Somewhere, you have physical schematic drawings that show all of the parts of your mechanical systems and overall asset environment. If you’re organized, you know right where these drawings are—good for you. If you’re lucky, they’re reasonably up-to-date. They might even be in an electronic format, like PDF. But, on their own, they’re insufficient as a basis for modernizing and streamlining the ways you monitor your assets.
That’s because machine learning operates on data, not analog drawing components. Even a PDF doesn’t cut it: It’s electronic, but it’s still just basically a photograph of the physical drawings, not a set of data. Meaningful analysis can’t be performed unless all of the descriptions in your schematic drawings are translated into ones and zeroes.
When we develop a digital twin for our customers, we take the details of your schematic—along with many other details we add in the course of learning about how your system elements rely on and impact one another—and we put it into a database. The thousands of rows and columns in the database model your installation as a sum of its many components, connections, and characteristics. Now, instead of relying on and interpreting a human-readable set of drawings, you have a dynamic digital entity with the power to support complex computations and analysis.
Creating a data base replication of your physical environment—a digital twin—enables you to:
At Tignis, we use physics-based machine learning algorithms along with the digital twin to yield concise analysis, tailored to your specific environment’s needs. Without a digital twin, you have no electronic medium for accurately capturing the myriad ways each system part relates to and impacts the other parts so that you can develop meaningful knowledge about them.