First results of an ai based classification of vibration data enabling preventive maintenance for rotatory flexible hose assemblies in parabolic trough collectors

Abstract

This study explores the integration of vibration analysis and AI to allow predictive maintenance of swivel joints in rotatory flexible hose assemblies (RFHAs). Four vibration sensors were installed and an AI-based aging classification using Support Vector Machine was conducted. For a reduced sample dataset, results showed a promising accuracy, reaching up to 100% for specific angle bins. Principal Component Analysis highlighted the interest of analyzing defined angle bins to reduce computational costs. The study underscores the potential of this method and emphasizes the need for further research to compare AI models and select specific features and angle bins.