A.I. Masterclass in Petrochem Using Artificial Intelligence to Improve Operational Efficiency, Safety and Sustainability

By Citalouise Geiggar, Ph.D and Vice President of Marketing at Radix North America

Extracting AI value and dispelling AI hype is not an easy journey for asset-intensive companies to take. Yet, when employed correctly, AI provides the ability to boost asset-intensive operations to new heights – making the journey worth it, as long as they follow the right path below.  

Recently, I sat down with my colleague, Natalia Klafke, EVP of Energy & Sustainability at Radix, to discuss her last article on how Radix uses AI to solve supply chain disruptions and global inflation challanges across asset-intensive industries. Natalia also shared with me additional insights on how Radix helps leaders leverage AI to streamline operations and it led to the insights below.  

Demystifying AI Value 

In a relatively brief time, AI technology at Radix has grown from using a simple language model only capable of responding to if/then statements into a deep learning program capable of solving complex problems using large language models that contain a vast majority of human knowledge. And the new generative AIs use that baseline knowledge to foster connections and make predictions that elude many users. 

This has allowed business leaders throughout the corporate world to test the limits of what AI can do for them. 

At Radix, we first started using AI - in areas like Petrochem - to help us understand the physics of the equipment being used, especially the thermodynamics and chemical reactions. This includes using AI to boost the first principal models we built, resulting in an ability to now predict failures without having to physically experiment with actual chemicals. 

Being able to predict potential failures is important when dealing with hazardous materials such as petrochemicals. In addition, maintaining accurate models can be a time-consuming task. A person can build optimization models in spreadsheets using tools such as Solver, but it can be a challenge to devote enough time and people to build the models within the required period. This is of course before you factor in maintaining the model and distributing the information to those who need it the most.  

In some cases, it's easier because you have online and connected instruments or continuous processes in the refining process. But when you go to, for example, what is the deactivation time for that specific catalyst, it is an extremely hard problem or variable to predict –and one made more accurately and effectively by AI.  

Asset-intensive companies operate a wide variety of machines with complex parts, are Radix is always looking for ways to protect their capital investments. By utilizing AI to better understand the condition of the various parts and their potential failure points, we helped prevent downtime, optimize maintenance schedules, and develop ways to use less fuel or energy.  

According to Natialia, “AI can avoid the 3:00 AM phone calls to come into the plant, perform maintenance, or find a vendor to perform maintenance at a moment’s notice. You want to optimize downtime. And if you manufacture specialty chemicals then you want to make sure you don't lose production because the demand for that product is so high. AI can help in these and other situations.”  

As a side note: the safety benefits, which AI provides, are also a real benefit for petrochemical companies. For example, Natalia mentions that there are programs we run at Radix where we utilize image processing to ensure that operators are wearing the correct personal protective equipment in restrictive areas. If the AI detects that an employee is not wearing all the right equipment, or is wearing it incorrectly, it can trigger an alert or an alarm until that employee fixes their gear. The same AI-based image processing software can also be used to ensure that escape routes and emergency access points are free of debris in case of an emergency.  

Despite AI still having a long development road ahead of it, companies like Radix are already using it to improve both safety and operational efficiency. Additionally, AI has given engineers in Petrochem and beyond more control over complex processes that have a substantial chance for error if mixtures or parameters are not precise and correct. Given its impressive early gains and the massive potential to help with future operations, we can expect many companies to follow Radix’s lead in AI application development and implementation.