Efficient Life Cycle Engineering relies on credible insight into how assets degrade in operation. Predictive maintenance has been identified as a key enabler, but its data management is often fragmented across heterogeneous, proprietary systems and does not readily feed back into design, maintenance planning, or end-of-life decisions. This paper presents and experimentally verifies an Industry 4.0-based approach that embeds predictive maintenance into a standardized digital-twin architecture, centered on the Asset Administration Shell, to close this gap. Firstly, to standardize data and enable intelligent processing, we upgrade our existing Industry 4.0 Component Stack to predict and manage key condition indicators, such as remaining useful life, together with event logic, and to employ our proposed Digital Product Passport that records thresholds, events, and maintenance outcomes in interoperable submodels. Secondly, to support context-aware decision making, we show how the upgraded stack maps event codes derived from cross-sensor telemetry to decision guards that trigger service tasks such as operation shaping or unplanned maintenance and is designed for continuous review and learning throughout the life cycle. The approach is implemented and verified on an unmanned aircraft system used as a surrogate asset in a maintenance, repair, and overhaul data environment, demonstrating that assets can locally interpret their operating context, autonomously request appropriate interventions, and expose the resulting evidence through standardized interfaces. For life cycle engineers, the approach transforms degradation behavior, interventions, and operation outcomes from isolated runtime logs into traceable, machine-readable information that directly supports design choices, maintenance strategies, and end-of-life planning.
Weiß et al. (Thu,) studied this question.