Despite naysayers who tend to believe that AI will hinder human interaction and, in some extreme projections, may eventually lead to a robot rebellion, Bruce Taylor, director of digital transformation for FurtherTec, said he believes AI “does have some very positive attributes.” Furthermore, Taylor said he has first-hand experience that shows it generates significant value.
The big value of AI is clearly within the power of computing, Taylor said.
“It can process a massive amount of information — different information types and attributes — in a very short period of time,” he said. “The resolution of the insights that are produced by AI can be generated very rapidly.”
Another positive advantage, Taylor said, is that once you turn it on, it can run 24/7/365.
“If you’re monitoring for developing conditions in equipment performance, it never sleeps, and the results are very consistent and repetitive,” he said.
On closer examination, Taylor admitted that some similarities exist between artificial and human intelligence. For one, they both require a certain amount of training.
“You have to establish a competency within AI systems, and it has to learn,” Taylor said. “Some of them are fairly rarified and difficult. The whole field of data science is still evolving.”
Data aggregation, Taylor said, is the normalization and contextualization of data.
“Instead of using the expression ‘data is gold,’ really, data is gold ore,” he said. “It still requires a lot of refining and processing to get value out. Even when you do clean and scrub the available data set, experience has taught that there are still biases and errors because the data will reflect how an operator has been running the system on a particular shift or crew.”
Bytes versus brains
“Refineries are never static. They’re always changing with turnarounds and engineering efforts that are continuously ongoing in the environment,” Taylor observed. “It requires a lot of time and effort to sustain these systems to make sure they’re reflecting current [demands of a site].”
The fastest supercomputer is still about 20-times slower than the human brain, Taylor said.
“The human brain can ‘rewire,’ innovate and evolve itself. Once AI is put in place, it does what it does very well, but it will continue to do that even if the [demands] change,” Taylor said.
Jerry Wood, system business analyst for Sinclair Oil Corp., argued that the transition to AI is smoother when the operations team is on board and accepting.
“Explain the ‘why’ to them,” Wood said. “They’re out there already, doing their rounds and checking on their equipment. When you show them the value of what AI can add to the equation, that really starts to bring the operators on board to participate in approving working with new technology.”
Wood suggested that operations workers be shown the steps that have been saved and reductions in running outside of the parameters, for example, in the application of AI to flare situations.
“Once they grasp the ‘why,’ they come on board and really work together with AI and machine learning,” Wood said.
Deidra Armstrong, director of operational excellence for Sinclair Oil Corp., agreed with Wood.
“Once they see AI is enabling them to focus on the right things at the right times, and they can spend their energy doing those things and doing them well, they will want more of it,” Armstrong said. “Achieving that balance is incredibly important to being able to see the value AI provides for them.”
However, not every solution and use case needs all of one thing or the other, Armstrong warned.
“You need to look at what’s right for that particular issue,” she said.
“Are you leveraging the maximum value out of the resources you already have?” Taylor asked. “When it comes to AI and human intelligence, you can make a case for just about anything. The question is ‘should you?’ Do you need to?”