In this episode, Michael Clark, Data Analytics and Information System Lead at Petronas Canada, shares the strategic roadmap used to dramatically reduce unplanned shutdowns across the company's natural gas compressor stations. Faced with over 15,000 unit shutdown events in a single year and disconnected IBM Maximo work order data, the team built a scalable analytics pipeline. By extracting daily state data directly from their SCADA systems into Databricks and visualizing it in Power BI, Petronas empowered its field and reliability engineers to pinpoint specific "bad actor" stations. Clark discusses the importance of normalizing shutdown rates, resolving nuisance trips like hypersensitive level switches, and standardizing fault codes across their operations. The session also outlines the company's future goals of validating SCADA events against Maximo work orders and building a comprehensive asset registry for over 210,000 assets to further advance their predictive maintenance capabilities.
