Simulation operation method for large-scale reservoir group in main stream and tributaries of river basin

Abstract

The invention discloses a simulation operation method for a large-scale reservoir group in a main stream and tributaries of a river basin, and belongs to the field of optimal operation of a hydropowersystem. The method comprises the following steps that (1) construction of reservoir operation function is conducted, analysis of relevant factors affecting reservoir outflow is conducted, correlationanalysis is conducted, and determination of input factors of each reservoir operation function is conducted; (2) construction of a neural network model is conducted according to the input factors ofthe operation function, optimization of neural network parameters is conducted by adopting an adaptive moment estimation algorithm, training is conducted on the constructed neural network by using historical operation data of the reservoir, and the trained neural network is used as fitting function of the reservoir operation function; and (3) according to the fitting function of the reservoir operation function, a spatial topological structure and the constraints of the reservoir operation, the simulation operation model of the reservoir group is established to simulate the operation process of the basin reservoir group step by step. By means of the simulation operation method, the fitting accuracy is improved significantly, and the operation plan of the large-scale reservoir group in themain stream and tributaries of the river basin can be more accurately described in the case of the unknown operation law.

Description

technical field [0001] The invention belongs to the field of optimal dispatching of hydropower systems, and more specifically relates to a method for simulating dispatching of large-scale reservoir groups in main and tributary streams of a river basin. Background technique [0002] With the successive completion and operation of large-scale reservoir groups in the basin, the evolution law of the hydrological process and the spatio-temporal pattern of the basin have changed. They belong to different power generation groups. Under the existing management system, real-time sharing of dispatching operation data cannot be achieved, which adds uncertainty to the formulation of downstream reservoir dispatching plans. Therefore, it is necessary to conduct simulation dispatching research on reservoirs whose dispatching conditions are unknown. [0003] At present, the most commonly used methods for simulating reservoir scheduling operations are scheduling diagrams and scheduling funct...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects of the prior art, the object of the present invention is to provide a large-scale reservoir group simulation scheduling method for main and tributary streams in the basin, which is based on the improved neural network method (Adam-DNN) of adaptive moment estimation to fit the reservoir scheduling function , which aims to solve the problem of traditional neural network falling into local optimal solution and gradient disappearance, and improve the fitting accuracy of scheduling function

Images