Prescriptive Maintenance of CPPS by Integrating Multimodal Data with Dynamic Bayesian Networks

The complexity and data-driven characteristics of Cyber Physical Production Systems (CPPS) impose new requirements on maintenance strategies and models. Maintenance in the era of Industry 4.0 should, therefore, advances prediction, adaptation and optimization capabilities in horizontally and vertically integrated CPPS environment. This paper contributes to the literature on knowledge-based maintenance by providing a new model of prescriptive maintenance, which should ultimately answer the two key questions of ""what will happen, when? and ""how should it happen?"", in addition to ""what happened?"" and ""why did it happen?"". In this context, we intend to go beyond the scope of the research project ""Maintenance 4.0"" by i) proposing a data-model considering multimodalities and structural heterogeneities of maintenance records, and ii) providing a methodology for integrating the data-model with Dynamic Bayesian Network (DBN) for the purpose of learning cause-effect relations, predicting future events, and providing prescriptions for improving maintenance planning.