Dynamic Predictive Maintenance Model Based on Data-Driven Machinery Prognostics Approach

As an increasing number of manufacturers realize the importance of adopting new maintenance technologies to enable systems to achieve near-zero downtime, machinery prognostics which enables this paradigm shift from traditional fail-and-fix maintenance to a predict-and-prevent paradigm has arose interests from researchers. Machine's condition and degradation estimated by machinery prognostics approach can be used to support predictive maintenance policy. This paper develops a novel data-driven machine prognostics approach to assess machine's health condition and predict machine degradation. With this prognostics information, a predictive maintenance model is constructed to decide machine's maintenance threshold and predictive maintenance cycles number. Through a case study, this predictive maintenance model is verified, and the computational results show that this proposed model is efficient and practical.