Closed-Loop Digital Twin, the Learning Never Stops

By design, a digital twin should continuously learn and improve through a constant feedback loop of real-time information. When utilized in conjunction with the Internet of Things (IoT) or Industrial IoT, a digital twin has the steady supply of information it needs to keep on learning and getting better over time.  This is crucial because nothing is static when it comes to manufacturing operations. Unfortunately, too many companies fail to keep their digital twin–which is supposed to be an exact representation of real operations—current.

Let’s look at a few key steps manufacturers can take to ensure that their digital twins are an accurate portrayal of their physical counterparts.

Empowering a digital twin to continue learning starts with connecting it to a constant stream of data. Fortunately, today’s manufacturing systems offer significant opportunities to capture data. From the strategic placement of IoT sensors on legacy equipment to the array of data available from programmable logic controllers (PLCs), manufacturers can collect whatever data they desire. And, the amount of data only continues to grow. In fact, IDC predicts that the collective sum of the world’s data will grow from the 33 zettabytes (ZB) collected in 2018 to 175ZB by 2025, for a compounded annual growth rate of 61 percent.1

However, simply having access to so much data can actually become a problem, especially if manufacturers capture the wrong data or don’t know what to do with it once they have it. And unfortunately, the complexity associated with mountains of data can quickly lead a manufacturer down the wrong path. The key to success rests with focusing on the data that provides the most meaningful insights about production, performance and the product.

Ultimately, the digital twin should provide research & development, engineering, manufacturing and all other stakeholders with a clear understanding of what needs to change to improve. If the data the manufacturer is collecting does not provide these insights, it might be time to shift focus and keep digging.

For manufacturers to compete in industry 4.0, they need to get new products as well as existing product iterations out the door before competitors capture the market. Access to data empowers the digital twin to evolve and continuously update to reflect any changes to the physical counterpart throughout the product lifecycle. Creating a closed-loop of feedback in a virtual environment enables companies to continuously optimize their products, production and performance.

Within a closed feedback loop, manufacturers can augment data with process-based information from the entire array of systems that play a crucial role within today’s enterprise such as manufacturing execution systems, enterprise resource planning (ERP) systems, computer-assisted design (CAD) models, customer feedback, integrated product sensors and complex supply chain systems.

The best way to close the feedback loop is to establish a digital thread. A digital thread provides an understanding of why various aspects exist, including a structure for looking at a virtual product or simulation. It allows teams to answer questions such as:

• How do changes impact product performance?
• Why is the production team inspecting a specific component?
• What is that inspection telling them?

Being able to quickly answer these questions empowers the type of fast iteration discussions that enable manufacturers to meet evolving market demands.

A digital thread also automates verification management. For instance, during the design process, if some changes are not communicated down, teams may be working with information that is no longer accurate. This is all too common in complete manual environments, but it should never be a scenario when the manufacturer is leveraging a digital twin.

A key part of the feedback loop is to ensure that those data streams are then securely delivered for aggregation and ingestion into a modern data repository, followed by processing and preparation for analytics. Today’s entire array of data processing tools and capabilities, including edge computing and artificial intelligence, can play a meaningful role in the analysis process. Manufacturers can then apply these insights back to the production environment through additive manufacturing, robotics, or other tools utilizing decoders and actuators.

The best way for a manufacturer to ensure that it is leveraging adequate and appropriate data (in real-time) is to regularly monitor the digital twin’s effectiveness. According to a recent Dataversity article, “The more you quantify the quality of how identical the digital representation is to the physical counterpart, the less risk when making physical alterations. As is always the case with monitoring, high accuracy and fidelity are how admins ensure their changes are beneficial and successful.”2

In many instances, manufacturers can address any inadequacies by revisiting the data collection component and refining which data streams it applies to the digital twin.

Bottom line: digital twins represent a significant opportunity for manufacturers to continually leverage data to improve their processes. However, this only works when following the steps to keep digital twins up to date with the physical environments they represent.