System and method for optimal power flow analysis

1. A method for determining a power flow of a power grid, comprising:

optimizing, using a processor, an objective function representing an operation of the power grid using a spatial branch and bound (BB) framework for determining iteratively upper and lower bounds of the objective function, wherein the lower bounds are determined using a semi-definite programming (SDP) relaxation of an optimal power flow (OPF) problem.

2. The method of claim 1 , wherein the optimizing comprises:

partitioning iteratively a feasible region of the OPF problem into a nested tree of regions corresponding to a BB tree, wherein the nested tree of regions includes a first region and a second region nested in the first region; and

determining the upper and the lower bounds of the objective function in at least some regions including the first and the second regions, wherein a solution of the OPF problem corresponding to the lower bound of the first region is an input to the SDP relaxation for determining the lower bound of the second region.

3. The method of claim 2 , further comprising:

updating a lowest upper bound of the BB tree with an upper bound of the second region, if the upper bound is less than the lowest upper bound of the BB tree;

updating a lowest lower bound of the BB tree with a lower bound of the region, if the lower bound is greater than the lowest lower bound of the BB tree and lower than the lower bounds of other regions of the nested tree; and

determining the power flow based on the lowest upper bound of the BB tree if a difference between the lowest upper bound and the lowest lower bound of the BB tree is less than a threshold.

4. The method of claim 3 , further comprising:

updating the lowest lower bound of the BB tree with the lowest lower bound of other regions, if the lower bound of the region is greater than the lowest lower bound of the BB tree and greater than the lowest lower bound of the other regions.

5. The method of claim 3 , wherein the upper bound solution for each node of the BB tree is checked for a satisfaction of a sufficient condition for global optimality before the lower bound problem is solved.

6. The method of claim 3 , wherein the lower bound solution for each node of the BB tree is checked for a satisfaction of a sufficient condition for global optimality and if globally optimal, is used to construct an upper bound solution.

7. The method of claim 2 , wherein the splitting is based on structure of elements of the power grid.

8. The method of claim 2 , wherein the SDP relaxation uses an alternating direction method of multipliers (ADMM) method.

9. The method of claim 8 , further comprising:

decomposing semi-definite constraints in the SDP into semi-definite constraints on smaller blocks comprising maximal clique subgraphs of the graph based on an electrical network.

10. The method of claim 1 , wherein the power grid includes at least one storage system, the objective function represents the operation of the power grid over time, and wherein the OPF is a multi-period optimal power flow (MOPF) problem.

11. The method of claim 10 , further comprising:

decoupling time-coupling constraints by dualization with an augmented Lagrangian formulation;

solving SDP problems corresponding to individual time-steps; and

applying an alternating direction method of multipliers (ADMM) method to converge the time-decoupled constraints.

12. The method of claim 11 , wherein the solving the SDP problem for each time-step comprises:

performing a clique decomposition of a graph associated with the power grid; and

applying the ADMM to the augmented Lagrangian formulation of the dual problem.

13. A method for solving an optimal power flow (OPF) problem optimizing an objective function representing an operation of a power grid, comprising:

partitioning iteratively a feasible region of the OPF problem into a nested tree of regions corresponding to a branch and bound (BB) tree, wherein the nested tree of regions includes at least a first region and a second region nested in the first region;

determining an upper bound of the OPF problem in the second region;

determining a lower bound of the OPF problem in the second region using a semi-definite programming (SDP) relaxation of the OPF problem, wherein a solution of the OPF problem corresponding to a lower bound of the first region is an input to the SDP relaxation;

updating a lowest upper bound of the BB tree with the upper bound of the second region, if the upper bound of the second region is less than the lowest upper bound of the BB tree;

updating a lowest lower bound of the BB tree with the lower bound of the second region, if the lower bound of the second region is greater than the lowest lower bound of the BB tree and the lower bound of the second region is lower than lowest lower bound of other regions of the nested tree;

updating the lowest lower bound of the BB tree with the lowest lower bound of other regions, if the lower bound of the second region is greater than the lowest lower bound of the BB tree and the lower bound of the second region is greater than the lowest lower bound of the other regions; and

determining the optimal power flow based on the lowest upper bound of the second region if a difference between the lowest upper bound and the lowest lower bound of the second region is less than a threshold, wherein steps of the method are performed by a processor.

14. A system for solving an optimal power flow (OPF) problem optimizing an objective function representing an operation of a power grid, comprising a processor for optimizing an objective function representing an operation of the power grid using a spatial branch and bound (BB) framework for determining iteratively upper and lower bounds of the objective function, wherein the lower bounds are determined using a semi-definite programming (SDP) relaxation of an optimal power flow (OPF) problem, wherein a solution of the OPF problem corresponding to a lower bound of a first region is an input to the SDP relaxation for a second region.