Rechnerische und experimentelle Untersuchungen der Schallabstrahlung bei Fahrzeuggetrieben

In der Arbeit wird das Betriebsgeräusch eines Fahrzeuggetriebes simuliert und an einem PC-Arbeitsplatz auralisiert. Die Bauteilverbindungs-Modellierung beeinflusst die Ergebnisse der Akustiksimulation entscheidend. Verschiedene FE-Modelle mit Schraubenverbindungen und Wälzlagern werden daher durch Messungen verifiziert. Die Untersuchungsergebnisse führen zu einer Entscheidungshilfe bei der Modellierung von Bauteilverbindungen. An drei Getriebegehäusen werden auf Basis von Körper- und Luftschallmessungen sowie mittels hochwertiger FE-Modelle umfangreiche Untersuchungen der Luftschallsimulation mit der DBEM, der IBEM, der Fast-Multipole-BEM und der FEM durchgeführt. Es zeigt sich, dass Schallkennfelder sehr präzise vorausberechnet werden können. Die Genauigkeit der Luftschallsimulation lässt sich abhängig von der Komplexität der abstrahlenden Oberflächen darstellen. Die Ergebnisse eines Getriebegehäuses mit Versteifungsrippen werden für Akustikmodelle mit und ohne Rippen ausgewertet, so dass der Einfluss der Vernachlässigung dieser Geometriedetails erkennbar wird. Ferner werden Möglichkeiten zur Verhinderung fiktiver Resonanzen bewertet und Aspekte zur Wahl des Simulationsverfahrens diskutiert. Insgesamt werden umfangreiche Möglichkeiten zur Verbesserung der Schallabstrahlungs-Simulation aufgezeigt.

The target of the thesis at hand is the improvement of the acoustic simulation of vehicle gearboxes. Initially, the operating noise of a gearbox at constant speed is simulated. A method is presented, which decouples the computationally intensive part from the speed-dependent part of the simulation. Furthermore, the auralisation of the calculated noise is done with a standard PC. The results are based on many assumptions, but the audible, synthetic sound of the virtual gearbox is authentic anyway.The results of the structure-borne noise calculation within the acoustic simulation are influenced decisively by the modelling of the joints between components. Thus, FE models with bolted joints and rolling bearings are verified by measurements stepwise, separately and systematically. After the classification of the models for bolted joints, several variants are derived and realised for three different gearbox assemblies with flange joints. The results of the modal analyses are compared with measurements in a wide frequency-range comprehensively. Since the assemblies reach approximately single-component-quality, a recommendation can be given for the modelling of bolted joints. The accuracy of the results and the significance of the bolted joints can now be estimated for similar assemblies a priori. The resilience of rolling bearings is modelled by linearised stiffness matrices. The FE models of two assemblies are verified by measurements of gearboxes with non-rotating shafts at relevant eigenfrequencies. Therewith, uncertainties of the rolling bearing modelling and important influences are estimable. Altogether, the investigations of the structure-borne noise simulation are helpful for modelling-decisions of joints between components. Thus, costly verifications of similar models can be reduced in the future.Extensive investigations of the sound radiation simulation are conducted for three gearbox housings with varying complexity of their radiating surfaces. DBEM, IBEM, Fast-Multipole BEM and FEM are applied for acoustic simulations in a wide frequency-range. The investigations are based on structure- and airborne-noise measurements along with high-quality FE models. Input values are updated with measurements to separate the deviations of the sound radiation simulation. The results show that soundfields can be predicted accurately with different methods. Other aspects for the choice of the appropriate simulation method are discussed comprehensively, for example the calculation times. The accuracy of the sound radiation simulation can be presented dependent on the complexity of the radiating surface. Acoustic models with and without ribs are evaluated for a ribbed gearbox housing to show the influence of neglecting these geometry details. Furthermore, different methods to avoid fictitious resonances for BEM calculations are evaluated. Altogether, the thesis at hand extends the possibilities of acoustic simulations in several fields.

Zitieren

Zitierform:
Zitierform konnte nicht geladen werden.

Rechte

Nutzung und Vervielfältigung:
Alle Rechte vorbehalten