Some elements of what we now call digital twins reach back into the early days of industrial computing. Data historians, PLC’s and instrumented processes have been a part of large industrial systems for some time. Process manufacturers in energy, resources, chemicals, food, and pharmaceutical markets have instrumented the environment for safety, efficiency, and regulatory compliance for as long as we have used computers in industry. This data provided some level of benefit in terms of meeting compliance requirements and improving maintenance and uptime through alerts and analytics, but it was also difficult to work with and expensive to deploy and manage.
Contemporary digital twins are an entirely different animal. While the PLC driven control systems work well in the refinery or plant, they don’t typically scale beyond those environments into the supply chain. In the supply chain, things move. They change hands. They are delayed or diverted. They may not ship at all. Almost all outcomes in the future are probabilistic and carry some level of performance risk. The term On Time In Full (OTIF) refers to the delivery of an order on time, including the full order quantity and this number is invariably less than 100% – sometimes a lot less. As companies embraced lean and sourced more and made less the challenges with gaining insight into actual supply chain performance became both more difficult and more important. During the current pandemic we saw many companies disrupted by performance failures in tier 2 and tier 3 suppliers that they had little to no visibility to. Lean, just in time supply chains became fragile, breaking quickly as these suppliers failed to react quickly to pandemic driven disruptions in supply, manufacturing, and logistics.
In an effort to better understand supply chain performance and processes, the market is looking to advanced technologies to continue generating improvements in performance. Today the focus is on IoT, 5G, and AI as the best ways to rapidly accelerate the improvement of data driven planning and execution in the supply chain. This trinity of technologies promises to dramatically reduce the costs and improve the availability of real time data in the supply chain and help companies digitally transform into more demand centric organizations. Becoming demand centric or demand driven is not a new concept. The concept of the Demand Driven Supply Network was developed by AMR research in the early 2000’s and became the basis of their Supply Chain Top 25 company rankings. Even with Gartner’s acquisition of AMR this program continues to this day and after 15 years we are still chasing what it means to be truly demand driven. This idea resurfaces, and gains traction among supply chain owners every couple of years, but we continue to struggle with the technology and data requirements to turn this vision into reality.
COVID has increased the focus on using data to create resilience in supply chains. This requires better data on demand and supply. Demand forecasting has also been subject to rapid change and multiple failures as a result of the pandemic. Rapid changes in customer demand, coupled with multiple failures in the supply chain have shown just how brittle our old models based on cost optimization can be. This is especially true in complex supply chain scenarios where we are attempting to create large multimodal models that replicate the physical world our resources, components and finished goods move through.
Supply chain digital twins create a unique opportunity to create a more intelligent way to model and plan for a variety of rapidly changing factors facing organizations trying to improve resiliency. Gaining a better understanding of multiple tiers in the supply chain (geo location, logistics, capacity, technology) help us better understand the resilience of our supply network. One of the most important advances brought by the digital twin over legacy approaches is the ability to add context to data to improve insight.
Understanding the characteristics, capacity, location and performance of tier 2 and tier 3 suppliers can bring a much richer and more accurate view of possible issues impacting the resilience of our supply chains. The same is true in terms of adding detail to customer profiles in terms of understanding changes in the demand side of the equation.
These models of the supply chain can include real-time data on many factors that affect supply chain performance. Models can consider weather, transportation delays, cold chain excursions, natural disasters, commodity shortages, currency fluctuations and hundreds of other real time and contextual data elements that impact performance. Building resiliency and agility requires a more nuanced and detailed set of data than tradition optimization engines provide, and supply chain leaders will look to digital twins to augment their traditional demand forecasting and supply chain planning applications.
Comments