“…Monitoring variable speed machinery becomes difficult due to the interaction between angle, and time dependent characteristics (Abboud et al , 2015). Some current approaches to the monitoring of non-stationary machinery include signal processing techniques such as order tracking, time-frequency domain analysis and cyclo-non-stationary signal analysis (Uma Maheswari and Umamaheswari, 2017; Zhao and Wang, 2011; Law et al , 2012; Prudhom et al , 2017; Abboud et al , 2017) as well as machine learning algorithms that look for patterns in the data that are the result of changes in condition(Goreczka and Strackeljan, 2012; Wang and Kanneg, 2009). Solutions to monitoring non-stationary machinery typically rely on extensive training data sets that attempt to cover the entire machine operating window, as well as a priori information about the failure signature.…”