Asset management leveraging technology is generally associated with leveraging data generated by sensors on machinery to manage the health of the equipment. This capability of leveraging the data captured from the floor to replicate asset conditions and processes drive manufacturing digital twin offerings like  SAP IoT, based on the SAP BTP platform. Hyperscalers like Microsoft and AWS also have their offerings, branded as Azure Digital Twins and AWS IoT TwinMaker. While we generally assign the term digital twins to these platforms, a comprehensive digital twin platform can be further broken down into a combination of physical digital twins , process digital twins and labor movement twins. But another way to think about asset management is to incorporate inventory tracking in this bucket as well. A seamless tracking of inventory across your warehouse can help you create a digital twin of your warehouse as well. But this digital twin will not be result of just tracking what product sits where in the warehouse. Process data, combined with human and equipment (like forklifts and bots) also need to contribute data. The fact is that process intelligence, when combined with other approaches, forms the foundation for realistic digital transformation. While this article focuses on how process intelligence can help build digital twins, which when combined with physical twins help build true digital twin capabilities, process intelligence solutions are strategic solutions that when used creatively in carefully designed architectures, can transform businesses. That is why it is one of the foundational themes of our upcoming, February 2023 research report, Building Resilient and Agile Supply Chains Leveraging Data, Analytics and Automation. In this article, we will explore what these two terms, physical and process digital twins mean and why there is a need to integrate these across the supply chain.

What is a Process Digital Twin?

As indicated above, physical-digital twins are created based on data from a physical object, like equipment. In this case, the data is from machine operations (like temperature, vibrations, RPM, etc.), whereas process digital twins is the virtual replication of your business processes.  This is where process automation and intelligence capabilities come in—using these technologies. A process digital twin is built using process mining and process discovery to leverage data generated through transactions in your systems. Process twins can identify process bottlenecks, re-engineer opportunities and create value through efficiency and productivity. But it is when they are integrated with physical twins that they help provide the complete picture.

Why End-to-End Integration is Imperative?

If you analyze the example architectures of the three solutions highlighted above, you will find that these platforms create the digital twin by combining physical and process digital twins. This will help enable the true "smart" factories and warehouses. SAP IoT Example Architecture from Intel https://www.intel.com/content/dam/www/public/us/en/documents/reference-architectures/sap-iot-reference-architecture.pdf Azure Digital Twin Platform Example Architecture https://docs.microsoft.com/en-us/azure/digital-twins/overview AWS Digital Twin Platform Example Architecture https://aws.amazon.com/iot-twinmaker/ A significant portion of that "smartness" comes from allowing the digital twin to do automated planning and adjustments. Merely highlighting anomalies has been something that has been in place in manufacturing for decades now. It is the ability to make adjustments to the process, in real-time, based on data from physical and process digital twins, that will eventually constitute a smart factory. An Illustrative Example Let us walk through a quick example to understand why this integration is imperative. Before that, we must bifurcate a smart factory's " smartness " into two distinct buckets. One can be tactical planning and is the bucket where the system is smart enough to take steps, like raising a flag or halting production when it detects an anomaly. The other bucket is "operational" smartness, where the twin can decide on operational planning aspects and feed inputs to the tactical planning algorithm when applicable. Without getting into the technical jargon, let us visualize what can be considered a basic operational fundamental "smartness" in your smart factory (beyond identifying anomalies). Say you use the same manufacturing line to manufacture three different products. Two of these three products, A and B, use the same raw material, X, as an ingredient. A large order from a very important customer for product A, came in late but had a close delivery date. Leveraging data from a process digital twin (order management in this case), the smart factory can prioritize the manufacturing plan in the short term. As you are already thinking, these are not the only inputs going into manufacturing schedule planning. But this is a simple example to highlight how smart factories can effectively benefit from process digital twins data. Eventually, you can build one master algorithm and multiple sub-algorithms supported by data from your physical and process digital twins to build truly smart factories. The good news is that all the tools and technologies you need to build such capability already exist.