The use of neural network to predict the behavior of small plastic pipes embedded in reinforced sand and surface settlement under repeated load

Abstract

This paper presents a feed forward back-propagation neural network model to estimate the vertical deformation of high-density polyethylene (HDPE) small diameter flexible pipes buried in reinforced trenches and settlement of soil surface (SSS) subjected to repeated loadings to simulate the heavy vehicle loads. The experimental data show that the vertical diametral strain (VDS) of pipe embedded in reinforced sand and SSS are dependent on relative density of the sand, number of reinforced layers and height of embedment depth of pipe. Therefore in this investigation, the value of VDS and SSS are related to the above parameters. In the developed neural network, the neurons of the input layer represent the relative density of the sand, number of reinforced layers and height of embedment depth of pipe. One neuron is used in the output layer to represent the value of VDS or SSS. In the entire test, the intensity of applied repeated loads is constant (5.5 kg/cm 2, equal to maximum traffic load). A database of 72 experiments from laboratory tests were utilized to train, validate and test the developed neural network. The results show that predictions of VDS and SSS using the trained neural network are in good agreement with experimental results. A comparative evaluation of artificial neural network (ANN) and regression model show that the predictions obtained from the neural network are better than regression model compared to those obtained with the experimental results.