Methodology for designing accelerated structural durability tests on agricultural machinery

• Accelerated structural testing was implemented on a four-rotor swather.
• A test design optimisation method was used based on the accumulated damage values.
• It proved difficult to reproduce the same loads to the rotors under swathing mode.
• A time acceleration factor of around 3.3 was achieved.
• Test design considering damage variance gave an acceleration factor equal to 3.1

Structural durability assessment is one of the last stages before an agricultural machine prototype reaches the market. Accelerated structural testing (AST) aims at reducing the time and resources required for this stage. According to existing AST methodologies, strain measurements are used to characterise machine loads under real-world operating conditions, and calculate resulting accumulated fatigue damages. An operation profile is defining the conditions to be monitored but also the target damages of the accelerated testing. Next, rainflow cycles are extrapolated to include non-measured high-amplitude loads. Finally, the machine prototype travels on suitable proving grounds to replicate real-world service loads. The number of laps required to reach the target damage values is the result of optimisation, given the fatigue damages accumulated during each lap.In this paper the above AST methodology was implemented on a four-rotor swather, which is an agricultural implement that drastically changes structure configuration during its working life, depending on its operating mode. Furthermore, recognising the fact that the damage accumulated during each lap varies, automated test facilities were utilised, and Monte-Carlo sensitivity analysis was introduced as part of the AST methodology, to study the effects of damage-per-lap variance on the required numbers of laps calculated via optimisation. When average values were used for lap damages, the total testing time was 1228 h with an acceleration factor of 3.3. However, conservative test design using the 99.9th percentile of the testing time simulation results, required 7.1% longer testing time, leading to a lower acceleration factor equal to 3.1.