Optimal Operation Of Pumping Stations Based On Genetic Algorithms And Artificial Neural Networks

China is a large country with complex natural condition. There are much inconsistency between the distributing of water resource and the requirement of industry, agriculture and life on time and space. To solve the inconsistency, many delivering water projects have been built until now. However, when running, them also need a great deal of energy. As the main of delivering water projects, pumping station consumes more energy sources than another proportion. So we must research how to efficiently operate pumping station and decrease the cost of delivering water.The problem of optimal operation is to conform the optimal movement patterns of pumping station under the condition of given flowrate and head in order to decrease the expenditure. In the pumping station of the plain area of our country, the most type of pump is adjustable-blade axialflow pump. When the water head and the flow are given, we can regulate the pump's blade mounting angle to control the discharge of pump in this pumping station that contains only one type of pump. We also can determine the number of running machines with machine's optimal combination. So we can reduce its energy consumption, and make it run optimally and economically.The paper introduces the actuality and the study means of optimal operation of pumping station. Aiming at the much cost of locale testing and the difficulty of actualization, a new conversion formula of the performance of prototype pump sets is established on the base of model testing. On the point of system, this paper establishes the mathematical model of Jiangdu Drainage and Irrigation Station and uses the method of Genetic Algorithms to study the economic rule of pumping operations. This paper takes the operation data of the Jiangdu Fourth Pumping Station as an example, analyses the optimal distribution of load in the machines ofpumping station. To make optimization operation based on actual condition of the series of pumping stations in 1999,339% could be saved comparing with its actual energy consumed in the same year. Aiming at the frequent change of flowrate and head and the slow calculational speed of the traditional methods, based on the result of optimal operation, this paper predicts the optimal movement patterns of the Jiangdu First Pumping Station by the method of Artificial Neural Networks. The successful application of this technique improves the rate of solving the problem, to realize the optimal operation of pumping station.