Optimal Power Flow By Newton Approach

TLDR: In this paper, a direct simultaneous solution for all of the unknowns in the Lagrangian function on each iteration is proposed, where each iteration minimizes a quadratic approximation of the Lagrangeian.

Abstract: The classical optimal power flow problem with a nonseparable objective function can be solved by an explicit Newton approach. Efficient, robust solutions can be obtained for problems of any practical size or kind. Solution effort is approximately proportional to network size, and is relatively independent of the number of controls or binding inequalities. The key idea is a direct simultaneous solution for all of the unknowns in the Lagrangian function on each iteration. Each iteration minimizes a quadratic approximation of the Lagrangian. For any given set of binding constraints the process converges to the Kuhn-Tucker conditions in a few iterations. The challenge in algorithm development is to efficiently identify the binding inequalities.