The development of reference working cycles for agricultural tractors

Abstract

Climate change and the current energy crisis are creating new challenges to agriculture and new technological solutions must be developed to increase agricultural machinery efficiency. Researchers and machinery manufacturers identified electrified powertrains as a possible solution to meet this demand. The development of field-effective electrified powertrains is challenging mostly due to the wide variability of operating conditions of agricultural tractors. While the automotive industry adopted reference driving cycles for the design and evaluation of hybrid powertrains, the tractor industry has not been able to easily record external load in real-world conditions as it requires dedicated systems that cannot be used under prolonged field usages. This study aims to provide a methodology for estimating a reference working cycle from a multi-year dataset using technologies available in current commercial tractors. Data were collected on a tractor used for 3 years of agricultural work. Data were first clustered into work states, then, for each state, signal features from on-tractor sensors were used to extract key factors to compute the reference work state. With an optimisation solver and a hidden Markov model, the reference working cycle that synthesised the real-world tractor use was calculated. This cycle was compared with established cycles for non-road mobile machinery. The new reference cycle better represented real-world tractor usage as it also complied with low engine operations, which are frequent in farming and mostly associated with machine setup. The new reference working cycle permits a reliable estimation of fuel consumption of real-world farming.