Application of natural language processing for spill reduction in an exploration and production company

This paper applies natural language processing (NLP) to analyze loss of primary containment (LOPC) incident reports in exploration and production operations. The study employs both rule-based entity extraction and machine learning approaches to identify causes and contributing factors of spills and leaks. The research demonstrates how NLP can process large volumes of incident reports quickly, improving data quality and enabling trend analysis. For PSM, this work is highly relevant as it addresses a key challenge in incident investigation: extracting meaningful insights from free-text descriptions to identify equipment failure patterns, prioritize interventions, and reduce spill frequency. The methodology helps safety managers identify systemic issues across multiple incidents that may not be apparent through manual review, supporting more effective root cause analysis and preventive measures.