Investigation of the acoustical behaviors of vehicle doors

Automotive market has been grown up day by day in both our country and the world. Increasing number of companies with different technologies are struggling to get a foothold for themselves inside the market. Developing technology is being used for supplying the customers not only with the standard safety and design requirements but also for more comfort and quality. Noise and vibration performance of the vehicles which is an open area to be developed is coming as the first subject being worked on within that comfort area. Increased vibration and noise performance of the vehicle also improves the customer perception of quality for the vehicle. For that reason, the subjective evaluations of the customers are being taken into consideration as a leading indicator for noise and vibration studies within the research and development departments of the automobile companies. One of the most important performances of the vehicle which improves the customer perception about the vehicle quality is the door slam noise of the vehicles. It is known that a tough and brief sound of door slam creates the idea in people's mind that the car is of good quality. This knowledge is the main reason for the reality that the studies for improving the impression that the door slam sound creates on customers are one of the subjects on which noise–vibration-harshness (NVH) teams of most of the global automotive companies' research and development departments are spending long time. This subject is being investigated under the sound quality topic which is one of the most important areas of the acoustic discipline. The first purpose of those studies is to understand which sound quality metrics effect customer perception in which way or at which levels of the sound quality metrics should be for a desirable door slam sound by ensuring the validation of the objective and measurable sound quality metrics with the subjective jury evaluations. The final purpose on the other hand is to specify the designs which enable the desired levels for the sound quality metrics. Although such type of studies have been performed until today, it can be seen that nearly all of these studies are comprised of experimental ones. Creating a Commputer Aided Engineering (CAE) methodology which is capable of leading the vehicle door design studies taking also the sound quality parameters into account during even very early phases of the product development processes would enable the studies in this field to be run faster, more effectively and less costly. The study, which is being presented in this thesis, was performed in order to achieve this goal. Firstly, a numerical model was developed which is capable of calculating the door slam noise in time domain. This numerical model was constituted by using the Finite Element (FE) model of the door of a vehicle, which had readily been prepared for manufacturing with a specified design. Two acoustic cavities representing the interior and exterior air were created and coupled with the structural model which consists of the door to be slammed with the cut body components it is connected to. The model was solved with Abaqus Explicit solver as transient and the transient sound pressure levels were calculated both at the ear point of the person closing the door outside the vehicle and at the right ear point of the driver inside. The physical car was also tested in a semi- anechoic room and the same type of sound pressure data was measured from the same points as in the virtual model. The validation study was performed by tuning the virtual model paramteters in order to optimally predict the experimental result. In the next step, some design modifications were made on this verified model and the analysis was re-calculated for each design change. The two of the most important sound quality metrics found in the literature survey, Loudness and Sharpness, were calculated for each parametric design change situations by using the transient sound pressure curves gathered from numerical calculations and for current situation by using transient sound pressure curves gathered from both numerical calculations and experimental measurement. By this way, the information that which design change affects the sound quality metrics in which way was gathered on the virtual model. For the current situation results, it can be said that, band – pass filtered transient analysis results were well correlated with experimental results. For the sound quality parameters, although Loudness sound quality parameter calculated from analysis result is much higher then the Loudness calculated from physical test results, sharpness sound quality parameter calculated from analysis and physical test results are close to each other. It can be inferred from the results that the virtual model can predict the frequency distribution well within the frequency band investigated (22 – 500 Hz) but the amplitudes for each frequency component is much higher than the physical sound. To sum up: The virtual model has the ability to predict the frequency components of the sound successfully but is not as much successful as that for damping the sound. The only study found in the literature for developing a CAE method to predict the transient door slam noise is one from the year 2005. It is dependent on some physical measurement results as input and it solves structural and acoustical equations with 2 discrete steps and 2 different analysis methods (FEM + BEM). For these reasons, the study presented in this thesis has the charatheristic of being first in the literature. Firstly, this study is the first CAE methodology to predict the transient sound pressure for door slam without dependence to any physical measurement input. Secondly, it can compute the explicit non – linear structural and acoustical equations at the same time in a coupled manner with just using Finite Element Method (FEM). All those indicateds are advantages of the methodology and each is a novelty brought for such a study to the literature. Additionally, gathering sound quality metrics from also the analysis result enables those metrics to be evaluated for any desired design change even in very early steps of the product development process. The disadvantage of the methodology on the other hand is that it takes long time to solve such a large-scale model with just FEM in one step in a coupled way.