Cascade hydropower station peak regulating method based on random continuous optimization strategy

Abstract

The invention discloses a cascade hydropower station peak regulating method based on a random continuous optimization strategy. The cascade hydropower station peak regulating method comprises the following steps: (1) generating an initial solution, selecting water levels in front of a dam from hydropower stations from upstream to downstream as discrete state quantities; taking a minimum power grid residual load mean square error as a target to carry out single reservoir peak regulating optimization calculation; (2) randomly and continuously carrying out optimized search step by step; and (3) updating iteration judgment: if reaching the tail of a time period, finishing a regulating process when a stopping condition of a successive optimization algorithm is reached, and outputting a final result, if the stopping condition of the successive optimization algorithm is not reached, setting a time period number to be a starting time period and turning into the step (2) to carry out the next time of iteration optimization, and if not reaching the tail of the time period, adding one to the time period number and turning into the step (2) to carry out two-phase regulation of the next time period. With the adoption of a constraint handling method provided by the invention, complicated cascade hydropower station coupling operation constraint of the cascade hydropower stations can be completely decoupled, and a constraint handling process of the algorithm is simplified; and a decision-making variable dimension is reduced, a reasonable optimization space is found and the method is suitable for solving a peak regulating problem of a large-scale cascade hydropower station group.

Description

technical field [0001] The invention belongs to the field of optimal operation of hydropower energy and optimal scheduling of electric power system generation, and in particular relates to a cascade power station peak regulation method based on a random continuous optimization strategy. Background technique [0002] The short-term peak shaving problem of cascade hydropower stations has the characteristics of multi-dimensional time and space, complex hydraulic and electric connections between reservoir groups, and huge optimization space. It is a typical large-scale nonlinear programming problem with multiple decision variables, high dimensions, and multiple constraints. For the solution of such problem models, conventional mathematical programming methods and intelligent algorithms are mostly used at present. However, mathematical programming methods are usually difficult to deal with complex constraints with a high degree of coupling, and are prone to the "curse of dimensio...

AI Technical Summary

Problems solved by technology

However, mathematical programming methods are usually difficult to deal with complex constraints with a high degree of coupling, and are prone to the "curse of dimensionality" problem. Although the model can be reduced in dimension by using step-by-step optimization techniques, the solution accuracy of traditional successive optimization algorithms strongly depends on Discrete precision of decision variables, reservoir group size and constraint coupling degree, the resulting problem of low solution efficiency limits its application in short-term joint peak-shaving scheduling of large-scale hydropower stations

Although the intelligent algorithm has been successfully applied in reservoir scheduling, its solution efficiency is also limited by the dimension of decision variables and the scope of optimization space, and it is easy to fall into premature and local convergence.

Images