Design and Application of Falling-Sill and Expanding Stilling Pool Based on BP Neural Network

Increasing abruptly expanding width and depth of the sill of bottom-diffusion bottom-flow stilling pool can effectively reduce bottom flow velocity in bottom of stilling pool and improve flow pattern of outflow water，and reduce length of stilling pool to a certain extent.Based on experimental research results，the BP neural network theory was used to establish BP neural network prediction model with sudden expansion width，sill depth and measuring point distance as model input parameters，and bottom flow rate as output parameter.Results showed that average relative error between experimental value and predicted value of predicted bottom velocity model parameters was less than 10%，and determination coefficient R2 reached to 0.9776，which meant that prediction model based on intelligent algorithm could complement a hydraulic model test research well.On this basis，effects of sudden expansion width，change of drop sill depth，different combinations of drop and expansion combinations on pool length of stilling pool were further given.Relatively speaking，effect of increasing burst width on reduction of stilling pool length was less than that of increasing fall depth.At the same time，increasing bursting width and fall depth could more effectively reduce required length of stilling pool.