This paper investigates fault diagnosis in batchprocesses and presents a comparative study of featureextraction and classification techniques applied to a specificbiotechnological case study: the fermentation processmodel by Birol et al. (Comput Chem Eng 26:1553–1565,2002), which is a benchmark for advanced batch processesmonitoring, diagnosis and control. Fault diagnosis isachieved using four approaches on four different processscenarios based on the different levels of noise so as toevaluate their effects on the performance. Each approachcombines a feature extraction method, either multi-wayprincipal component analysis (MPCA) or multi-way independentcomponent analysis (MICA), with a classificationmethod, either artificial neural network (ANN) or supportvector machines (SVM). The performance obtained by thedifferent approaches is assessed and discussed for a set ofsimulated faults under different scenarios. One of the faults(a loss in mixing power) could not be detected due to theminimal effect of mixing on the simulated data. Theremaining faults could be easily diagnosed and the subsequentdiscussion provides practical insight into theselection and use of the available techniques to specificapplications. Irrespective of the classification algorithm,MPCA renders better results than MICA, hence the diagnosisperformance proves to be more sensitive