Data-Driven Simulator: Redesign of Chickpea Harvester Reels

Abstract

Conventional redesign methodologies applied on the grain harvester headers for the mechanical harvesting of chickpeas cause its progress to not be as rapid and technological. This paper presents a hybrid modeling-optimization methodology to design harvester reels for efficient chickpea harvesting. The five fabricated headers were tested in both real and virtual modeling environments to optimize the operational parameters of the reel for minimum losses. Harvesting losses data gathered from chickpea fields over ten years of trials were fed into a fuzzy logic model, which in turn was merged with simulated annealing to develop a simulator. To this end, simulated annealing was used to produce combinations of reel diameter and number of bats, to be fed into the fuzzy model until achieving a minimum harvesting loss. The proposed model predicts the reel structure measured in-field evaluation, which fits well with the previously established mathematical model. A significant improvement in harvesting performance, 71% pod harvesting, validates the benefits of the proposed fuzzy-simulated annealing approach to optimize the design of grain harvester headers.