Control (management)

Product (mathematics)

Process (computing)

Engineering

Reliability engineering

Operating system

Computer science

Risk analysis (engineering)

Business

Operations research

Maintenance actions

Mathematics

Preventive maintenance

Political science

Control chart

Warranty

Operations management

Artificial intelligence

Geometry

Control limits

Machine learning

Covariate

Reliability Engineering & System Safety

The development of condition monitoring technology initiates a new tendency towards incorporating condition-based maintenance (CBM) into warranty policies to reduce the manufacturer's expense over the warranty period. However, CBM is not always more cost-effective than time-based maintenance (TBM), or vice versa. This paper investigates three preventive maintenance (PM) policies for a product sold with a free repair warranty and examines under which situation each policy is optimal. The product deteriorates with both age and a time-varying covariate and its failure rate is described by Cox's proportional hazards model. The covariate can only be revealed by condition monitoring. In policy 1, no inspection is performed and PM actions are carried out periodically. Continuous inspection is performed in policy 2 where a PM action is recommended once continuing operation is not cost-effective. Policy 3 applies periodic inspection to select appropriate maintenance action at each inspection epoch. The warranty servicing costs are analyzed in the Markov decision process framework based on a discretization method. The policy structures of the CBM policies are proved to be of control-limit forms. A practical numerical example is provided to illustrate the proposed approach. The results can help the manufacturer design effective warranty policies.

Comparison of three preventive maintenance warranty policies for products deteriorating with age and a time-varying covariate