Reservoir forecast dispatch grey dynamic programming method

Abstract

Grey dynamic programming method of reservoir forecast dispatching, disperse storage of reservoir, the dispatching period is divided into n periods, use dynamic programming reverse recursive method during the entire scheduling process to calculate generating capacity generated by reservoir on the time-interval; and sequential calculate optimal generating capacity from the second to the last time-interval, gain the optimal dispatching path from calculating the optimal generating capacity at the last time-interval. The invention considers the uncertainty runoff forecast, its runoff forecast value is gray number, making the final calculated the optimal capacity is also a grey number, it appears to increase the uncertainty, but their counterparts dispatching curve is exclusive, and more conducive to the implementation of reservoir dispatching programme by production department, meet the actual situation requirement of engineering; compared with stochastic dynamic programming, this method considers forecasting value of whole calculation time-interval, make fully use of known conditions to make the dispatching curve more close to the actual ideal curve.

Description

technical field [0001] The invention relates to a reservoir optimization scheduling method, in particular to a gray dynamic programming method for reservoir forecast scheduling. Background technique [0002] Reservoir scheduling is carried out under the condition that it is difficult to accurately predict the future runoff, and the process of runoff change is quite complicated, and it is difficult to describe it accurately. For reservoir scheduling, the dynamic programming method is widely used at present. The essence of dynamic programming is to divide and conquer and solve redundancy. For recurring sub-problems, they are solved only when they are encountered for the first time, and the answers are saved, and they can be directly quoted when they are encountered again without having to solve them again. Designing a standard dynamic programming algorithm can usually be carried out according to the following steps: (1) divide the stage; (2) select the state; (3) determine th...

AI Technical Summary

Problems solved by technology

The deterministic dynamic programming method thinks that the runoff forecast value is a definite value, but different forecast methods have different forecast values, and some even vary greatly, which often makes production units at a loss, and this processing method that assumes that the future reservoir runoff is known is difficult. In line with the actual engineering situation, only an idealized optimal solution can be obtained; in the stochastic model, according to whether the runoff is related and whether there is forecast, it can be divided into the following four situations: ① The natural inflow flow of the reservoir is independent of each other, and according to certain probability distribution; ②The natural inflow flow of the reservoir is independent of each other, only the forecast value of the current period, and other periods are calculated according to the probability distribution; ③Runoff in adjacent periods is related and has no forecast; ④Runoff in adjacent periods is related and has A period of time forecast, stochastic dynamic method requires a large amount of historical runoff data, which is difficult to achieve in new power stations

Images