Research On Coordinated Planning Of Integrated Electricity And Natural Gas Systems Considering Uncertainty

The high proportion of new energy sources has transformed the power system from a deterministic system to a strongly uncertain system,and from a single power system to a multienergy coupled integrated energy system.In this context,the traditional deterministic optimization methods and the strategies of independent planning and operation of each energy system are no longer applicable,and the development of the integrated en ergy system optimization methods considering uncertainty becomes an inevitable trend.In this paper,we investigate the coordinated planning method of the integrated electricity and natural gas system considering uncertainty.Multiple uncertainties are taken into account,such as the source and load uncertainty,the spatial correlation of wind and solar uncertainty between different regions,and the development uncertainty that gradually appears between different planning stages.To cope with different uncertainties,a distributed robust planning model,a multi-region distributed cooperative planning model,and a multi-stage stochastic planning model are established and solved by corresponding algorithms,respectively,as follows:Firstly,the distributed robust optimization model is established and solved by using the column and constraint generation method for the uncertainties at both sides of the source and load in the integrated electricity and natural gas system.A whole-system planning model is constructed,which realizes the coordination of new installation and expansion of power grid lines,natural gas pipelines,substations,city gates,photovoltaics,wind turbines,gas turbines and energy storage devices.Based on the 1-norm constraint and the ∞-norm constaint,the fuzzy set of probability distribution of typical source and load scenarios is established,and the distribution robust optimization model is solved by using the column and constraint generation method.Secondly,for the spatial correlation of wind and solar power between different regions,a scenario reduction method considering spatial correlation characteristics and a multi-region distributed coordinated planning model are established and solved by using the analytical target cascading algorithm.In order to consider the spatial correlation of wind and solar power in the scenario cluster process,the particle swarm algorithm is used to optimize the initial k-means cluster center,to obtain the typical scenario with similar spatial correlation characteristics to the original scenario.Taking the typical scenario as input,a multi-region coordinated planning model is established to carry out the planning of power grid lines,natural gas lines,substations and gas gates in each region.Based on the decomposition modeling principle,the multi-region coordinated planning model is decomposed into transmission grid sub-problems,distribution grid sub-problems and gas network sub-problems,and the analytical target cascading algorithm is used to solve each sub-problem in a distributed iterative manner under the premise of satisfying the consistency constraints of shared variables.Finally,for the development uncertainty that gradually appears between different planning stages,a multi-stage stochastic planning model is developed and solved using a distributed approximate dynamic programming algorithm.The multi-stage stochastic planning model takes into account both the long-term uncertainty in the development process and the short-term uncertainty in the operation level,and realizes the cooperative planning of power grid lines,natural gas lines,substations,gas gates,photovoltaics,wind turbines,gas turbines and energy storage devices.In the solution process,the multi-stage stochastic planning model is reformulated as a Markov decision process with sequential decision characteristics,where the investment variables are set as "wait and see" variables,and the decision is made stage by stage as the development uncertainty is revealed during the planning period.The distributed approximate dynamic planning algorithm converts the multi-stage stochastic planning in the form of the Markov decision process into a single-stage deterministic planning that is easy to solve by decoupling the temporal and spatial dimensions to achieve the problem solution.