Study On Reservoir Optimal Operation Based On Water Quality Intelligent Prediction

The management and utilization of water resources are improved more and more scientifically and effectively through numerous constructions of water conservancy. However, these constructions also bring series of environmental problems. In order to decrease and clear those impacts on the environment, plenty of studies have been focused on this field, which makes great sense to the research of reservoir optimal operation which focused on water quality.This paper is mainly based on a real watershed management instance - Xiangxi River which is located in Three Gorges Region. The objective of this study is to establish an effective prediction model for simulation the river water quality change together with reservoir optimal operation. The main contents of this research are as follows:1. Generate 2-Dimensional hydrodynamic numerical model.In this model the finite element method is used to generate a study area, and water dynamic models are used to predict basic fluid dynamic fields when Three Gorges Dam's water level evaluates up to 175m. Hydrodynamic characteristics are also analyzed by calculating fluid dynamic field under different operation modes. At last a conclusion is got that peak regulation can enhance water level fluctuation in reservior and branch region and also can improve water exchange.2. Generate 1-Dimensional water quality model.In this model we get a summary of concentration change law of Xiangxi River water quality under Three Gorges Powerplant's different peaks. And water quality change rate is introduced to describe water quality change, after that make certain the factors and their weights on pollutant transpiration/diffusion in tributary rivers which is caused by Three Gorges. Also valley load flow, peak regulation, level in front of dam and concentration boundary impacts on branch water quality are analyzed in depth.3. Water quality prediction model is formed by using improved BackwardPropagation Artificial Neural Network(BP-ANN) algorithm whose samples come from characteristic real data, and its model is based on sample training and error analysis. As a result, we get a stable and high precision model.4. Reservoir optimal operation model based on water quality intelligent prediction is established. Based on Three Gorges Dam's Reservior optimal operation principle, objective and multiple constraints the water quality pridiction model is established. And Genetic algorithm is used in this model to get a solution. Through analyzing the optimal operation's impacts on water quality based on the aim of improving electricity generation and water quality, the reservoir optimal operation scheme is achieved to maxmize the comprehensive benefits when in non-flood season.