Autonomous Fault Diagnosis and Root Cause Analysis for the Processing System Using One-Class SVM and NN Permutation Algorithm

This paper presents an autonomous, self-learning machine learning framework for fault detection, classification, and root cause analysis in complex chemical processing operations. The system integrates an incremental One-Class Support Vector Machine (SVM) to capture unknown process deviations and a Neural Network (NN) with a permutation algorithm to extract the precise variable contributions driving a fault. By dynamically updating a fault dictionary in real-time, this AI methodology automates the technical aspect of incident investigations during abnormal situations, helping safety engineers isolate the initial process abnormalities before they propagate into major accidents or catastrophic failures. Note: "Support Vector Machines" were a precursor to transformer models, the latter being the basis for today's frontier models.