Joint optimization of opportunistic maintenance and repair channels assignment for continuous process manufacturing systems subject to stochastic durations and epistemic uncertainty

For continuous process manufacturing systems (CPMSs) such as steel, the fact that the production process cannot be stopped places higher demands on machines' stable and reliable operation. The break between adjacent production batches is used as the maintenance “opportunity” to improve machines' performance. The selected maintenance actions for machines are assigned to finite and heterogeneous repair channels during each “opportunity”. However, existing studies ignored the internal epistemic uncertainty related to degradation and the external randomness of production and maintenance durations, which could lead to overestimation of system profit and reliability. Therefore, a novel joint optimization model of opportunistic maintenance and repair channels assignment for CPMSs with the above two uncertainties is proposed. The uncertain maintenance opportunity under stochastic production and maintenance duration is characterized as an uncertain opportunity time window (OTW). The optional maintenance actions during uncertain OTWs include do nothing, imperfect maintenance, and perfect maintenance, which take various times and costs to be executed by different repair channels. Moreover, a stochastic fuzzy flow manufacturing network (SFFMN) is established to evaluate machine reliability under stochastic and epistemic uncertainties. The joint optimization model to maximize the expected system profit is addressed by the tailored dual population estimation of distribution algorithm based fuzzy fitness (DPEDAbFF). Finally, a case study of the hot rolling manufacturing system validates the proposed approach in reducing unplanned downtime risks while improving profitability.