Bir ağır iş kamyonu üzerinde transfer yol analizi ve orijinal kuvvet belirleme yaklaşımı

Most of the mechanical systems are composed of different subsystems coupled by several links. Any excitation acting on the system is divided into several internal forces which propagate through these links or so called transfer paths. This thesis presents the use of experimental Transfer Path Analysis (TPA) in identifying the transmission paths and their contribution for a heavy duty truck in order to estimate the vibration and noise transmitted from the cabin and engine mounts. TPA uses the operating forces to calculate the response by multiplying them with the corresponding transfer functions. For that reason, predicting the operating forces accurately have become a major research topic and can be done using different methods. One of the methods that is commonly used is called "matrix inversion method" and it is used when the transfer paths include rigid connections or the mounts are very stiff compared to the receiving structure. When matrix inversion methods are used, the errors in the experimental measurements and poor conditioning can often lead to unreliable results. This thesis focuses on developing new techniques to improve the force identification process in TPA. The required data for TPA is measured on a heavy duty truck. The operating response values at the desired locations (steering wheel accelerations and sound pressure level inside the cabin) are also measured on the truck to verify the validity of the predicted values. First, classical TPA method is employed in order to predict the sound pressure level inside a prototype heavy duty truck cabin and the vibration level on the steering wheel. To validate the predicted results, coherence of the collected data and the condition numbers of FRF matrices are investigated so that the accuracy of the predicted results can be quantified for the frequency band of interest. The predictions of the total response are compared with the experimentally measured data such that the coherence and condition number related observations are validated. Then, a comparison of response predictions as a result of performing Transfer Path Analysis using the matrix inversion technique with different matrix arrangements, all of which are based on coherence knowledge and cross coupling omission is presented. Finally, a new method of improving the quality of the matrix inversion procedure and lowering the condition number of the matrix by selective omission of the cross-coupling terms is presented. It will be shown that the results of the new method are better compared to those of the Moore-Penrose pseudo-inverse technique also having the benefit of not dealing with selection of the singular value rejection threshold.