Robotic System Reliability Analysis and RUL Estimation Using an Iterative Approach

This paper presents a novel methodology to evaluate robotic system reliability and Remaining Useful Life (RUL) integrating FMECA (Failure Modes, Effects and Criticality Analysis), life data analysis and data-driven & model-based methods. Starting from the FMECA analysis, the methodology proposes to identify the main critical components of new parts or systems, using life data analysis. A database collects and shares data directly from the field on similar systems and applications. Data are stored and managed via a web-based interface, the user may obtain them in real time as needed, when a modification in the robot or production cells occurs. Information are captured through a set of appropriate sensors, selected and located studying historical life data. From this dataset, RUL of components may be estimated using data-driven methods and model-based approaches. Then, the RUL results are shared with ERP systems to optimize production resources and maintenance activities and with FMECA again, to improve new projects in a closed loop. A preliminary application of the methodology is proposed on an anthropomorphic robot integrated in a production cell. This research is a part of PROGRAMs: PROGnostics based Reliability Analysis for Maintenance Scheduling, H2020-FOF-09-2017-767287.