Research and Application of a Novel RPCA-SVME based Multiple Faults Recognition

In the modern industrial process, the likelihood of the occurrence of multiple faults is higher than that of a single fault Comparing with single faults, the multi-faults problem has higher coupling and complexity, thus it is quite important to establish an effective multi-faults recognition model to ensure process safety. In this paper, a multi-fault recognition model based on reconstructed principal component analysis (RPCA) algorithm and support vector machine ensemble (SVME) classifier is proposed to satisfy the needs. First, obtain the principal component information from the original high-dimensional data space. Second, to solve the loss of local feature information, reconstruct the local structural error of the feature space through the inverse mapping matrix, and then align the error to obtain the reconstructed coordinates. Third, based on the One vs. One (OvO) ensemble strategy, an SVME classifier is constructed for multiple faults recognition. Finally, to verify the performance of the proposed RPCA-SVME model, the simulation experiments are made on a Circle dataset and the Tennessee Eastman process (TEP). The comparison results show that the proposed method can guarantee higher diagnostic accuracy and macro F1 score.