Analysis of repetitive mechanism signatures

Abstract A new signal decomposition for repetitive signatures is introduced. This is based on the concept of derived time series, produced by rearranging the original signal by using points one period apart. Time domain averaging is shown to be a special case of such rearrangement, and a second technique of variance analysis is presented. Applications to rotating machinery show that the decomposition can be very sensitive to various generating mechanisms, making them suitable for signature analysis of mechanical systems.