Articles

Heavy-duty powertrain systems resource multiple sensors collecting streams of data during production. Sensor data streams, originating from various production machines, simultaneously encounter the challenging situation of limited resilience to sensor quality issues in terms of sensors being non-linear, outdated, spawning false alarms, or going offline. Alleviating this critical situation requires a reinforcement technology to assure resilience, with both verified lineage-based approaches and a observe-and-repair mechanism of limited quality. Moreover, legislation advances aimed to guarantee higher future operational safety put additional pressure on the overall manufacturing sensor monitoring frame. For heavy-duty powertrain systems harnessed for combustion engines, an integrated technology based on automatic data lineage and quality governance detection of production machine sensor data streams resource the frame using a unified provenance model and ontology supporting the explicit definition of sensor quality rules. Provenance models automatic integration form the basis that enables the monitoring of multiple machine resources. Used quality rules enable the detection of bad-quality sensor segments while, in parallel, additional quality rules based on observed data control supported by human intervention capabilities repair gaps during production. The proposed automated technology provides an additional layer of resilience that improves the reliability of sensor data stream quality and contributes to the adherence of operational safety future rules. The implementation case study focuses on sensors deployed in powertrain systems production that modern industries must follow to ensure next-generation production safety levels.