Advanced Optimization Algorithms for Hydropower Dispatch

Project ID: 486
Principal Investigator: David Harpman
Research Topic: Improved Power Generation
Priority Area Assignments: 2011 (Climate Adaptation)
Funded Fiscal Years: 2010 and 2011
Keywords: None

Research Question

On at least an hourly basis, Reclamation's powerplant operators determine the most economic combination of generator units to operate and set their generation level. This is known as the economic dispatch problem. Substantial gains can be realized from even small improvements in dispatch efficiency. For example, increasing the generation efficiency at Glen Canyon Dam by 2 percent would yield over $4 million (in 2004 dollars) annually.

Recently, a number of new optimization heuristics have been described. Examples of these new technologies include particle swarm optimization (PSO), simulated annealing (SA), genetic algorithms (GA), ant colony optimization (ACO), extremal optimization (EO), differential evolution (DE), tabu search (TS), and the cross-entropy (CE) method. These technologies do not rely on traditional calculus-based approaches but instead are based on innovative search techniques drawn from biological and physical processes. Although computationally intensive, these methods can solve difficult constrained optimization problems, like the economic dispatch problem, quickly and reliably.

We plan to develop computer code for several of the most promising of these new optimization approaches, apply these algorithms to example economic dispatch problems, and systematically assess their performance. The goal of this effort is to identify algorithms which can help guide the hydropower dispatch decision and improve efficiency. Improved dispatching efficiency will result in the generation of more electric power using less water--benefitting all water and power users.

Need and Benefit

The majority of the supervisory control and data acquisition (SCADA) systems at Reclamation hydropower plants do not contain an optimization package to assist with dispatch decisions. Powerplant operators use professional experience and judgment to achieve powerplant efficiency, often relying on the "equal loading" rule of thumb. The operator also tries to minimize start/stop cycles and adjust generation rates while avoiding operation within each unit's rough zone(s).

The SCADA systems at a few of Reclamation's larger and more modern hydropower plants have been retrofitted with optimization software. Reclamation has installed WaterView2000, a commercial optimization package, at Grand Coulee. An optimization routine developed by Steven Stitt, a retired Reclamation employee, has been used at Hoover Dam for several years and was installed at Yellowtail in 2008. Experience has shown that these programs can result in increased power revenues and potentially reduced maintenance and mechanical equipment repairs because operation within rough zones and start and stop cycles is minimized.

The benefits of this proposed research can be extended to powerplants in all Reclamation regions. The goal of this research project is to identify newly emergent algorithms suitable for the real-time solution of the hydropower economic dispatch problem. These algorithms can be embedded in a powerplant's SCADA system and help guide the dispatch decision process. Improved dispatching efficiency will result in the generation of more electric power using less water--benefitting water and power users, as well as the American taxpayer.

Contributing Partners

Contact the Principal Investigator for information about partners.

Research Products

Bureau of Reclamation Review

The following documents were reviewed by experts in fields relating to this project's study and findings. The results were determined to be achieved using valid means.

Advanced Algorithms for Hydropower Optimization (Phase 1 Report) (final, PDF, 3.5MB)
By David Harpman
Report completed on March 21, 2012

The advent of personal computers in the mid-1980s gave rise to an era of unparalleled advances in heuristic optimization research. These new optimization algorithms are not based on traditional calculus-based approaches, but instead have their origins in physical and biological processes. Three promising evolutionary algorithms (EAs) were identified from the emerging literature; the real coded genetic algorithm (RCGA), differential evolution (DE) and particle swarm optimization (PSO). These EAs

Not Reviewed

The following documents were not reviewed. Statements made in these documents are those of the authors. The findings have not been verified.

Briefing for CSU Research Collaborators 4-27-2011 (interim, PDF, 346KB)
By David Harpman
Publication completed on April 27, 2011

Powerpoint briefing on research progress for CSU research collaborators. Presented at Colorado State University on 4-27-2011.

Advanced Optimization Algorithms for Hydropower Dispatch Powerpoint briefing for S&T Office (interim, PDF, 190KB)
By David Harpman
Report completed on

Last Updated: 9/19/16