Abstract: Economical operation of a hydropower station is an energy-saving measure to improve water energy utilization efficiency. The load distribution of units in the plant is based on the minimum water consumption, and generally solved using dynamic programming. The Three Gorges hydropower plant features a large number of units and bears a large load; the problem of dimension disaster would easily occur if dynamic programming is adopted. This paper develops a dimensionality reduction method for unit commitment problem based on the group theory. We group the plant’s units by operation characteristics, and calculate the number of essentially different startup and shutdown state combinations and the corresponding combinations based on the Burnside lemma. Then, we use dynamic programming to solve the time-space optimization after dimensionality reduction. Application of this new method to the economical operation of the plant shows it reduces computational cost significantly and responds quickly to the change in the operating conditions of the hydropower system.

Key words: unit commitment, dynamic programming, dimension disaster, group theory