Predicting the draught requirement of tillage implements in sandy clay loam soil using an artificial neural network

Abstract

A 5–9–1 artificial neural network (ANN) model, with a back propagation learning algorithm, was developed to predict draught requirements of different tillage implements in a sandy clay loam soil under varying operating and soil conditions. The input parameters of the network were width of cut, depth of operation, speed of operation, soil moisture content and soil bulk density. The output from the network was the draught requirement of the individual tillage implement. The developed model predicted the draught requirement of mouldboard plough, cultivator and disk harrow with an error < 6.5% when compared to the measured draught values, whereas the American Society of Agricultural and Biological Engineers (ASABE) equation predicted these draught values with an error > 30%. Such encouraging results indicate that the developed ANN model for draught prediction could be considered as an alternative and practical tool for predicting draught requirement of tillage implements under the selected experimental conditions in sandy clay loam soils. Further work is required to demonstrate the generalised value of this ANN in other soil conditions.