(4b) Key Features for Successful AI Applications | AIChE

(4b) Key Features for Successful AI Applications


Thorat, S. - Presenter, Ingenero Inc.
Jaiswal, N., Ingenero
Nair, R., Ingenero Inc.
Producing Ethylene has many challenges due to process complexities caused by the large temperature and pressure differentials required to crack and recover products. These complexities create tremendous opportunities for AI/ML to address these challenges efficiently. AI/ML models develop a mathematical model of the relationship between measured variables in the manufacturing process that are unique to the particular unit and can rapidly adjust the unit to optimized operation. When accurately applied AI/ML can drive significant and highly valuable process improvement.

The challenge is to utilize technology appropriately to ensure success and capture the desired benefits. Effective utilization of AI can achieve greater operational efficiencies, maximized capacity utilization, lower production costs, improved quality, reduced emissions, enhanced safety and increased reliability. The uses are numerous, but success is highly dependent on necessary key elements for AI applications that are often ignored.

A smart analytical AI engine is expected to generate actionable intelligence. Information must be generated that can be easily interpreted and applied either directly or indirectly by effectively combining human intelligence and control systems. Dashboards from an AI solution must be easy to use and accessible to enhance decision making across all levels of plant operation. Considering the complexities in the ethylene process, having these “Smart” dashboards with interactive levels to provide the requisite insights to enhance decision making and drive operations to an optimal level is particularly important.

A critical process change in an Ethylene plant will impact operations at multiple points. Optimal response requires a multi-dimensional solution covering the wide spectrum of operations to achieve an overall synergistic response. To ensure this synergy and make the most effective solution, a single modeling technique may not be adequate. A hybrid modeling approach which utilizes the power of AI/ML and 1st principal models together, can often address the intricacies of the ethylene process more effectively and with the highest accuracy.

AI solutions create maximum impact due to their ability to deliver insight on a real time basis. The model predictions must be quick and accurate. Compromising on accuracy is not an option as each action has its own consequences, especially in ethylene process. To ensure accuracy, AI solutions need to be widely compatible to interact with various data sources for connecting the dots precisely. Development and utilization of soft sensors to generate the critical and accurate data insights and confirm accuracy can be critical.

Checking financial impact of process change and customizing the AI solution to the address the manufacturing problem is critical to capture value. Ensuring the solution is addressing the most financial impactful areas is key.

This paper will present the various improvement opportunities possible from AI/ML solutions application in the ethylene process. The critical features of an AI/ML solution necessary to ensure success and maximize benefits will be explained using ethylene case examples. The focus will be on highlighting critical aspects necessary for successful development and deployment of real time AI applications in ethylene.



This paper has an Extended Abstract file available; you must purchase the conference proceedings to access it.


Do you already own this?



AIChE Pro Members $150.00
AIChE Emeritus Members $105.00
Employees of CCPS Member Companies $150.00
AIChE Graduate Student Members Free
AIChE Undergraduate Student Members Free
AIChE Explorer Members $225.00
Non-Members $225.00