Dynamic Marginal Cost Optimization Of Microgrid Considering Confidence Of Renewable Energy Generation

Renewable energy,such as wind power and photovoltaic,is developing rapidly.The consumption of renewable energy in a single microgrid is greatly affected by meteorological conditions and cannot coordinate with other microgrids to absorb the unbalanced distribution of renewable energy in each microgrid.In order to make full use of renewable energy sources in large areas,multi-microgrids can be used to solve this problem.Microgrid uses energy production,energy storage and energy conversion equipment to achieve the balance between supply and demand of energy in the region.（1）In the process of energy production,especially the renewable energy power generation process of multi-microgrids,the accuracy of meteorological prediction plays an important role.At present,great progress has been made in the research of weather prediction in China,but the prediction problems have not been completely solved so far.In order to improve the consumption capacity of renewable energy in the microgrid cluster area,we consider deploying sensors in the microgrid cluster area,adopt the iterative four partition deployment method,and propose the combined spatiotemporal attention mechanism network and Multilayer Perceptron（C-STAN-MLP）model.According to the measurement data of meteorological sensors in the microgrid cluster area and the prediction data of ruitu system,Predict the meteorological elements of multiple groups of meteorological sensors deployed near renewable energy sources in the region.This hybrid model has two parts: one is spatiotemporal attention mechanism network（STAN）and the other is multi-layer perceptron model.The two prediction results are fused by step-by-step fusion method to improve the prediction accuracy of natural resources such as scenery.（2）According to the previous renewable energy power generation and operation data and the existing concept of power generation prediction accuracy,this paper puts forward the confidence index of renewable energy power generation at a specific time,which is used to describe the minimum possible value of renewable energy output at that time.Furthermore,this paper puts forward the equivalent unit model of renewable energy output.According to the predicted value,actual value,wind and light abandonment data and load data of renewable energy output of microgrid,a robust estimation model of renewable energy power generation confidence and its probability distribution based on different load peak periods is proposed,and a solution algorithm based on robust programming problem is established.Finally,the simulation model is established,and the system’s parameters and operational data are used to conduct the daily dispatching of the renewable energy generation confidence model.The results show that the amount of renewable energy consumption has been significantly improved and achieved good results.（3）In view of the dynamic marginal cost of micro grid operators and sub microgrids,based on the multi-microgrid（MMG）structure and its component mathematical models,a cost model of renewable energy sources,combined heat and power（CHP）units and energy storage systems is established.It is particularly pointed out that the renewable energy source is equivalent to the virtual unit’s participation in the day ahead scheduling,and the dynamic cost optimization model is established.The dynamics of this model is reflected in that it is "dynamically" specified by the sub microgrid operator according to its risk tolerance ability.In the simulation analysis,in the microgrid with CHP unit,under the scenario of thermal power determination,it can not only meet its own energy demand,but also reduce the cost.A dynamic marginal cost model considering the confidence of renewable energy power generation is established.Compared with previous studies,the micro grid dynamic marginal cost considering the confidence of renewable energy power generation is more conducive to reduce the installed capacity and power generation of conventional thermal power units,achieve good economic benefits,and can effectively meet the needs of users.