Massive Contingency Analysis with High Performance Computing
EMSL Project ID
37300
Abstract
The objective of this proposal is to implement and evaluate new load balancing scheme for performing massive contingency analysis of power grid with high performance computing, to better understand power system cascade failure and to improve power grid reliability. The proposed research directly supports the mission of the DOE office of Electricity Delivery and Energy Reliability in enhancing reliability of the electric infrastructure and facilitating recovery from energy supply disruptions.Contingency analysis is a key function in the Energy Management System (EMS), which assesses the ability of the power grid to sustain various combinations of power grid component failures based on state estimates. As electricity demand continues to grow and renewable energy increases its penetration in the power grid, the personnel operating and managing the power grid are facing some fundamental challenges. Contingency analysis evolves from previous "N-1" analysis (i.e., failures of one component) to "N-x" analysis (i.e., failures of multiple components). Given the sheer number of contingency cases in the problem space and the real-time requirements for power grid operations, today’s industry algorithms and tools are not able to handle comprehensive contingency analysis. Therefore, parallel computers or multi-core computers become the solution for massive contingency analysis.
Our previous work established a frame work of “N-x� contingency analysis with a dynamic load balancing scheme based on a shared task counter updated by atomic fetch-ad-add operations. Case studies of massive up to 300,000-contingency-case analysis were conducted and archived about 500 times speedup on 512 processors. Though more processors would improve the speedup, but they also increase the possibility of counter congestion which could compromise computational performance. Hence, multiple counter scheme and task stealing are proposed to reduce counter congestion time to further increase speed-up. The initial Matlab simulation results show very promising results of using multi-counter scheme. However, the effectiveness of evaluating the proposed research will be limited the processing capacity of the computer.
The super computer, Chinook, in EMSL is a powerful computing resource representing both leading and emerging classes of high-end computer architectures. It provides us a great resource for this task. The proposed research will rely on the Chinook’s capabilities and, if successful, will improve wide-area situational awareness, help power grid operations and ultimately lead to better reliability in the power grids.
Project Details
Project type
Limited Scope
Start Date
2009-08-19
End Date
2009-09-30
Status
Closed
Released Data Link
Team
Principal Investigator
Team Members