Efficient MILP formulations for AC optimal power flow to reduce computational effort

AC optimal power flow (ACOPF) is a nonlinear and non-convex problem that might include various constraints. The methods in the literature for solving this problem may suffer from premature convergence and high computational effort. In this study, ACOPF is formulated as a mixed-integer linear programming (MILP) that allows problem modeling with a closer look at the limitations of actual operating conditions along with other common constraints. This article contributes to literature by presenting efficient MILP methods of the ACOPF problem in terms of various actual constraints, such as valve loading cost, prohibited operating zones, and the multi-fuel units. The proposed methods aim at reducing the computational effort with a sufficient improvement in finding the optimal solution. Effectiveness and accuracy of the proposed methods are evaluated on the standard IEEE 30-bus, 118-bus, and the polish 2746-bus test systems. Numerical results confirm the validity and quality of the proposed MILP methods with convexified various actual constraints (MILP_CVAC) for the ACOPF problem. A significant reduction is found in computing time to solve the ACOPF problem with the proposed MILP_CVAC methods due to the use of strong MILP solvers and the convexity of the model, without affecting the accuracy. In some cases, the results show several hundred times shorter computation time compared to nonclassical methods.