A reliability engineering case study of sugarcane harvesters

Abstract: The present study aimed to analyze factors associated with the equipment failures of the sugarcane harvester, whose machineries has high importance in the harvest process and cost involved. Part of the data was originally provided by a company located in the countryside of Sao Paulo State, from two machines, collected from January 2015 to August 2017, corresponding to 2.5 crops. The overall dataset was obtained from three different sources: a stop-tracking system, which provides the track of a preventive and corrective maintenance historical of the analyzed equipment; telemetry data of the equipment, captured through embedded computer systems, installed in the machine’ type under study, which provide information on its operation; and meteorological data from the Brazilian National Institute of Meteorology. Multivariate analyzes were used such as principal components and multiple regression models, therefore creating a model for prediction considering the next equipment’ break, then pointing to causes of process failures. Thus, the results point to some improvements concerned with individualized reliability scheme in order to reduce the number of corrective stops given the equipment.摘要：本研究旨在分析与甘蔗收割机（sugarcane harvester）设备故障相关的影响因素，此类机械设备在收割作业中占据核心地位且涉及高昂成本。部分原始数据由位于圣保罗州乡村的一家企业提供，涵盖两台设备2015年1月至2017年8月的运行记录，对应2.5个收割季。本次完整数据集取自三个不同渠道：其一为停机跟踪系统（stop-tracking system），可记录所分析设备的预防性维护（preventive maintenance）与纠正性维护（corrective maintenance）历史；其二为设备遥测数据（telemetry data），通过安装于目标机型内的嵌入式计算机系统（embedded computer systems）采集，可提供设备运行相关信息；其三为巴西国家气象局（Brazilian National Institute of Meteorology）的气象数据。研究采用主成分分析（principal components）、多元回归模型（multiple regression models）等多元统计分析方法，构建了用于预测设备后续故障的模型，并据此明确了收割流程故障的成因。研究结果表明，可通过推行个性化可靠性维护方案，减少设备的纠正性停机次数，从而优化作业流程。