Digital Twin Strategies: Manufacturing tips for successful implementation

Digital Twin Strategies: Manufacturing tips for successful implementation
Digital Twin Strategies: Manufacturing tips for successful implementation
According to Gartner, 75% of organizations implementing IoT already use Digital Twins or plan to within a year as it is emerging as an important part of a predictive analytics strategy. However, most organizations haven’t made it past their pilot stage and definitely aren’t using it at scale. So, why aren’t organizations grasping at the opportunities the Digital Twin provides for manufacturers? As Thomas Edison once said, “opportunity is missed by most people because it’s dressed in overalls and looks like work.”
 
To make it more appealing, vendors are offering up existing solutions to support the Digital Twin that are “easy” and “simple to use.” For example, vendors provide models from CAD or simulation, and represent them as an exact representation of the product or asset that was made or operates in the field – even though much has changed since these were created.  
 
The definition of Digital Twin has been well vetted over the last few years, yet it seems when it’s time to pilot projects, it is forgotten. Organizations need to strive for the exact, individual configuration of the product or asset, yet the solutions they are implementing do no such thing. By working out these kinks, everything else―including performance monitoring, IIoT, predictive analytics and operational simulation―will fall into place, creating tremendous value for the organization.
 

Why Organizations Have Not Successfully Implemented the Digital Twin

 
Simply put, a model―CAD, simulation, engineering bill of materials (eBOM) or manufacturing bill of materials (MBOM)―is not a Digital Twin. A Digital Twin is first created when the manufacturing of the product is complete, which is indicated when serialized components are recorded. All other information created before that point is related to the engineering parts and is considered Digital Thread data that can be accessed by the individual Digital Twin’s configurations.
 
When it comes to implementation, organizations fail to focus on the technologies that are required to know exactly what is needed for collecting related operational and performance data. They also miss out on viewing what the asset actually looks like at any given point in time. Tangible business value is created at this point.
 

Why Predictive Analytics and IIoT Efforts are Doomed to Fail

 
If organizations don’t have context within their Digital Twin configuration, when it comes time to implement technologies, such as IIoT and predictive analytics, they will only be generating clues, not accurate results. For example, organizations can collect all the operational and performance data about an asset that they want, but as soon as an asset is removed from the system, the operational history related to that asset continues with it. If the data follows the system itself, it will completely change the context of the system monitored. 

For instance, when looking at the operational performance of an aircraft, the initial information when beginning monitoring is absolute―especially when there is ample context to begin with. But, when parts are replaced with new parts, this changes the operating life parameters. As a result, unless you continually update the context of the aircraft, algorithms are thrown off resulting in immense cost. Results may point to maintenance requirements, when in reality, no intervention is required—wasting labor hours and system productivity.   

For predictive analytics or IIoT to be effective, the context of the asset and system is required. Organizations need to have the technologies and processes in place to replace the operational data from the old asset with the operational data from the new asset in order to get an accurate representation of what is being monitored. Organizations should be taking the physical configuration of that asset into consideration as part of their predictive analytics algorithms to get the right answer, not just estimate an aggregate average.

The Secret to Making the Digital Twin a Part of a Manufacturer’s Technology Strategy

 
Manufacturers need to change their philosophy and build the foundation of each asset and system they want to manage. Due to the many uses for Digital Twin across the product lifecycle, they also need a flexible/dynamic data model ingrained in the technology they choose. Only a dynamic data model will support the evolving needs and various configurations of Digital Twins, based on the use case. 

Organizations should start any Digital Twin strategy by capturing and managing the actual physical configuration of the asset. That means they need to know the serial numbers of components that were assembled into this end product, how many bolts were used to attach a component, what torque value was applied to those bolts and what versions of software are installed. All these things, and much more, should come together to form the digital representation of the physical configuration of the asset―a process lacking in the industry today. 

Organizations need to know what assets they have within the systems they are operating and maintaining. It doesn't do any good to know that there is an excess temperature warning, if we don’t really know what the asset is or how it will react. The model number of the asset, the serial number, the maintenance history and the operating history of that asset should be known before determining if that “excess” temperature reading is really “excess” at all.   

Conclusion – Where to Start and What Value Should I Expect?

 
Where to start largely depends on where the user is in the lifecycle. If they are a maintenance shop, they should start by keeping accurate records of inspections, tracking what the assets look like when they come in and making sure data is digitally recorded, not on paper. Specifically, data should live in databases that can be accessed via Open APIs, are searchable, inspectable and supports a dynamic data model to build the context for any Digital Twin use case that’s to follow.
 
For manufacturers, users should make sure their “as-built” configuration is as accurate and as detailed as possible. Then, they can start extending that “as-built” configuration out into the test shop, installation and delivery, commissioning and operation divisions. Lastly, an asset’s journey across the lifecycle will need a solution to capture and manage changes as things happen to it, such as replacement parts, maintenance and upgrades.
 

About The Author


Jason Kasper, joined Aras Corporation in April 2017 and is a Product Marketing Manager with his primary focus being maintenance, repair, overhaul (MRO), manufacturing execution systems and their importance within the product lifecycle. Jason has over 20 years of experience in working with customers to develop enterprise software solutions. www.aras.com
 

Click Here for More Information

Did you enjoy this great article?

Check out our free e-newsletters to read more great articles..

Subscribe