Intelligent Control Technology Of Urban Open-channel System

Water resource is a vital strategic resource.Open-channel is an essential part of urban drainage system,which has a profound historical significance for the reasonable management,scheduling and use of water resource.So,in order to guarantee the daily water supply,it is important to construct and control the urban drainage system reasonably.In the aspect of open-channel facility construction,China lags behind other developed country.Most of the pump stations need manual duty.In the aspect of scientific research,most of the domestic and overseas researches only take the control of water level,flow or volume into consideration,while neglect the control cost relating to the energy consumption.On the other hand,because the difficulty of obtain the precise parameter and the randomness of the rainfall,most of the control algorithms based on the model are hard to apply in the practical engineering.This thesis has done some studies which are related to the problem mentioned above.1.In order to reconstruct the backward manual duty pump station,this thesis designed a solu-tion which is used to build an urban open-channel remote monitoring system.This thesis illustrated the structure of the system and elaborated the design process of configuration platform,interme-diate database and ODBC data channel in center monitoring server.Eventually,this solution was applied in the practical engineering,which has achieved the goal.2.For the case of the open-channel model and the predicted rainfall are both available,this thesis proposed an improved online dynamic programming control algorithm for single open-channel,in which the objective is to follow the desired water level and minimize the control cost related to the energy consumption.By applying the state feedback and the rolling updating meth-ods,the improved dynamic programming can decrease the influence of rainfall prediction error.This thesis did a simulation,which showed that the improved dynamic programming algorithm is better in the aspect of water level following and energy saving.3.For the case of the open-channel model are unavailable with the disturbances of rainfal-l and urban drainage discharge,this thesis proposed an adaptive dynamic programming control algorithm for multi open-channels network,in which the objective is also to follow the desired water level and minimize the control cost related to the energy consumption.The adaptive dy-namic programming controller is a close-loop structure,which contains action neural network and critic neural network two RBF neural networks.The adaptive dynamic programming algorithm was testified in simulation by compared to an traditional PID algorithm.The result shows that the proposed scheme outperforms the traditional PID controller in the aspect of water level following and energy saving.