Terence Connelly

The Use of in Vehicle STL Testing to Correlate Subsystem Level SEA Models

For the assessment of vehicle acoustics in the early design stages of a vehicle program, the use of full vehicle SEA models is becoming the standard analysis method in the US automotive industry. One benefit is that OEMsand Tier 1 suppliers are able to cascade lower level acoustic performance targets for NVH systems and components. Detailed SEA system level models can be used to assess the performance of systems such as dash panels, floors and doors, however, the results will be questionable until test data Is available. Correlation can be accomplished with buck testing, which is a common practice in the automotive industry for assessing the STL (sound transmission loss) of vehicle level components. The opportunity to conduct buck testing can be limited by the availability of representative bodies to be cut into bucks and the availability of a transmission loss suite with a suitably large opening. In addition, the temporary fixture used to mount the test structure does not typically provide representative boundary conditions. Boundary conditions are particularly important when the in-vehicle performance is dependent on point connections at hinges / locking mechanisms and flexible seals at the perimeter of the system such as found in doors or lift-gate. In this paper the use of in-vehicle testing on prototype vehicles to assess the STL for the vehicle systems and the SEA models are discussed. Results are presented for a dash panel, floor system and door systems. The measurements are compared with predictions from system level SEA (Statistical Energy Analysis) models of the test systems. It is concluded that subsystem level SEA models can be used to assess the design and correlate well with STL measurements from in-vehicle tests.

A Study of NVH Vehicle Testing Variability

At certain key stages in the vehicle development process, prototype vehicles are available for NVH testing. This testing fulfills two functions: primarily it is used to assess the status of the vehicle to the program NVH performance targets, but it also provides an opportunity to validate the vehicle SEA model. These single vehicle test events provide a snapshot of the NVH performance but do not provide any understanding of the variability of the NVH performance, which is due to many factors: components, build or assembly and test setup variability. SEA models can be used to estimate the vehicle level variability, if the variability of the interior components is understood, but there is limited data available to confirm the accuracy of these predictions. In this paper we examine the repeatability and reproducibility through a standard gage R&R study of Engine Noise Reduction (engine NR) and Tire NR testing. We also examine the variation in NR from vehicle-to-vehicle through testing on six compact sedans of the same vehicle line. A more limited chassis dynamometer study on three of the six vehicles is also presented to look at typical running vehicle measures such as articulation index, loudness and overall SPL for tire and powertrain noise. Although the sample population is small, we have extracted estimates of variation in body noise reduction due to test setup, actual component and build variation and the implications for predicting variability through SEA models.