基于GA-BP神经网络的明渠流速测点优化研究.

The flow measurement of irrigation channels is mostly based on the current meter method. A large number of flow velocity meters in the channel section can effectively ensure the accuracy of flow measurement, but it also increases the difficulty of flow measurement. Taking a section of water supply open channel of Taiyuan Pumping Station as the research object, based on the measured data of multi-line and multi-point current meter, the cross-section flow rate was set as output, and the flow rate of measuring points at different positions was taken as input to train the GA-BP neural network model, and the verification results were compared with the calculation results of flow rate area method. The results show that the GA-BP neural network only needs the flow velocity of the measuring points at four specific locations to obtain the prediction results that meet the accuracy, effectively optimize the number of measuring points in the open channel, and ensure the accuracy of flow measurement. It provides a new idea for the application of related models in channel flow measurement. [ABSTRACT FROM AUTHOR]