The Use of Digital Twins to Optimize Operations

I first came across the concept of a digital twin when I learned, many years ago, how to use CadSim software, but at that time it was referred to as a dynamic simulation model. So what is a digital twin? – is it just an updated name for this?

Apparently, the term was used first in 2002 by Michael Grieves, then a professor at the University of Michigan. The term has evolved over time, but it is clear there are three important components of a Digital Twin:

an accurate, detailed model of the process;
an evolving set of data relating to the process;
a means of dynamically updating or adjusting the model in accordance with the data.

In the last 10–15 years, advances in the use of data historians, wireless Internet of Things sensors, cloud computing, “big data”, and AI have made digital twins more feasible and valuable in industrial settings. Digital twins incorporating both physical models and machine-learning models have now become a valuable tool in industrial process simulation, enabling real-time monitoring, optimization, and prediction of complex systems. In pulp and paper, digital twins can help simulate what-if scenarios, troubleshoot problems, improve efficiency, reduce waste, and ensure consistent product quality. Sensor data and AI models can be used to predict equipment failure before it occurs. A digital twin is also a valuable operator training environment.

Some of the areas that can benefit from being included in a digital twin:

Predictive Maintenance: monitor assets such as digesters, refiners, and paper machines, and predict failures in pumps, valves, or rotating equipment to avoid unplanned downtime;
Energy and Resource Efficiency: model steam and electricity usage to find opportunities for reducing energy costs, and optimize water and chemical usage to meet sustainability goals;
Quality Control: predict paper characteristics (e.g., tensile strength, brightness, smoothness) based on process variables.

It should be stated, however, that the setup and use of digital twins has its challenges. There is a high initial investment of equipment and time, and a need for high-quality, real-time data as well as the building and validation of models based on various AI strategies. It requires expertise in both IT and pulp and paper process engineering, plus a commitment to proactively use and maintain the model (“use it or lose it”). Managers of facilities with a digital twin should make sure to track value created by it, to justify its continued use.

Recently, I had a ride in San Francisco with Waymo, an autonomous taxi service that is a subsidiary of Alphabet. It was an impressive experience. Waymo vehicles use a very detailed 3D model of the cities where they operate (currently San Francisco, Los Angeles, Austin and Phoenix). They are also fed by five lidar sensors, 29 cameras and six radars on every car to scan their environment. It took over ten years to develop this process of getting customers from A to B without a driver, and digital twins were invaluable in speeding up this development.

Digital twins can be a powerful way to navigate the complex challenges of modern manufacturing. As sensor technology, AI technology, and connectivity continue to evolve, the role of digital twins will becoming increasingly important in industrial innovation.

Martin Fairbank, Ph.D. Martin Fairbank has worked in the forest products industry for over 35 years, including many years for a pulp and paper producer and two years with Natural Resources Canada. With a Ph.D. in chemistry and experience in process improvement, product development, energy management and lean manufacturing, Martin currently works as an independent consultant, based in Montreal. He has also published Resolute Roots, a history of Resolute Forest Products and its predecessors over 200 years. This email address is being protected from spambots. You need JavaScript enabled to view it.