As artificial intelligence tools, particularly large language models (LLMs), continue to evolve, their potential for supporting technically complex domains such as pipeline integrity management is rapidly gaining attention.
In a new technical paper first presented at the 2026 Pipeline Technology Conference, Penspen digital transformation experts and leads from THEIA, Penspen’s Digital Integrity Management Solution, introduce a structured initiative to develop and evaluate a retrieval-augmented generation (RAG) system built on domain-specific knowledge of pipeline integrity.
The research has involved the development of a RAG architecture using curated datasets, technical standards, failure case studies, inspection technologies (ILI, CIPS, DCVG), and asset lifecycle data. The system, developed by Persona Dynamics and integrated as an advisory capability within Penspen’s THEIA platform, provides contextually grounded responses by retrieving relevant technical documentation before generating outputs.
The model’s responses are evaluated through a series of technical challenges posed by independent subject matter experts. These challenges increase in complexity and cover corrosion threat assessment and fitness-for-purpose.
Validating AI for Pipeline Integrity Management
Scoring is conducted using a quantitative framework designed to assess three core criteria: Technical Accuracy, measuring the degree to which responses are correct and compliant with applicable standards; Actionability, assessing the practical reliability of recommendations for decision support; and Request Dependency, evaluating the extent to which response quality is affected by the specificity or structure of the user query. Each output is rated for its usefulness and risk of misdirection, providing objective metrics on the trustworthiness of LLM-generated guidance in safety-critical applications.
The outcomes of this evaluation inform the development of LLM-driven tools that aim to safely augment engineering workflows, improve consistency in integrity assessments, and support the training of early-career engineers, while identifying the current limitations and governance requirements for AI in high-stakes infrastructure contexts.
Complete the form to access the full technical paper.
Download technical paperRelated Insights
Subject Matter Expert Evaluation of Pipeline Integrity Trained, Agentic Artificial Intelligence Based Technical Advisory System
As artificial intelligence tools, particularly large language models (LLMs), continue to evolve, their potential for supporting technically complex domains such as pipeline integrity management is...
Artificial Intelligence in Pipeline Integrity: What the Evidence Actually Says
Where can AI be trusted in pipeline integrity management? In the August issue of World Pipelines, Penspen digital experts explore the evidence from Penspen’s evaluation of Aura, an AI-based...
From Chokepoints to Corridors: Why Energy Security Now Depends on Infrastructure Resilience
In the 31st issue of Dar Magazine, Nigel Curson, EVP of Technical Excellence, explores why resilient infrastructure will play a defining role in Africa's long-term energy security and economic...
Industry Report – Capability Gap Analysis for Newbuild Offshore CO2 Pipeline Transportation Systems
Penspen's new report, commissioned by the North Sea Transition Authority, examines the challenges posed by transporting carbon dioxide through newbuild offshore pipelines, and addresses the steps...