Digital twins are highly effective tools for optimizing manufacturing processes without wasting time and resources. They can drive a lot of value for manufacturers across sectors, from predicting quality in real time to removing the necessity for costly physical tests.
But first things first, what is a digital twin?
In the vaguest sense, anything can be a digital twin. You can build a basic digital twin in Excel by taking some data and making a chart and saying it represents production. That's why you'll see lots of companies claiming to offer "digital twins"—and even jobs for "digital twin engineers"—but they don't all offer the same services or value.
Adding to the confusion, the phrase has come to mean different things in the process industry and in discrete manufacturing. In discrete manufacturing, it often refers to a CAD model of an individual product. In process automation, on the other hand, it's frequently used to describe the methodology where you build a virtual copy of a process (not a product), which can then be used to run the factory better.
For all these reasons, we specifically avoid buzzwords like "digital twin," because they tend to create confusion without providing value. But the above definition—building a virtual copy of a process to make your factory more efficient—is most closely aligned with what we do at Fero.
3 ways to drive value with digital twins
Now that we've got a standardized definition, let's look at three key ways that you can use a digital twin to drive value in your factory:
1. Predict quality issues in real time
The value of this is huge. If you can predict quality issues, you can prevent them and improve the quality of your finished product, in addition to saving costs and reducing waste. One Fero customer found that they were able to skip 98% of end-of-line product tests by using a digital twin to predict the quality of the final product in real time.
Most industrial ML software providers offer this functionality, but the value you'll see is determined by the accuracy of the predictions. A line running through a few datapoints can be used for prediction. But is it a good prediction? Does it provide confidence bands and other measures of statistical confidence? Does it take into account relationships between hundreds of other variables? Do you trust it enough to make key decisions based on it? A truly valuable digital twin helps you not only predict quality issues in a vacuum, but understand the root causes of those issues so you can optimize your process.
2. Skip plant tests by simulating hypotheticals
Delaying your regular production to run plant tests is both costly and time-consuming. With a digital twin, you can simulate hypothetical situations without the need for plant shutdowns. For example, you can come up with alternative recipe and process formulations, or figure out how to change set points to account for ambient weather conditions or certain devices being taken out of service.
In addition to saving costs, this helps manufacturers make progress on their sustainability goals by eliminating the energy use and emissions from regular testing. When you run a plant test, the entire plant continues to emit greenhouse gases, without a finished product to show for it. The learnings that result from such tests can be replicated within a digital twin, without replicating the emissions.
3. Learn the most effective process improvements for your factory
In the above examples, a digital twin provides immediate ROI. But it can also give engineers a long-term tool to explore the production process in real time, so they can make ongoing process improvements using their own domain knowledge.
On the factory floor, it can be difficult to see what's going on. With a digital twin, engineers can easily examine all the factors affecting production. Fero's digital twin tracks relationships between hundreds of parameters. When Fero is used as a digital twin, you can get even more benefit from Fero's explainable machine learning engine. Engineers can learn about their factory and improve their process—for example, understanding what the optimal temperature is for a certain step, or what complex interrelated factors cause a particular quality issue.
Overall, a digital twin can drive a lot of value, whether you're looking to improve cost and production efficiencies, push forward emissions reduction goals or all of the above. Just be sure you know what you're actually signing up for.