Simultaneous fault diagnosis based on multiple kernel support vector machine in nonlinear dynamic distillation column

Although numerous works have been done, most of the studies in fault diagnosis are limited to single type at a time. Majority reported literature do not extend root cause for simultaneous faults specifically distillation column. However, an industrial system is susceptible more than one time, which may or be interrelated. These only reduce performance but also increase computational complexity algorithm. In this work, therefore, multiple kernel support vector machine (MK-SVM) algorithm proposed diagnose developed MK-SVM algorithm, multilabel approach based on various functions has utilized classification faults. Dynamic simulation pilot-scale column using Aspen Plus® used generating data normal and faulty operation. Eight different types considered, including valve sticking reflux reboiler, tray upsets, loss feed flow, composition, temperature changes. faults, combination two, three, four introduced evaluation The result showed that high detection rate (FDR) 99.51% very low misclassification (MR) 0.49%. MK-SVM–based better with F1 score >97% all combinations Moreover, it observed shows single, multiple, as compared other established machine-learning algorithms.