Complex Eigenvalue Analysis of Railway Wheel/Rail Squeal: Advanced Study

Wheel/rail squeal noise of trains is one of the challenging problems to be solved. Several models such as
self-excited instable vibration due to negative damping, flutter instability, hammering effect due to local
defects, etc. were proposed, but a successful model for the squeal problem has not been presented until
now. In this study, experimental and numerical methods were applied. Squeal noise and acceleration of
the rail were measured in a depot when a subway train was running back and forth at a speed of 20,
25 and 30 km/h. Squeal frequencies above 80 dB(A) were 480, 1220, 2200, 4340, 5420, 8860, 9640,
and 9960 Hz, which were consistent with natural vibration frequencies of the rail or wheel. A new
numerical approach was presented to investigate the basic mechanism of the wheel/rail squeal noise, using
complex eigenvalue analysis by the finite element method. The positive real parts of the eigenvalues reflect
self-exciting instable vibration, which is closely related to the occurrence of squeal noise. The effect of
parameters such as friction coefficient, wheel/rail contact position, axle load, etc. on the instable vibration was
examined. The instability of the vibration system was sensitive to the stiffness of rail support. In lateral
creepage, when the adhesion coefficient was less than 0.1, instable vibration modes did not occur. In
longitudinal creepage, when the friction coefficient was high enough, instable vibration modes were
generated. The predicted unstable modal shapes were different from those in lateral creepage. The
longitudinal creepage caused the natural vibrations of the axle. Numerical predictions could explain
many of the field test results.