Large-scale conveyor belt system maintenance decision-making by using fuzzy causal modeling

Abstract

Conveyor belt systems have been significantly developed for decades and are playing a critical role in todays large-scale continuous transport systems. Traditional conveyor belt monitoring focuses on catastrophic failure. Failure alarms and maintenance decisions are submitted separately without revealing relationships of monitored events. Causal modeling such like Bayesian methodology provides intuitive and mathematically sound tools to understand complex relations between uncertain variables and failure causes. However to derive inference knowledge for validating causal modeling is difficult. This paper introduces a causal modeling methodology based on Bayesian inference to diagnose failure situation and decide relative maintenance operations for large-scale conveyor belt systems. Fuzzy logic is applied to estimate the likelihood density function which is usually hard to be obtained for causal inferences. This methodology is applied as a maintenance decision-making process in intelligent conveyor belt monitoring system. An application of indicating the main failure cause and suggesting maintenance operation for conveyor belt emergency braking system is presented.