摘要

This paper develops a short-term model for the optimal operation of hydropower reservoirs. The model incorporates Muskingum channel routing, flow ramping constraints attributable to navigation requirements, and constraints on turbine operating regions, up/down time, and startup number. Lagrange multipliers are introduced to decompose the primal problem into a hydro subproblem and many individual unit subproblems, which are solved by the improved simplexlike method (SLM) and dynamic programming (DP). The Lagrange multipliers in the dual problem are updated by using an improved subgradient method, and the dual solution is modified to be feasible to the primal problem by using an efficient iterative policy. The method proposed has advantages in dealing with multiple identical units by defining united-generators to give a lower dimension and less execution time of the hydro subproblem. In a case study, the derived solution is very close to the optimal one, with a distance in benefit less than 0.55%. DOI:10.1061/(ASCE)WR.1943-5452.0000142.