An Ensemble Fault Diagnosis Approach for Multimodal Process

To improve the accuracy of fault diagnosis for multimodal process, an ensemble fault diagnosis approach based on IJITL-WEMD-RLSSVM (short words of, improved just-in time-learning (IJITL), empirical mode decomposition with window (WENID), recursive least squares support vector machine (RLSSVM)) is presented. Firstly, the corresponding data are found in historical data through IJITL method and the small sample data are output. Then WEMD method is used for data preprocessing of the small sample data. Finally, RLSSVM classifier is trained. To verify the validity, the proposed method is applied to a wastewater treatment process. The experimental results show that the proposed IJITL-WEMD-RLSSVM method is superior to the conventional method in speed and diagnosis accuracy for multimodal process.