The PI System fuels innovation and collaboration for Covestro’s global teams
- ChallengeCovestro’s data was isolated in local servers and difficult to access.
- SolutionMove data from local PI System™ servers to a standardized central architecture based on Asset Framework.
- BenefitsEasy access to data now allows for enterprise-wide collaboration, dynamic visualization, and advanced analysis.
Covestro realized that just because its engineers and researchers are spread throughout the world does not mean its data has to be. Before implementing the PI System, the company’s data was stored in local silos inaccessible to teams in other offices. When you have more than 16,000 employees spread across 30 global offices, that’s a problem. The company realized that, as one of the world’s largest chemical and polymer producers, it needed to do better. The PI System made Covestro’s data easily accessible across the organization, fueling collaboration and cross-site innovation on data visualization, machine learning, and more.
Breaking down silos with naming conventions
A multidisciplinary team from across Covestro led the company’s ProDAVis (Process Data Analysis & Visualization). The first step toward an integrated data environment in the PI System was standardizing tag-naming conventions. Once the tags were standardized, Covestro could unlock the full power of Asset Framework to contextualize its data and create a virtual model of the company’s assets.
For a large, complex organization like Covestro, having a standard naming convention is a huge first step. With that infrastructure in place, anyone in the company can monitor performance and process flow on any piece of equipment on any plant floor in real time, eliminating the silos that previously bogged things down.
Data without sensors
Sensors are always the best way to collect data from the field, but they’re also expensive and, as Covestro found, not always necessary in every location. The ProDAVis team used the PI System to simulate processes and detect anomalies based on context-rich digital modeling. These simulations accounted for missing data points and eliminated the need for expensive sensors in some situations. Every sensor saved represents funds that the company can use to advance other business goals.
“Imagine the potential savings when considering that an installed field device measurement can cost several thousand euros,” said Esteban Arroyo, a global data analytics expert at Covestro. “This integration allows us to detect the deviation between measured and simulated variables, which may indicate process anomalies.”
In addition to discovering process anomalies, Covestro uses PI Notifications to alert the ProDAVis team when something goes awry. Engineers also monitor PI Vision dashboards that seamlessly integrate with tools like Seeq and TrendMiner, available on a self-serve basis. Putting rich data in the hands of the people who need it, no matter where in the world they are, opens up new pathways for innovation and collaboration and creates opportunities for pattern recognition that did not previously exist.
Machine learning in the data cockpit
The PI System helped Covestro streamline every layer of its data pyramid, even advanced analytics tasks like scenario testing for root cause analysis and forecasting to enable predictive maintenance. Engineers can deploy machine learning and artificial intelligence tools from within the PI System. One such tool allowed engineers to model the effects of different conditions on the heat-transfer coefficient of a heat exchanger in order to decide on the best course of action. “You specify the condition, you see the effect on the target,” Arroyo said.
Additionally, a clear, simple central navigation cockpit in PI Vision allows Covestro’s team to access system data in different ways depending on their needs. Users can choose Asset View, which allows quick access to specific data and KPIs by types of equipment, or Hierarchy View, which organizes asset data by plant and location. Either way, the data is always available in the same reliable place.
“Before ProDAVis, our experts had virtually hundreds of island solutions for data analysis,” said Arroyo. “Our solution was to provide them with a set of seamlessly integrated tools, both self-service and machine learning, where they can learn models, analyze process behaviors, and share their insights with other experts.”
For more information about Covestro and the PI System, watch the full presentation here.