2025 Spring Meeting and 21st Global Congress on Process Safety
(7c) Enhancing Ethylene Plant Operations through Generative, Predictive, and Prescriptive Applied AI Solutions
Author
Pratap Nair - Presenter, Ingenero
The integration of generative capabilities into predictive and prescriptive analytics within Applied AI solutions has significantly augmented the operational efficiency of a leading petrochemical manufacturer. This innovative approach resulted in a remarkable 134 KTA increase in production, 2,475 GBTU in energy savings, a reduction of 164 KTA in CO2 emissions, a 3% decrease in overall greenhouse gas emissions, and the prevention of unplanned shutdowns at least three times within the first year across various production units, including Olefins, LDPE, LLDPE, HDPE, EO-EG, Ammonia, and Methanol plants.
An effective Applied AI solution necessitates the meticulous orchestration of several key components:
- Selection of strategic use cases tailored to specific operational challenges
- Identification and integration of relevant data from sensors and advanced smart sensors, ensuring redundancy and capturing additional insights on unmeasured parameters
- Robust data handling and processing methodologies
- Application of advanced analytics leveraging appropriate algorithms to facilitate predictive, prescriptive, and generative analytics
- Comprehensive presentation of insights, recommendations, and forecasts, alongside descriptive analytics
- Capability for scenario analysis to support diagnostics and optimization efforts
- Seamless integration with existing or reengineered workflows
- Automation of action implementation through integration with advanced process controls (APCs) at the shop floor level
- Provision of remote access for experts to monitor and analyze data
- Utilization of collaboration software to enhance team communication and decision-making
This paper will delve into the necessity and functionality of each component outlined above, alongside the enabling technologies and challenges encountered during the implementation of these Applied AI-based digital solutions across diverse geographical settings. Insights drawn from practical experiences will illustrate how value is generated through this comprehensive approach, ultimately advancing the operational landscape of ethylene production.