominated sorting genetic algorithm to improve the traditional genetic algorithm,to achieve global searching,combined with the strong local search ability of tabu search,the non dominated sorting genetic algorithm and tabu search algorithm are organically combined,a hybrid non dominated sorting genetic tabu search algorithm with strong global search ability and local search ability is established.NSGA-TS hybrid algorithm improved the convergence speed of the algorithm,and the comprehensive cost in the objective function is considered the balance of network loss and line investment,the operation reliability of the power grid and the cost of carbon emissions,then the multi-objective programming model of the coordinated development of distributed power grid and power grid is constructed,at last,the NSGA-TS hybrid algorithm is applied to power grid planning example.(3)Put up the analysis of the factors affecting the investment ability of power grid enterprises based on Knowledge Mining Technology.Firstly,the index system of influencing factors of investment ability is constructed and the influence factors of investment capacity are analyzed by calculating the similarity between the factors.Then grey correlation clustering analysis is used to classify and analyze the influencing factors,and the corresponding classification rules are used to eliminate the redundant attributes and indexes by setting the threshold.Finally,use the path analysis to explore the path and and the influence effect between the key factors and investment ability,obtain the direct effect,indirect effect and comprehensive effect value between the independent variable and other independent variables and the independent variable and dependent variable,to prepare the investment capacity forecast and analysis for power grid enterprises.(4)Set up the generalized predictive model of the investment capacity of the power grid enterprises based on the system dynamics under the multi objective constraints.Considering the investment capacity of power grid enterprises from net profit,financing and depreciation three point of view,and the conditions such as asset-liability ratio and minimum own capital ratio,combining the theory of system dynamics,this paper put forward the investment capacity prediction model based on system dynamics.The model has been successful applied in province electric power company.In order to improve the universality of the model,through decomposing the investment ability step by step,this paper establishes a generalized forecasting model formula,and the five generalized predictive curves which can reflect the essential characteristics of power grid enterprises are obtained.For the first time we can use a single comprehensive index(Investment Ability)to reflect the operating level and investment strength of the power grid enterprises,and provide the image and intuitive investment control curve for investment decisions.On the basis of the generalized predictive model curve,this paper carried out the risk factors strength test for grid investment,comprehensive research on power grid integration risk factor for dynamic decomposition,received an effect degree of significant risk factors and risk.In view of the risk factors,this paper proposed the control plan of the power grid enterprises under different scenarios scheme,to dissolve the risk of uncertainty.(5)Set up a new model of complex multi-attribute evaluation of complex power grid based on the grey target theory of fuzzy cumulative prospect for initial screening of investment projects and an optimization model of complex power grid investment based on stochastic weight optimization cuckoo algorithm(RWCS).Based on the analysis of power demand forecasting and the calculation results of complex power network planning model,this paper constructs a multi attribute evaluation index system of investment projects from the aspects of technology,economy and society.This paper introduces the cumulative prospect theory,and puts the two concepts of risk preference and expected utility into the evaluation.At the same time,for the problem that the attribute value is interval valued intuitionistic fuzzy problem,a new model of stochastic multi-attribute comprehensive evaluation based on fuzzy theory of grey target is proposed.According to the final score,the single item investment priority ranking is obtained.According to the portfolio investment of construction project,using modern portfolio theory,this paper constructs the investment portfolio optimization model based on the random weight cuckoo algorithm under the constraints of power demandthe,the reliability level and investment ability of the power grid enterprises.In order to improve the convergence speed of the algorithm,the method of updating the position of bird's nest with random weights is used to solve the problem of slow convergence speed and conduct the empirical application.The empirical research shows that the model provides an applicable decision method for the optimal investment decision of complex power grid and construct the complex power grid optimization investment decision management process based on power demand and investment ability.