Digital twin is the phrase used to describe a computerized (or digital) version of a physical asset or process. The digital twin contains a sensor or sensors that collects data to feed the asset or process model. In short, the digital twin concept combines the ideas of modeling and the Internet of Things (IoT).
The digital twin concept has most often been applied to assets. A piece of machinery generates data on vibrations, heat, pressure, and other things as well. That data is used to predict a machines failure and to apply preventative maintenance to make sure that an unplanned failure does not occur.
In terms of supply chain planning, factory machinery is the key area where an asset’s failure leads to increased manufacturing costs and service failures for customers. The supply chain model does not look to predict asset failures, but it does seek to use the digital twin maintenance model’s inputs to improve factory scheduling. The asset model relies on machine learning to improve the forecasting of machine downtime over time.
AspenTech is an example of one supplier of supply chain planning solutions that seeks to use these inputs. AspenTech also sells asset maintenance solutions. Aspen Mtell is a low-touch machine learning solution they say can accurately forecast a hyper compressor failure in a low-density polyethylene (LDPE) process. In many industries, scheduling around machinery maintenance would not be all that difficult. You have the maintenance crew fix the machine at night, and the schedule proceeds as planned.
The chemicals industry is different. It is hard. A small number of raw materials can be transformed into hundreds of thousands of final products. The manufacturer does not just produce products, but also coproducts and byproducts. These byproducts can be sold to other companies or used internally in the production of other final products. Optimal production proceeds by monitoring not just the physical process, but the chemical properties of the materials being produced. There are production wheels that contain rules about the sequences in which chemical grades can be produced and the constraints that must be respected. There are expensive, heavy and complex manufacturing assets that can cover the full spectrum of production operations: continuous, semi-continous or batch. Shutting down and then restarting the process is expensive, time consuming (think days, not hours), and has environmental, health, and safety implications.
A hyper compressor’s job is to build up pressure that is needed in the conversion process. Compressors may be called upon to apply up to 50,000 pounds of pressure per square inch to the process. That puts a lot of strain on the machinery. These compressors typically go down many times a year. The ability to mitigate this problem is worth millions of dollars to chemical companies.
As one example, Aspen Mtell provides more than 25 days of advance warning of a central valve failure. For example, on January 3rd, Mtell might tell a planner that an asset failure is likely on or shortly before January 19th. This can allow for scheduling less expensive maintenance downtime rather than reacting to unplanned downtime. Aspen Plant Scheduler can then be used to schedule the planned downtime options. The schedule optimizer can trade off customer commitments, inventory holding costs and manufacturing costs. A chemicals manufacturer will have more options if they have more than one reactor that can provide the desired chemical grades. Nevertheless, there will be different costs associated with the production options. In short, detailed scheduling is a complex optimization problem involving demand priorities, manufacturing economics, and production sequencing constraints.
The AspenTech offering is the first example of a solution that includes optimized production scheduling based on an integrated digital twin maintenance model that I have seen. In many industries, this solution would be overkill. Not here. Other asset-intensive industries, like power and metals & mining could similarly obtain significant value from optimizing maintenance across the supply chain. Critical, costly equipment is key to environmental, health and safety in these industries. And in heavy process industries with complex turnarounds, these solutions can save companies millions of dollars a year.