Effect of tail-rotor torque variation on vibration at helicopter tail-rotor drive-train

Condition monitoring and fault diagnostics in rotorcraft have significant effect on improving safety level and reducing operational and maintenance costs. Most of the commercially available diagnostic tools for rotating machinery are based on vibration analysis. In this paper, the effect of torque variation on vibration is investigated at the tail rotor drive train (TRDT) of the Apache helicopter. Vibration used in this study is measured at different torque values simulating different loads on the tail rotor. At each load value, vibration is measured using four sensors distributed over the tail rotor drive train of a dedicated AH-64D helicopter test bed. The proposed method depends on simulating a finite state machine where each state represents a load value among five different applied loads. Within each state, TRDT condition is modeled using parameters extracted form cross correlations among the four accelerometer sensors. Fuzzy logic is then used to cover the transition regions between states. Fuzzy logic uses three features which are selected using a proposed search algorithm from large number of features calculated from the vibration bispectrum. In order to quantitatively evaluate the proposed method, collected vibration signals are divided into training and testing samples. Training samples were used to train the fuzzy logic system, while testing samples are used to evaluate its performance. The proposed algorithm shows an interesting performance in determining the vibration behavior during torque variation on the TRDT parts.