To reduce downtimes and extend the lifetime of components, fault detection and identification become more important in production plants. Sensors and other information sources can be deployed for condition monitoring and fault diagnosis. In this contribution, first a review on the application of Short-time Fourier transform, continuous and discrete wavelet transform, Wigner-Ville distribution, and empirical mode decomposition to strip rolling mills is given. The suitability of these data analysis methods in the field of non-stationary signals is analyzed. Secondly, these methods are applied to hot strip mill data to detect deviations in strip travel that affect product quality or cause downtimes. Their capability to record characteristic fault features is evaluated. The results show that applying wavelet transform and empirical mode decomposition methods are able to detect process changes, here related to irregular production conditions and results. (C) 2015 Elsevier Ltd. All rights reserved.