📢 New Publication Featuring GEAR Lab Contribution in Process Safety and Environmental Protection!

We are pleased to share a new peer-reviewed publication in which GEAR Lab member Dr. Lei Zou contributed as a co-author:

🎯 “Fine-tuned large language models for natech analytics: Evidence from two decades of Texas chemical emission incidents”

🔍 Study Overview

Natural-hazard-triggered technological accidents (Natechs) remain a significant yet under-examined risk for chemical process industries. Although decades of incident records exist, they are largely unstructured and challenging to analyze at scale.

This study introduces an automated, LLM-driven analytical framework that leverages fine-tuned large language models to systematically analyze long-term chemical emission incident reports across Texas (2004–2024).

Using the Texas Commission on Environmental Quality (TCEQ) dataset, the framework performs three key tasks:

1. Classifies whether an incident is Natech-related and identifies its primary hazard.

2. Extracts unit–issue pairs describing which industrial components failed and why.

3. Generates concise, evidence-style justifications directly from narrative text.

🏷️ Keywords

Large language models, Natech, natural hazards, industrial safety, chemical emissions

📄 Read the Full Paper:👉 https://www.sciencedirect.com/science/article/pii/S0957582025012406

🎉 Congratulations to all authors, and to Dr. Lei Zou for representing the GEAR Lab in advancing AI-driven Natech analytics and industrial safety research!