| With the rapid development of our country's economy,the continuous improvement of people's living quality and the increasement of social consciousness of water saving,energy saving,people's expectations for better water supply system become higher.How to save the cost as much as possible on the premise of stable water supply has became the focus of social concern.As important parts of the water supply system,seeking a more scientific and reasonable operation mode for pumping station and clear water tanks is necessary,and also meet the further development of water supply system in our country.At present the management level of pumping station in our country is relatively backward,the traditional scheduling mode that depends on experience management has been unable to meet the growing demand for the quality of water supply,optimizing the operation of pumping station is facing new challenges.Based on water supply system of Wuhu Sanshan District,this paper builds forecasting model of hourly water consumption,analysis model of water supply network,optimal scheduling model of pumping station and joint optimal scheduling model of first pumping station,secondary pumping station and clear water tanks by the methods of BP neural network,genetic algorithm,dynamic programming algorithm and so on.A systematic study of all kinds of contents involved in optimal scheduling of pumping station has been made in this paper.The main contents of this paper are as follows:(1)Forecasting of hourly water consumption: according to the change rules of hourly water consumption of Wuhu Sanshan District and various influence factors,a forecasting model of hourly water consumption of Wuhu Sanshan District has been built by BP neural network.The model adopts three-layer network structure to give a scientific forecast of water consumption in the future.The model uses the previous 24 hours' data of hourly water consumption as input,and the data computed by the model for the next hour as output.With the practical data of Wuhu Sanshan district,it can be proved that the model has higher forecasting accuracy.(2)Analysis model of water supply network: according to the locations of the pressure monitoring points in Wuhu Sanshan District,this paper selects BP neural network to build the analysis model of water supply network of Wuhu Sanshan District.The model contains a network structure of three layers.The hourly average pressure of seven pressure monitoring points and the hourly water flow of Sanshan waterworks are used as input and the output is the hourly average pressure of Sanshan waterworks.Verified by the practical example of water supply network of Wuhu Sanshan District,the model can simulate the network running conditions,which can be used for optimal scheduling.(3)Optimal scheduling of pumping station: this paper researches the modeling method of optimal scheduling models about first pumping station and secondary pumping station.The optimal scheduling models of first pumping station and secondary pumping station of Sanshan waterworks are built under the constrained conditions of water flow,water pressure,performance of pumps,and so on,which can be solved by the method of genetic algorithm.(4)Joint optimal scheduling of the first pumping station,secondary pumping station and clear water tanks: considering the operational characteristics,the full use of the difference between peak and valley price for electricity and the storage vulume in clear water tanks,a hierarchical optimal scheduling model of first pumping station,secondary pumping station and clear water tanks is proposed.The model is built with the objective of minimizing the daily total operational electric cost and the on-off frequency of pumping units.Solved the model with both dynamic programming algorithm and genetic algorithm.The optimal operation scheme of fchajuirst pumping station and secondary pumping station is determined.The calculation results of the practical example show that the economic benefits of the optimal scheme are greatly improved. |
PDF Full Text Request