Research Of The Optimal Scheduling Method For Active Distribution Network With High Penetration Distributed Generations Integration

The distributed generation system such as photovoltaic and wind power has the advantages of clean and low-carbon,as well as flexible interface.With the advancement of China's energy structure reform,distributed generation systems have been widely used in power grids.Integration of high penetration DGs in active distribution network(ADN)can reduce the large-scale and long-distance transmission of electric energy,and supply power to users nearby.This will make up for the shortages of centralized generation.However,DGs have inherent characteristics such as randomness,volatility and intermittentness,as well as scattered installation locations.These characteristics bring great challenges to the traditional distribution network operation mode.Therefore,it is necessary to deeply analyze the coordinated interaction mechanism between DGs and other scheduling resources,and explore reasonable optimization scheduling methods.These are of great significance to improve the consumption of renewable energy and realize safe and economic operation of ADN.Thus,this paper conducts an in-depth study on the optimal scheduling method for ADN with high penetration distributed generations integration.The main contents are as follows:To solve the uncertainty of renewable energy output in ADN,an adjustable robust optimal scheduling was proposed based on improved uncertain boundary.Firstly,adjustable robust optimization method was introduced and a robust optimization economic scheduling model was established for ADN which contains compressed air energy storage system.Through linear duality theory and Lagrange transformation,the robust optimization model with uncertain variables was turned into mixed integer optimization problems,which contains only certain variables and can be solved by conventional solutions.Meanwhile,a decision making method for the price of robustness was proposed based on an improved uncertain boundary.Then quantitative analysis was also done for the maximum confidence of constraints with uncertain variables.The proposed improved boundary improves the conservativeness of the existing robust boundary and has better solution characteristics.Finally,The results illustrate the availability and reasonability of the proposed strategy.To realize coordination interaction between DGs and demand side,an optimal scheduling strategy considering uncertainties of both source and load was proposed.Firstly,according to the consumer psychology model,a demand side response characteristic curve was established,which is based on uncertainties of both response magnitude and response boundary of demand side.Then normal cloud model was introduced to give a unified description of the demand side response characteristics.This model can simultaneously reflect randomness and ambiguity of the response deviation.On the basis,an optimal scheduling model considering uncertainties of both source and load was proposed.As the model is difficult to solve directly,the forward cloud generator was adopted to generate random scenes.Then several typical scenes was obtained by the scene reduction technique,based on which the optimal scheduling model was solved.Finally,The results illustrate the availability and reasonability of the proposed strategy.For integration of multi-type DGs into ADN,a coordinated interactive scheduling strategy under the framework of AC/DC hybrid ADN is proposed.Firstly,different types of DGs and other scheduling resources are integrated to construct a typical system structure of AC/DC hybrid ADN.By analyzing the structural characteristics of the system,a hierarchical scheduling architecture of AC/DC hybrid ADN was proposed.Based on this,a two-stage hierarchical scheduling model was established,which was solved by a hierarchical algorithm based on discrete wind-driven optimization.The model can minimize the operating costs of AC/DC sub-networks while reducing system losses.Finally,The results illustrate the availability and reasonability of the proposed strategy.Integration of high penetration DGs has a great impact on the electricity market trading model and optimization method in ADN.To solve this problem,a cooperative evolutionary game strategy for electricity trading stakeholders in ADN under consortium blockchain framework was proposed.Firstly,by analyzing behavior characteristics of electricity trading stakeholders(ETSs)in ADN,a electricity market trading system was established under consortium blockchain framework.This system pre-authorizes each ETS as a semi-open local node.Through the dynamic selection of alliance node,distributed storage of transaction data can be realized and no additional t hird-party regulatory agency is required.Then detailed certification methods for electricity trading in AND was proposed.On this basis,a cooperative evolutionary game model of ETSs was established,and a multi-objective evolutionary algorithm based on decomposition was utilized to solve this model.The solution result can serve as a basis for generating smart contracts and realize decision-making optimization for electricity trading in ADN.Finally,the case studies illustrated the availability and reaso nability of the proposed strategy.In view of the unbalanced boundary power and insufficient utilization of grid resources resulted from traditional separation scheduling schemes,a coordination optimal scheduling strategy was proposed to optimize the transmission and distribution(T&D)system from an overall perspective.Firstly,taking the minimum power supply cost as the objective function,a basic scheduling model was established and the corresponding constraints were discussed.In order to reduce complexity of solving the aforesaid model,an improved concurrent subspace optimization(CSSO)algorithm was introduced,and then a universal model of coordination optimal scheduling for the T&D system was established.This universal model regards the T&D grids as concurrent subspaces for disciplinary analysis,and completes the optimal design unified in the system layer.In addition,construction method of response surface approximation model based on radical basis function(RBF)neural networks was proposed,in order to imitate the discipline analysis process.The results illustrate the availability and necessity of the proposed strategy.