Failure prediction is one of the key challenges that have to be mastered for a new arena of fault tolerance techniques: the proactive handling of faults. As a definition, prediction is a statement about what will happen or might happen in the future. A failure is defined as "an event that occurs when the delivered service deviates from correct service." The main point here is that a failure refers to misbehavior that can be observed by the user, which can either be a human or another computer system. Things may go wrong inside the system, but as long as it does not result in incorrect output (including the case that there is no output at all) there is no failure. Failure prediction is about assessing the risk of failure for some time in the future. In my approach, failures are predicted by analysis of error events that have occurred in the system. As, of course, not all events that have occurred ever since can be processed, only events of a time interval called embedding time are used. Failure probabilities are computed not only for one point of time in the future, but for a time interval called prediction interval.

PurposeThe purpose of this paper is to optimize the preventive maintenance based on fault tree (FT)–Bayesian network (BN) reliability for sugarcane harvester machine as a fundamental machine in the sugar industry that must be operated failure-free during a given period of the harvesting process.Design/methodology/approachTo determine machine reliability using the algorithm developed based on mapping FTs into BNs, the common failures of 168 machines were carefully investigated over 12 years (2007–2019). This algorithm was then used to predict the harvester reliability, estimate delays by machine downtimes and their consequences on white sugar production losses that can be reduced by optimizing the preventive maintenance scheduling.FindingsThe optimization of preventive maintenance scheduling based on estimated reliability of sugarcane harvester machines using FT–BNs can reduce white sugar production losses, the operation-stopping breakdowns and the downtime costs as a crisis that the sugar industry is facing.Practical implicationsMachine reliability gradually decreased by 31.08% approximately, which resulted in a working time loss of 26% in the 2018–19 harvesting season. In total, the white sugar losses were estimated as 204.17 tons for burnt canes and 114.53 tons for green canes. The losses of the 2018–19 harvesting season have been 11.85 times greater than the first harvesting season. The proposed maintenance interval for critical subsystems including the hydraulic, chopper and base cutter were obtained as 1.815, 1.12 and 1.05 h, respectively.Originality/valueIn this study, a new approach was used to optimize preventive maintenance to reduce delays and their implications upon costs in time, inconvenience and white sugar losses. The FT–BNs algorithm was found a useful tool that was over-fitting of failure occurrence probabilities data for sugarcane harvester machine.

No abstract

No abstract

No abstract