Move fast and break (simulated) things

The following conversation was overheard among two control room operators as they ran through the start-up sequence at an LNG plant

“Hey, what happens if we leave this feed valve at full open?”

“I don’t know… probably overrun the flare system.”

“Let’s try it and find out!”

They did try it and they did find out. Yet, these are not famous last words from two reckless control room operators. The plant did not make the evening news, and the manager did not have to worry about how his emails would look in the newspaper.

This plant invested in a digital twin that included plant behaviour in the form of dynamic process models. The two operators were able to follow their idea and find out what would happen in a scenario that was unlikely but very possible. The events played out in a completely safe environment where the only consequence was greater knowledge.

A mantra adapted Move fast and break things. That tech industry slogan is generally credited to Facebook founder Mark Zuckerberg. Coders are encouraged to try new approaches as quickly as possible, even to the point where they might cause other problems. Freedom to fail is one of the reasons tech industry start-ups have been able to launch with stratospheric success, but it is not so easily adapted to the industrial world. You can be cavalier about failure when the worst consequence is fewer cat videos. Bad code can be rewritten for just the cost of the coder’s time. Steel is not so easily corrected, and poor decisions in industrial plants can put people in harm’s way.

Digital twins that include process behaviour and all 1D, 2D and 3D plant data can be used to create a sandbox. Engineers and operators can try new things and get a realistic idea of the consequences for both economics and operating performance. Most ideas will fail, but that failure will lead to maturity in both design and operations.

Simulation across the lifecycle To get the most value out of the digital twin, it is important to start early. Process simulations provide a way to evaluate alternative plant configurations long before designs become physical and change becomes expensive. Plants have run complete virtual commissioning exercises to uncover and mitigate thousands of punch-list items prior to real-world start-up. First feed enters a mature plant, so processes can ramp up to nameplate capacity quickly with no surprises.

The value of the digital twin does not end with a successful start-up. Operational excellence is a journey, not a destination. A fully realised digital twin helps plants capture organisational knowledge so that every operator can gain the experience of their most knowledgeable colleagues. Unusual or high-consequence real-life scenarios can be replayed in the simulator for better recognition, understanding and response.

Cloud brings digital twin within everyone’s reach

Plants in the energy industry require many different companies to design, build and operate. Their lifetimes are measured in decades. The digital twin can form the thread connecting these stakeholders over time, but only if accessible. Owner-operators should move to the cloud to turn engineering data into intelligent, actionable insights. Digital twins in the cloud create business resilience and allow project managers to match deployment and licensing to project needs.

Simulated problems, real improvements

Let us return to that LNG plant and our two operators. Before the simulation, they both suspected an open feed valve would overwhelm the flare system. They expected to prove the plant had an expensive problem. As it turned out, the simulation proved otherwise. With a better understanding of process thermodynamics and some smart operating choices, they could show there was no credible risk of exceeding flare capacity.

While they did not find the expensive problem they expected, they did uncover a real issue. The flowmeter downstream from the feed valve was not capable of measuring the fully open flow rate. If their scenario occurred in real life, plant operators would get incorrect flow readings and might not understand what was happening in time to fix it. They were able to spec in a properly sized flowmeter at minor expense. In plants as in people, maturity is often about avoiding bad situations. Digital twins with process behaviour provide the safe environment needed to move fast, break things and learn how to fix them.

About the Author

James Wade is Process Simulation and Operator Training Expert at AVEVA. He uses his 20 years’ experience in engineering, sales, and marketing to help people discover better ways of problems in the industrial world. James completed his mechanical engineering degree at Carnegie Mellon University and his MBA at New York University.

Featured image: ©Simone_n