Research On Power Forecasting And Dynamic Environmental Economic Dispatch In Power System With Photovoltaic Power Generation

At present,human beings are facing many problems such as energy crisis and environmental pollution,and adopting renewable energy instead of traditional fossil energy is the necessary way and development trend to solve these problems.With the gradual maturity of renewable energy utilization technology,China's renewable energy has experienced rapid development,and the proportion of photovoltaic power generation in the power grid has been increasing year by year.Compared with conventional energy,photovoltaic power generation has the advantages of clean,convenient,safe and reliable,short construction cycle,easy to large-scale industrialization,etc.,and has now become an important part of the sustainable development of energy and ecological environment.However,natural factors such as meteorology have a large degree of influence on PV power generation,resulting in randomness and fluctuation of PV power generation,which makes PV power generation susceptible to uncertainty factors,and these uncertainty factors create great challenges for power system operation and grid dispatching.Load forecasting and PV power forecasting are the prerequisite and foundation of power system optimization and dispatching,and the forecasting accuracy and forecasting effect of load and PV power can be improved through reasonable and effective forecasting technology.On the basis of the load forecasting data and PV power forecasting data,it is of great significance and importance to make full use of solar energy,improve the utilization rate of PV power generation,reduce the impact and influence of power system with grid-connected PV power generation,and build an ecological and environment-friendly society.In this context,this paper conducts short-term load forecasting and PV power forecasting for "light-fire" co-generation power systems with grid-connected PV power generation,and conducts in-depth research on environmental economic dispatch of power systems based on accurate load forecasting data and PV power forecasting data.The main work and innovation results are as follows.(1)A short-term load forecasting model based on quantum fireworks algorithm optimized support vector machine(QFAW-SVM)is proposed.Firstly,the problem that the parameters of support vector machine(SVM)are not easily optimized in the process of short-term load forecasting in power system,which leads to less accurate load forecasting results,is analyzed.To address this problem,this paper establishes the quantum fireworks algorithm to optimize the SVM model.The QFAW-SVM forecasting model is established to forecast the load of the "light-fire" co-generation system during the optimized scheduling period.The actual load example analysis verifies the rationality and effectiveness of the proposed model.The proposed method can avoid the problem that the parameters of the SVM model are not optimized enough,and effectively improve the short-term load forecasting effect of the SVM model,providing a new method for the accurate short-term load forecasting,and providing a load data basis for the subsequent environmental and economic dispatch of the "light-fire" co-generation system.(2)To address the problem that the volatility and randomness of PV power generation affect the accurate forecasting of PV power,and the problem that the forecasting model based on traditional machine learning algorithm has a single method and the forecasting accuracy is not satisfactory,an composite forecasting model(ICEEMDAN-IF-MPSO-SVM)based on improved adaptive noise-complete ensemble empirical modal decomposition algorithm(ICEEMDAN)combined with support vector machine(SVM)is preposed.First,the ICEEMDAN algorithm is used to decompose the original PV power into a series of intrinsic mode functions(IMFs),which are adaptively divided into corresponding groups.Secondly,an improved particle swarm algorithm(MPSO)is used to optimize the SVM model parameters,and the MPSO-SVM forecasting model is established to forecast the fluctuating components of PV power represented by the corresponding IMF groups,and finally,the final forecasting results are obtained by superimposing and reconstructing the forecasting results of each group.The generality and superiority of the forecasting model proposed in this paper are verified by an actual example analysis.The proposed model reasonably decomposes the fluctuating PV power signal,effectively reduces the influence of fluctuating components in the PV power signal on the power forecasting accuracy,better suppresses the uncertainty of PV power generation,and improves the forecasting effect of PV power.(3)To address the problem that the forecasting model based on traditional machine learning algorithm ignores the learning process and cannot well explore the correlation between PV power and its influencing factors in the process of PV power forecasting,a composite forecasting model(OVMD-IPSO-LSTM)based on the optimized variational modal decomposition algorithm(OVMD)combined with long and short-term memory network(LSTM)is proposed.Firstly,the number of modes and penalty factors of the conventional variational modal decomposition algorithm(VMD)are optimized by using a parameter optimization method based on maximum of feature frequency and mutual information,and the improved algorithm(OVMD)is used to decompose the PV power into a series of intrinsic mode functions(IMFs),secondly,a particle swarm algorithm(IPSO)with improved inertia weights is established to optimize the LSTM model parameters.The IPSO-SVM forecasting model is established to forecast the fluctuating components of PV power represented by the each corresponding IMF,and then the forecasting results of each fluctuating components are superimposed and reconstructed to obtain the final forecasting results.The effectiveness of the proposed model is verified by an actual example analysis.The proposed method can effectively solve the noise problem of PV power data signals,smooth the PV power signals,reduce the interference of fluctuating components on PV power forecasting,suppress the influence of PV power uncertainty on power forecasting,and realize the accurate forecasting of PV power.(4)A combined forecasting model of PV power based on the entropy method is proposed.Three single forecasting models,namely,genetic algorithm optimized wavelet neural network(GA-WNN),ICEEMDAN-IF-MPSO-SVM and OVMD-IPSO-LSTM,are used to forecast the power of actual PV power plant separately,and then the entropy is used to determine the weighting coefficients of each single forecasting model,and the combined forecasting model based on the entropy is established.The forecasting results of the actual example analysise show that the combined forecasting model proposed in this paper,further improves the forecasting accuracy and forecasting effect of PV power above the ICEEMDAN-IF-MPSO-SVM forecasting model and OVMD-IPSO-LSTM forecasting model in Chapters 3 and 4,and provides a new idea for the combined forecasting of PV power,It also provides a basis for the subsequent environmental economic dispatch of "light-fire" co-generation system.(5)A dynamic environmental economic optimization dispatching model of power system with PV power generation is proposed.Combining the load and PV power forecasting technology and forecasting data,the model is established for the economic operation and environmental protection of the "light-fire" co-generation system with grid-connected PV power generation.The dynamic environmental and economic scheduling model of the "light-fire" co-generation system is established,and the dual optimization objectives of the lowest cost of power generation and the lowest emission of polluting gases of the co-generation system are adopted,and the fast non-dominated ranking genetic algorithm with elite strategy mixed with improved particle swarm algorithm(NSGA-II-PSO)is used to solve the optimal scheduling model,and the multi-objective Pareto optimal solution set of the optimal scheduling of the co-generation system is obtained,and the optimal compromise solution is obtained by the principle of maximum satisfaction.Finally,in the dynamic scheduling of the "light-fire" co-generation system,the integrated optimization of environmental and economic indexes is achieved.