Covestro is one of the world’s top 20 global chemical manufacturers, with production sites in Europe, Asia, and the United States. Sensor Health and Data Quality have been areas of focus for Covestro for several years. In order to identify potential problems in this area, Covestro wrote their own PI ACE application called Redundant Sensor Diagnostics (RSD). While this internally-developed diagnostics solution provided some interesting insights into the state of the data and the health of the sensors, it fell short of being able to identify failing sensors, especially in the area of redundant sensors. After several high-profile and costly failures, Covestro made it a priority to find a partner that could leverage the data stored in PI historians with a solution specifically addressing sensor health. The goal of the initiative was to minimize operational risks and reduce the extensive financial impact of those events by focusing on preventing unplanned outages and lowering maintenance costs.
Covestro has approximately 1.6 million relevant tags across 30 sites. In the last several years, Covestro has focused on building an extensive PI-AF Asset Class Template Library that is available globally. In this session, we will delve into the shortcomings of the home-grown solution and demonstrate the benefits of the off-the-shelf application we are using today.
By leveraging sensor PI System data, and in combination with AVEVA Predictive Analytics, Covestro’s goal is to reduce unscheduled downtime and lower maintenance repair costs (including labor & materials) by identifying proactively and systematically sensor and asset health issues.