Digital Twin - Bits Mimic Atoms for Better Plants (Revised)

Digital Twin - Bits Mimic Atoms for Better Plants (Revised)

Using a Digital Twin enables process optimization, better control room operator and field operator learning and retention for greater safety, reducing energy consumption, better plant and skid design, etc. The term Digital Twin is new, but the underlying concept is not new. Just imagine what this can do for operational excellence in your plant. So what really is a Digital Twin, and how exactly does it help the plant? Why all the excitement now? Here are my personal thoughts:

Digital Twin is the digital model of a physical thing (e.g. a whole plant, pump, or reservoir). I see two parts to a Digital Twin:

  1. Form model
  2. Functional model

All of the above are not new. They have existed individually for decades. The most interesting fact is that they are now combined and used together, and the synthesis brings new exciting results. This is what’s new about digital twin.

Structural Model

2D drawings and 3D CAD structural models of shape and form of plants and equipment have been used for engineering for decades to detect pipe clashes etc. but are now being used in new ways. Reservoir model is another example of a digital model representing the shape of an asset.

Field Operator Learning

Industrial Virtual Reality (VR) learning is putting 3D CAD structural models into new use for digitization of learning to ensure competency. The 3D CAD structural model is imported into the 3D VR environment and enhanced with textures and actual plant equipment shapes rather than symbolic representations. Manual procedure steps are also engineered into the VR system. Thanks to the immense computing power of gaming graphic processing cards, the plant environment is rendered with a high degree of realism including depth perception, shadows, and with a first-person perspective changing as you turn your head or squat down. A very immersive experience. A student puts on the VR goggles and it is like walking into the real plant as it has been engineered to be a digital replica of the plant. You move up and down stairs and monkey ladders. You can interact with equipment, like starting and stopping pumps, opening and closing valves. The student can now practice manual tasks and various scenarios in the virtual environment until they get it right every time. It is like a modern video game. Some plants are already using VR for learning.

The VR environment rendered from the 3D CAD structural model can be integrated with the process simulation model. The combination of the form (plant design) model and functional (process) model together provides a highly realistic learning experience where actions like starting and stopping pumps or opening and closing valves changes the flows and pressures etc. just like in a real plant. If your plant is old, designed without 3D CAD, the virtual environment for VR learning can be reverse engineered. The VR environment can even be integrated with control room Operator Training Simulators (OTS) such that a combined team of field operators and control room operators can work together just like they do in real life.

Design Verification

VR is also used for digitalization of plant and skid design allowing the end-user to walk into and experience the plant or skid before it has been built. This helps in ensuring the usability of the plant to make sure it is not too cramped to move around or to remove equipment like valves for overhaul etc.

Simulate twice, weld once.

Functional Models

Functional models like dynamic process simulation which digitally mimics behavior has been around for decades for process optimization and control room operator learning but are now being used in new ways. Dynamic process simulation is now also used in combination with VR learning to increase the realism of field operator learning with process changes in response to actions taken.

Process Optimization

Dynamic process simulation has been used for decades to test out different setpoints for pressures and temperatures as well as control strategies to optimize the process to achieve one objective or another, such as minimal energy use, or producing more of a particular grade etc.

Control Room Operator Learning

Dynamic process simulation is also used for control room Operator Training Simulators (OTS) for years, to practice certain control room operator tasks like startup, shutdown, or plant upset etc. for competency. The teacher sets up the scenario and monitors to see if operator response correctly and is able to handle the situation. It is very much like a flight simulator.

A new exciting development is that control room operator learning is now combined with field operator learning by integrating the OTS and VR systems such that control room student and field operator student can work together on tasks where such collaboration is required.

Bits vs. Atoms

The future is digital. A digital twin is about digital (bits) mimicking the physical (atoms) world. The digital model concept of digital twin is not new, but the pieces of information are now being used together in new ways. Use of digital twin in a process plant can improve field and control room operator performance for greater safety and reduce energy cost.  You don’t have to deploy all these digital twin solutions at the same time or for the entire plant. You can deploy them one by one, starting in one area of the plant. Well, that’s my personal opinion. If you are interested in digital transformation in the process industries click “Follow” by my photo to not miss future updates. Click “Like” if you found this useful to you and Share it with others if you think it would be useful to them.

Ganapathy Subramaniam

Vice President - Engg services(Control system, Digitalisation, Performance Monitoring solutions)

3y

1) Digital twin(with VR) for operator training is an extension of the OTS which has been in the industry for quite some time.. The need and necessity for this extension has to be decided by the plant owners based on the budget they have. 2) Digital twin for predictive maintenance/failure of equipment is understandably not required as we have lot of energy management tools in the market to do that 3) But digital twin for process performance monitoring( for example boiler) which has potential for APC(ex :combustion control) and performance(heat exchanger performance) will be of interest I guess. Although OTS for a boiler simulates all these performance degradation of boilers over a period of time is not covered in OTS. This is where I believe a dynamic Digital twin might be of help. But the efforts involved in developing this kind of digital twin is enormous and it has to be maintained by skilled personnel to keep it working over a period of time. This is where the gap comes in for a Digital twin according to my experience.

Like
Reply

I have made some revisions to this essay removing some of the original text. A digital twin looks like and behaves like its physical twin; that is, it is a custom model for simulation; a 3D structural model for the plant look, and a process model for the process behavior. There are many tasks that do not require a digital twin to be digitally transformed, and many software applications used for digital transformation are not a digital twin. The industry went overboard in ‘twin washing’, including this essay Predicting equipment failure, root cause analysis, and many other functions do not require a digital twin. Apps predicting pump failure and diagnosing the root cause, computing efficiency, or for energy management has an internal model, but it is not simulation so it is not a digital twin A digital twin is a model for simulation. You don’t need to simulate to predict. That is, you don’t need a digital twin to predict. There is no need for a digital twin for predictive analytics or efficiency monitoring on a compressor, turbine, or pump etc. or for energy management. A digital twin has high upfront cost and high maintenance cost, so it should only be developed for the use cases which really needs it: OTS and VR A digital twin is modeling and simulation; 3D structural model or process model. Thus the digital twin does not need live data Storing documents and a history database are also not digital twins

Willian S.

Project manager, industrial facilities, automation, power energy.

6y

Great mimic to explain digital twin

Eduardo Paciencia Godoy

Professor and Researcher at UNESP - São Paulo State University

6y

Raul Katayama Mansano veja esse assunto

Like
Reply
Andrey Murashko

Sr. Manager, IT/OT Architecture

6y

Jonas, to my opinion, Form Model is secondary. Civil building industry has come along with BIM concept which in effect is a data model (an information model of a building), and 3D shape is derived from that model as a subset of data. When we want to simulate a certain aspect of a process plant, we need that data model in order to link variables (live data stream) to respective plant's elements. 

To view or add a comment, sign in

Insights from the community

Others also viewed

Explore topics