Metamodel-based design of alluvial channels at incipient motion subjected to seepage

摘要

The design of an alluvial channel affected by seepage requires information about five basic parameters: particle size, water depth, energy slope, seepage velocity, and average velocity. The conventional approach to predicting the incipient motion in an alluvial channel cannot be applied in the case of a channel affected by seepage. Metamodelling techniques are nowadays widely used in engineering design to simulate a complex system. Here, a metamodel is described which employs the radial-basis function (RBF) network to predict the seepage velocity and energy slope based on experimental data under incipient motion conditions. It was found that the model fits experimental data very well and provides predictions for the design. With the help of the metamodel generated by the RBF network, design curves based on the RBF metamodel are presented for use in designing an alluvial channel when it is affected by seepage.