Frequency Domain Order Tracking and Synchronous Average for Rotating Machinery Diagnostics

When we deal with rotating machinery, in order to monitor the characteristic and quality of the rotating parts, common method is to sample the accelerometer signals using discrete time axis with constant time interval. However, in most of the cases, rotational speed of the machine is nonstationary. Therefore, the characteristic frequencies of interest (speed of the shafts, gears and bearings) are not constant. So, signals sampled evenly in time cannot be used directly for condition monitoring. In order to deal with these problems we need algorithms that are robust to changes in rotational speed. In this regard, Order Tracking and Synchronous Average are well-established algorithms. Main idea of the Order Tracking is to find the specific sampling coordinates that result in signals sampled at constant angular increments, no more at constant time increments. OT also allows further averaging (Synchronous Average) to easily separate deterministic, periodic component and non-deterministic one. This separation for example allows for bearing characteristic signals to be distinguished from the gear contribution, fostering diagnostics. In this paper, a Novel Frequency Domain algorithm is proposed and its performance is compared with Benchmark algorithms as well.