Strategic Planning for Power System Restoration

This paper considers the power system restoration planning problem (PSRPP) for disaster recovery, a fundamental problem faced by all populated areas. PSRPPs are complex stochastic optimization problems that combine resource allocation, warehouse location, and vehicle routing considerations. Furthermore, electrical power systems are complex systems whose behavior can only be determined by physics simulations. Moreover, these problems must be solved under tight runtime constraints to be practical in real-world disaster situations. This work is threefold; It formalizes the specification of PSRPPs, introduces a simple optimization-simulation hybridization necessary for solving PSRPPs, and presents a complete restoration algorithm that utilizes the strengths of mixed integer programming, constraint programming, and large neighborhood search. 1 Background & Motivation Every year seasonal hurricanes threaten coastal areas. The severity of hurricane damage varies from year to year, but significant power outages are always caused by seasonal hurricanes. Power outages have significant impacts on both quality of life (e.g. crippled medical services) and economic welfare. Therefore, considerable human and monetary resources are always spent to prepare for and recover from power threatening disasters. At this time, policy makers work together with power system engineers to make the critical decisions relating to how money and resources are allocated for preparation and recovery of the power system. Unfortunately, due to the complex nature of electrical power networks, these preparation and recovery plans are limited by the expertise and intuition of the power engineer. Furthermore, the National Hurricane Center (NHC) of the National Weather Service in the United States (among others) is highly skilled at generating ensembles of possible hurricane tracks but current preparation methods often ignore this information. This paper aims to solve this disaster recovery problem more rigorously by combining optimization techniques and disaster-specific information given by NHC predictions. The problem is not only hard from a combinatorial optimization standpoint, but it requires modeling of a complex physical system (i.e. the electrical power network) which is a challenging sub-problem. The electrical power industry has developed several tools for modeling the power system’s behavior (e.g. T2000, PSLF, Powerworld, PSS), each with its own strengths and weaknesses. Furthermore, the electrical power industry recognizes there is not a single model for understanding the behavior of an electrical power network. For that reason, this work seeks to build solution procedures that are independent of any specific electrical power simulation tool. The paper considers the following abstract disaster recovery problem: How to store supplies throughout a populated area to minimize the amount of time each customer is without electricity after a disaster has occurred. It makes the following technical contributions: 1. It formalizes the Power System Restoration Planning Problem (PSRPP). 2. It proposes a Constraint Programming and Simulation Hybrid System for optimization of complex network-flow systems