Modeling and optimization using artificial neural network and genetic algorithm of self‐propelled machine reach envelope

Abstract

Self‐propelled machinery, which is a crucial component of agricultural mechanization, is essential for enhancing production. Few farming operations involve the operation of riding‐type self‐propelled machinery for a prolonged period of time (6–8 h/day) and in a fluctuating weather condition in field condition. To alleviate the operator's workload, an ergonomically designed self‐propelled reach envelope was developed using various configuration simulator systems. To model the process and achieve the desired output for the development of simulator of self‐propelled machinery, artificial neural network (ANN) technique was employed. The simulator was tested using a range of control lever positions (0–23) and engine speeds (1600 and 2000 rpm). The study investigated how these process parameters influence the oxygen consumption rate (OCR) of female farm workers and determined the optimal operating parameters corresponding to the 100% OCR of female agricultural workers. Genetic algorithm (GA) was used along with the developed ANN model to forecast the OCR of agricultural workers. The experimental mean ± SD values were found 0.79 ± 0.04 L/min and predicated ANN values were found to be 0.776 ± 0.06 L/min for OCR, respectively. It is evident that not all operating parameters have the same effect on OCR. The training, testing, validation, and overall data sets employed in the ANN model and yielded correlation coefficients (R2) of 0.50889, 0.44229, 0.39029, and 0.48323, respectively. The goodness of fit was another crucial factor in evaluating the ANN tools. The ANN model demonstrated the lowest mean‐square error value that indicates higher precision and predictive power. The results satisfy the most suitable optimal parameters for the OCR of female workers operating the self‐propelled machinery simulator.