Driving Safety Excellence: Strategies for Minimizing Land Transportation Incidents in the Oil and Gas Industry

Abstract

This research presents a smart accident risk appraisal framework tailored for the oil and gas transportation division, where overwhelming vehicle occurrences pose critical operational and security concerns. The study leverages a machine learning model—specifically, a Random Forest Regressor—trained on normalized transportation data to foresee mischance likelihood based on key factors such as truck weight, driver encounter, travel timing, and maintenance status. The demonstration showed high predictive precision with an R² score of 0.95 and a mean absolute error of just 3.42%. A user-friendly web application was created utilizing Streamlit, empowering logistics staff to input trip subtle elements and immediately get hazard expectations. The app is improved by joining Failure Mode and Effects Analysis (FMEA), giving relevant experiences, proposing remedial activities, and visualizing Risk Priority Number (RPN), which has decreased in recent times and after mitigation. Also, the app incorporates usefulness for importing updated Excel records, permitting continuous advancement and versatility to changing conditions. The project adopts the Six Sigma DMAIC (Define, Measure, Analyze, Improve, Control) framework to ensure a structured approach to risk management and process improvement, aligning the AI development process with proven quality improvement methodologies. To guarantee long-term control and reliability, the project presents a strong framework, including data updates, standard operating procedures (SOP) standardization, retraining plans, and monthly reviews. A key proposal includes extending the dataset to incorporate driver scheduling data to capture fatigue-related dangers. The venture demonstrates how AI-driven decision support can significantly improve transportation security in high-risk industries when combined with structured risk management strategies like Six Sigma.

Date of Award	2025
Original language	American English
Supervisor	Saed Amer (Supervisor)