AI-based Quality Prediction and Operation Optimization Platform for EDC Cracker

Recognized for Field-Validated Work in High-Temperature Process Intelligence

SIMACRO is pleased to share that we received the highest evaluation score and were selected as the top-ranked project in a competitive Korean government-funded industry program focused on AI-based EDC cracker modeling and operations. 

Why EDC Crackers Are a Hard Problem

For high-temperature cracking processes, the internal states that govern performance cannot be directly measured during operation. Coke buildup, temperature profiles along the radiant tubes, and pressure drop must instead be inferred from boundary conditions. As a result, decisions regarding cracking severity, decoking schedules, and energy management have historically relied on operator experience and reactive decision-making.

This is where a digital twin platform provides significant value. ProcessModel V™ enables users to build, deploy, and operate models that estimate unmeasurable process variables—commonly referred to as soft sensors or soft-sensor variables—by combining first-principles models, AI/ML models, and plant data. By making these internal process states more visible, users can make better-informed decisions. For example, more accurate estimation of coke buildup can help optimize decoking schedules, reducing unnecessary downtime while minimizing fouling risks. In another real-world application, temperature profiles along the radiant tubes were estimated and optimized, resulting in energy savings of up to 6%.

Four to Five Years of Field-Validated Work

Over the past four to five years, SIMACRO has worked with global energy and chemical companies to build and validate real-time digital twins for high-temperature cracking processes. That experience has established field-validated expertise in EDC cracker reaction dynamics and operational trade-offs that conventional approaches cannot reliably address.

The Next Step: An AI Agent for EDC Cracker Operations

The goal now is to operationalize this expertise through an AI Agent purpose-built for EDC cracker operations, capable of autonomously optimizing the process in real time. We are targeting rapid commercialization by 2027.

We are proud to have been recognized for this work and look forward to delivering on it.

Written by Hyejin Cho
SIMACRO media team | media@simacro.com

Next
Next

Webinar with Peter Reynolds