ABSTRACT
In this paper, signal processing is integrated with inverse techniques to solve inverse
problem for internal parts of machinery in which, sources are unknown and
inaccessible. Explosions of an internal combustion engine produce powerful pulses of
energy which cause the engine vibration as a response. Attenuation of engine vibration
is one of the important problems for NVH engineers. Engine vibration can affect the
whole chassis vibration and interior noise level. Engine vibration is the dominant source
of noise emissions (less than 200Hz) inside the vehicle compartment. In this research,
inverse techniques are implemented to estimate unknown amount of vibration sources.
Tikhonov regularization carried out to stabilize solution of ill-posed problem. L-curve
method is also utilised to find the regularization parameter.
In comparison with transfer path analysis, this method doesn't need to separate the
observation part from the system. Despite the coherence based methods, it is practical
in low frequency analysis of engine with missing inputs and outputs.