The Research Of Two-step Optimal Operation In Municipal Water Supply System

Energy saving and optimal operation of urban water supply system oriented the development of water industry in recent years.At present,the management of most waterworks in our country is relatively backward,and human-based operation is still the main mode of dispatch.The unbalanced distribution pressure and high hydraulic pressure at night result in obviously energy waste.Research on optimal operation of water supply system is aimed at ensuring residents using water as normal,optimizing pipe network pressure distribution and saving water supply energy consumption.It is of great significance for improving the economic benefits of water supply companies and the scientific management of water supply networks.Based on macroscopic model of pipe network,this paper studies the two-step optimal dispatch of water supply system,including urban water demand forecast,hourly macroscopic models of pipe network,the first-step optimal dispatch model and the second-step optimal dispatch model.Firstly,the water demand forecast model was established based on the least squares support vector regression algorithm.The water demand data of seven consecutive days were used as training samples,and the data of the eighth day was used as forecasting target.The prediction accuracy of the tested model was in line with the requirements of optimal scheduling.An hourly macroscopic models based on BP neural network was established.According to the variation law of 24-hour water demand data,one day was divided into six periods to meet the actual dispatch requirements of the pump station.The data of pipe network pressure monitoring points and outflow rate of waterworks at each period are taken as input,and the factory pressure of waterworks is taken as output to complete domestication process of macroscopic models.Taking the total cost of water supply as the optimization objective,an optimal scheduling model is established based on the balance between water supply and demand of pipe network,pressure constraints of pressure monitoring points and macroscopic model of pipe network.The optimal scheduling model is transformed into non-constraint problem by penalty function method and solved by genetic algorithm to get the outflow rate and pressure of pump station as the second-step optimal scheduling target.The second-step optimal scheduling takes the lowest operating cost of pump station as the objective function,and the optimal pump combination scheme was solved by genetic algorithm.Under the precondition,the water demand and hydraulic pressure are satisfied,the pumps are all operated in high efficiency section and the reduced energy waste.Finally,taking the Cangzhou water supply system as an example,the two-step optimal scheduling model is tested.The optimized pipe network pressure distribution is balanced and the pump station power consumption is lower,which verifies the practical value of the model.