A New Intelligent Model for Computing Crack in Compacted Soil-Biochar Mix: Application in Green Infrastructure

There have been studies analysing water retention and cracking in soil-biochar mix, but there is lack of model for estimating cracking in these kind of mixture. These models can be useful as it can be directly inputted in governing equations of seepage analysis for further analysing stability of green infrastructure. The objective of this study is to develop a model for computing crack as function of water content, suction and biochar content. Neural network-based modelling was adopted to achieve the objective. A series of experiments were conducted to quantify cracking in soil biochar composite using a novel crack intensity factor as a function of water retention and soil suction in five different soils with varying biochar content (i.e., 0%, 2%, 5%, 10% and 15%). The biochar was obtained from an invasive weed (i.e., Water hyacinth). The data obtained from the experimental study was then used for developing model using Artificial neural networks (ANN) technique. A single ANN model was developed and validated with the testing data. The model was found to give satisfactory performance. This model can be useful in improving the water balance calculation in green infrastructure as well as agricultural fields (subjected to extreme drying-wetting season).