Optimized Fuzzy SVM with Chaotic Henry Gas Solubility Algorithm for Fault Identification in Rotating Machinery

Reliable and accurate fault diagnosis in rotating machinery is vital for minimizing unplanned downtime, reducing maintenance costs, and ensuring operational safety in industrial environments. Traditional diagnostic approaches depend heavily on manual feature extraction from vibration signals, which can be time-consuming, expertise-dependent, and prone to missing subtle fault patterns. This study presents a novel hybrid framework—IDL-OFSVM—that combines Intelligent Deep Learning (IDL) with an Optimized Fuzzy Support Vector Machine (OFSVM) for automated fault classification. Vibration signals are first transformed using the Continuous Wavelet Transform (CWT), and deep features are extracted via the lightweight MobileNet architecture. The Chaotic Henry Gas Solubility Optimization (CHGSO) algorithm significantly enhances the classification model's performance, which effectively tunes the FSVM parameters. Experimental evaluations on benchmark datasets show that the proposed method achieves 99.8% training and 99.7% testing accuracy, outperforming several state-of-the-art approaches. Beyond technical accuracy, the framework offers practical advantages, including reduced dependency on domain expertise, suitability for real-time monitoring, and potential integration into predictive maintenance systems. These benefits make the IDL-OFSVM model a promising solution for industrial fault diagnosis applications, where reliability, speed, and scalability are crucial.