The Benders Decomposition And Coordination Method For Reference Network

Transmission system planning guarantees the reliable supply of power. With the rapid development of renewable energy, interconnections of power grids, and China marketization reform, the transmission system planning problem faces more and more challenges.The reference network model yield the optimal transmission capacity of transmission lines by simultaneously minimizing the system operation and transmission cost. The reference network can be used to assess the existing network, and it can also quantify the congestion cost and the optimal investment strategies. Compared with the existing power grid structure, it can find out the weak links of the system. For an existing line, the reasonable margin can be determined by comparing its existing capacity with that in the reference network model. For the future, combined with load forecasting and power planning, gradual transmission system planning can be implemented:according to the evaluation of transmission system by reference network, the optimized decision model with optimal transmission capacity will be set up, which is dynamic, process considered. This model can make sure balance between power supply and demand. As a result, the transmission system will meet the long-term marginal cost, that is, to maximize the social benefits with the transmission system planning gradual.For a large-scale power system, the reference network model is very complex, because tremendous unknown variables and line outage events must be considered. This paper uses the Benders decomposition algorithm to solve the model. The entire problem is decomposed into a series of simpler sub-problems which include Benders master problem and Benders sub-problems. The master problem and sub-problems are linked by Benders cuts.In order to improve the efficiency of algorithm, several strategies are implemented. Firstly, the power allocation factor is applied to calculate the line flow based on DC flow. Secondly, the dominant events are selected and all other non-dominants are ignored during the Benders decomposition algorithm. Lastly, the constraints of Benders sub-problems are further reduced. The case study demonstrates that the computation efficiency is greatly improved.