Energy-efficient control of hydrostatic transmission of a front-end loader machine using machine learning algorithm and its sensitivity analysis

Abstract

This paper highlights development of 2-DOF and 3-DOF energy management controllers using a Machine Learning (ML) algorithm for closed-circuit hydrostatic transmission of Front-End Loader (FEL) machine. Objective of energy management controller, is to keep prime-mover, pump and hydro-motor in its efficient zone irrespective of variation of load-speed at drive end. Based on duty cycles of the transmission and sensitivity analysis of drive components, GPR algorithm shows better prediction (within 3%) of optimized commands of hydro-motor and pump displacements and their operating speeds, which results in its efficient operation. Validated model of hydraulic transmission components is integrated with verified model of IC engine to investigate saving in fuel consumption through designed controllers. Comparative study between 2-DOF and 3-DOF controlled transmission systems infers that fuel consumption for engine in 3-DOF controlled system decreases up to 34% of that of 2-DOF with a 20%–30% relative increase in overall drive efficiency.