Multi-uncertainty reservoir scheduling rule extraction method and system

Abstract

The invention belongs to the field of reservoir scheduling operation, and discloses a multi-uncertainty reservoir scheduling rule extraction method and system. The multi-uncertainty reservoir scheduling rule extraction method comprises the steps: calculating an optimal reservoir scheduling scheme according to historical runoff; obtaining probability distribution of reservoir forecast incoming water through a probability forecast model; taking the upstream water level of the confronted-period reservoir and the incoming water probability forecast of the confronted-period reservoir as input variables; and establishing a Bayesian neural network model by taking the optimal decision water level of the reservoir in the next time period as an output variable, and training the Bayesian neural network model by adopting a simulation-based training mode to obtain a reservoir scheduling rule considering incoming water uncertainty and model parameter uncertainty at the same time. According to the multi-uncertainty reservoir scheduling rule extraction method, forecast incoming water, uncertainty information of the forecast incoming water and uncertainty of scheduling model parameters can be fullyconsidered, and reservoir scheduling rules are extracted according to a historical optimal scheduling scheme; and the obtained reservoir scheduling rule considering runoff uncertainty and model parameter uncertainty can provide decision support for real-time scheduling personnel.

Description

technical field [0001] The invention belongs to the field of reservoir dispatching operation, and in particular relates to a method and system for extracting multi-uncertainty reservoir dispatching rules. Background technique [0002] At present, the closest existing technology: deterministic reservoir optimal dispatching is based on the optimal solution under the condition of perfect runoff forecast for the entire dispatching period, and the runoff entering the reservoir at each period is regarded as a deterministic runoff process. However, the deterministic runoff process during the dispatching period cannot be obtained in actual dispatching, so deterministic reservoir optimal dispatching is difficult to apply to the real-time dispatching process. Reservoir scheduling rules can make scheduling decisions based on the state of the reservoir in the current period, and are more widely used in the real-time scheduling process. [0003] The current dispatching rules are usually...

AI Technical Summary

Problems solved by technology

However, due to the randomness and uncertainty of runoff and the uncertainty of model parameters, the decisions obtained by the current dispatching rules are accompanied by unknown uncertainties and risks. How to consider the randomness and uncertainty of runoff in the process of extracting dispatching rules Uncertainty and uncertainty of model parameters have become the difficulty of current research

[0004] To sum up, the problems existing in the existing technology are: (1) Deterministic reservoir optimal dispatching is based on perfect runoff forecasting conditions, which is too ideal to be applied to real-time dispatching problems

[0005] (2) The extraction technology of the existing dispatching rules does not fully consider the forecasted water and its uncertainty information and the uncertainty of dispatching model parameters, resulting in poor reservoir dispatching accuracy

Moreover, it cannot provide data information support for reliable reservoir scheduling

[0006] (3) When the existing statistics, regression or machine learning methods consider the input uncertainty, most of them first use the trained model to make collective predictions for different input conditions, and fail to consider the input uncertainty in the model training. Certainty

Images