How AI Is transforming refinery construction

Artificial intelligence (AI) can help enhance refinery construction by leveraging its ability to control and enhance data.

By Anna Liza Montenegro October 14, 2022
Image courtesy: Brett Sayles

Refinery construction is often viewed through the lens of utilitarian design. But anyone who’s worked in the industry can attest to the fact that this by no means implies stagnancy or an unwillingness to work with new technologies. However, refinery construction is an incredibly competitive field that needs to contend with a constantly shifting market. Doing so often means that people within every area of the field will need to work with advanced technologies. And some of the most impressive examples of this principle have come from an unexpected source – artificial intelligence (AI).

What is artificial intelligence?

AI is an important and often misunderstood element within the IT industry. Most people conjure up images from science-fiction when they hear about artificial intelligence. Those with some more practical experience in the subject might think about the digital assistant on their smartphones or self-driving cars. But even those practical examples gloss over some of the most important aspects of artificial intelligence.

The true nature of I can be most easily grasped by considering another name for the subject – machine learning (ML). It might be tempting to classify AI by a single instance of it within your own life. However, judging AI by a single implementation is somewhat akin to judging humanity by a single individual’s job performance. One of the most important aspects of humanity and AI is their ability to learn and adapt to new information. Humans can learn about new subjects and leverage their innate advantages within those fields. Artificial intelligence can do the same. AI is best understood by looking at how it operates rather than a specific example of that process.

AI uses ML to take in data relevant to specific fields. This is in contrast to traditional programming which focuses on fairly inflexible scripting. You can think of this in terms of a decision-making flow chart. A traditional program might examine a specific data point to match it to a designated rule-based decision-making tree. The script might specify that if the value is X then the program needs to add Y to it and then execute function Z. The speed of a modern computer means that traditional programs can execute thousands of simple comparisons in a split second. But this process lacks any real ability to learn or make decisions about scenarios that don’t fit into a pre-defined script.

Leveraging AI within the appropriate domains

It’s important to keep in mind that AI is generally highly specialized. This specialization is why you typically use specific AI for specific tasks. AI designed to drive a car is extremely good at working with every element of that car. But the AI won’t be able to tackle any of the tasks that Alexa or Siri handles with ease. And of course, the reverse holds true as well. You usually can’t even move specialized AI around within specific subdomains. For example, Siri wouldn’t be a good match for an Android device’s unique operating environment. Even specific usage scenarios tend to be more or less compatible with any given AI. This principle also holds true for work in refineries.

AI as a whole holds tremendous promise for the refining industry. Machine learning and artificial intelligence can be leveraged for nearly any task related to digital operations. In fact, many experts are pushing toward an ultimate goal of nearly autonomous factories and refineries that are efficient and even self-optimizing. AI can be seen as an important building block within the larger infrastructure of refineries. The importance of AI to the larger-scale operation of refineries will almost certainly grow at a steady pace.

Refinery construction can and generally should be separated into distinct operations. Enterprise-level IT typically uses the word solution to describe individual packages whether that means hardware, software, or the more typical bundling of both into a singular entity. It’s best to consider the role of AI in refinery construction within that context of problems and solutions. Every element of a refinery’s construction and operation consists of problems. And AI requires continual evaluation as a potential solution for those problems or usage scenarios. However, one area in refinery construction has shown particular promise as a venue for integration with AI – computer-aided design (CAD).

Computer-aided design and AI

CAD is an indispensable element of modern architecture and construction. CAD has largely supplanted standard drafting within most industries, and people tend to think of it within that context. One of the reasons for CAD’s popularity stems from the fact that it leverages all of the features unique to computers. Almost any digital system can be integrated into computer-aided design and drafting.

You can find a tremendous variety of CAD software packages on the market. One of the most important elements of these CAD packages is what computer systems they can integrate with and how they can make use of that system’s data. This is especially relevant in the context of machine learning. As you’ve seen, one of the most important elements of AI is its ability to take in data and actually learn from it. People often discuss machine learning in the context of semi-autonomous agents like the personal assistants used in smartphones. But this type of learning is also relevant to processes that act as an assistant within the design process.

The importance of data and design

You probably think of digital data in the context of specific files inside your computer. However, in the context of AI, it’s better to think of almost everything in the world as data. Consider a situation where you’re using CAD for refinery construction. The process will involve countless decisions based on conditions relative to geographical conditions, budget, building material, and a wide variety of other factors. Now think back to the earlier discussion of logic and decision making. Many of these design decisions fall under the larger banner of fuzzy logic.

You don’t necessarily think of manufacturing methods for refinery construction, an area’s substrate, and cost concerns in the context of a pre-defined decision tree. You’ll instead use a combination of past experience and intuition to decide how all of those factors work with each other. This type of fuzzy decision-making is difficult for standard programs. But it’s well within the wheelhouse of AI that acts as a predictive expert system that can leverage large amounts of existing data.

Now imagine this type of help as an integrated feature within the CAD package you’re using for refinery construction. It might seem like a futuristic innovation that won’t be ready for years to come. But in fact, this exact type of generative design AI has already been implemented within some iterations of CAD software.

Construction, simulation, and optimization

This is a lot of information to take in, but step back and consider how all of these tools might come together within a project involving refinery construction. You might begin the process using CAD software in a manner that wouldn’t seem too out of place to someone trained with a drafting table. But now imagine the possibilities inherent to intelligent simulation using AI.

As you have new ideas you can put them to the test within the CAD system. You can try out different materials against the larger topology. You can instantly test how well various ideas integrate into the larger assembly process. And in general, the AI can act as a predictive agent to show you how well concepts might work in real-world situations. You could even use the AI within a CAD system to optimize solutions that you’ve settled on.

Looking toward the future

AI and machine learning are still developing at a rapid pace. CAD software has already demonstrated just how powerful integration of design and AI can be. But as amazing as these enhancements are, you can expect far more for the future. Refinery construction is a complex process involving a wide variety of different situations and components. And this is exactly where both AI and computer-aided design shine most brightly. AI is a powerful tool for the present, but it also holds considerable potential for the future.

– Downtown Ecommerce Partners (DEP) is a CFE Media and Technology content partner.

Original content can be found at Oil and Gas Engineering.

Author Bio: Anna Liza Montenegro is a trained architect and an accomplished marketing professional in the architecture, engineering, and construction (AEC) industry, she possesses both strategic and execution of marketing initiatives, go-to-market plans, and execute product launches. At Microsol Resources, she develops the marketing strategy, brand management, digital marketing, and other demand generation activities for Microsol’s strategic partnerships with Autodesk, McNeel Rhino, Bluebeam, Enscape, Chaos Group V-Ray, Panzura, Ideate Software, FenestraPro, and other partners. When not marketing, she loves spending time with her kids, traveling, and summers in Maine.