College of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
A two-level model of station-network cooperative optimization of integrated energy system for park is established. The upper-level model is an improved p-median model, and the decision variables are the location and capacity of energy stations and network layout. Considering the demand response characteristics of electric and heat loads, the lower model coordinates the interruptible load with the energy station supported by the power-heat-gas multi-energy flow. The greedy-variable neighborhood-cobweb (GVNC) algorithm is used to solve the model, which can improve the computational efficiency and obtain the global optimal solution. Second-order cone relaxation (SOCR) is used to transform the non-linear programming problem into a second-order cone programming (SOCP) problem which is easy to solve. Convex-concave procedure （CCP） is used to solve the problem of accuracy recovery after relaxation. The example shows that considering the influence of source-load interaction, the operation cost of the system can be effectively reduced, and the planned capacity of the energy station can be more economical and reasonable.