The operational reliability of aeroengines is vital to civil aviation safety; however, bearings and other key components are prone to failure under harsh operating conditions. In real-world monitoring data, severe class imbalance often leads conventional fault diagnosis methods to be biased toward majority classes, limiting their ability to identify critical faults. To address this issue, this paper proposes a robust anomaly detection framework that integrates the Synthetic Minority Oversampling Technique (SMOTE) with a Broad Learning System (BLS). SMOTE is first applied to generate synthetic fault samples in the feature space, thereby balancing the data distribution and reducing bias. The b... More >