automotive body design
have on the ride shake performance. The CAE model results were then analysed to identify the causes of vibration peaks and design studies were performed to minimise them. Even without FE or other simulation
An in-house code from automobile supplier Autoneum: VisualSISAB
simulations from modal analysis, which extracts the dynamic characteristics of a vibrating system – in this case an automobile body and its sub-assemblies. Some up-front analysis with FE code is always necessary, and FE codes continue to dominate in the low frequencies (to roughly 500 Hz). For high frequencies, Heers agrees that SEA codes are the best approach. But then, he points out, there’s what people in the industry refer to as ‘the mid frequency gap’, from 500 Hz to roughly 1,000 Hz, where no method has yet to dominate because it involves a coupled structural mechanics/acoustics problem where traditional FE models are far too large and other methods don’t accurately represent the interplay between the body and acoustics. In the last decade, however, the AMLS
(automated multilevel substructuring) method has become dominant in this area. It’s an FE-based technique that can solve the modal space to 1,000 Hz. The most recent release of Abaqus from Simulia, Dassault Systèmes’ brand for realistic simulation, includes a parallelised direct dense solver in the frequency response phase that speeds up that method significantly. Heers cites the example of a body model with the acoustics cavity, computing a very detailed frequency response up to 1 kHz, including all modes up to 1,500 Hz. Previously, this took 16 days on a single core, comparable to competitive products, making it impractical for actual use by auto manufacturers. With the newest release, however, it takes 14 hours (four hours for the AMS run and 10 hours for the frequency response) when run on four cores. Heers adds that while studies above 1,000
kHz are important for powertrain analysis, that range is quite adequately captured with FE methods (due to much stiffer structures than car bodies). He also noted that there is sometimes a disconnect between what software can do and what customers want to
8 SCIENTIFIC COMPUTING WORLD
accomplish. He points out that while our ears can react to frequencies as high as 15 kHz, most sound energy is far below that; an opera singer has a high note of 1 kHz, the highest note of a flute is 2 kHz.
software, it’s possible to find unwanted noise and vibrations. Such is the experience of Jean-Louis Ligier, an R&D manager at Renault. When an engine slows down, components produce noise and vibrations that can lead to their deterioration. To study the problem, he modelled the engine with a system of five differential equations in Maple software. Not only could he enter the equations as if writing them by hand, ‘the fact that I can do symbolic calculations allows me to do optimisations that are virtually impossible with other software’. Ligier was looking for shock, which is considered to occur when the relative angular displacement of the crankcase to the frame is greater than a certain value. Using the equations, he plotted the relative displacements of components and checked when the value
ONE OF THE MAJOR CHALLENGES IS FINDING ENGINEERS WITH EXPERTISE IN BOTH AERODYNAMICS AND STRUCTURAL MECHANICS
A sit-in simulator Taking modelling to an even higher level, Bentley Motors wanted to understand the noise, vibration and harshness (NVH) implications of installing alternative powertrains in existing and new vehicles – principally those concerning the control of road-induced shake. For this, Sound & Vibration Technology (SVT), working together with David Fothergill of DJ Fothergill Consulting, developed a full-vehicle noise and vibration simulator that provides calibrated vibration stimuli to the driver, in addition to sound stimuli generated by Desktop NVH Simulator software from Brüel & Kjaer. The full-vehicle model was developed using the HyperWorks suite from Altair Engineering. The complete static vehicle is positioned in front of a projection screen, and the driver’s inputs are used to calculate vehicle parameters as it is ‘driven’. The sound and vibration stimuli for the occupants are calculated in real time and presented to them during a virtual ‘drive’. Altair’s HyperWorks CAE suite was chosen
to construct and run the system model. HyperMesh was used to read the CAD data, form the geometric databases and construct some of the component models. All the models were formulated as Radioss bulk data decks, which were read into HyperMesh and solved using Radioss. With that FE simulator, a subjective impression was gained of the effect that changing the power-unit mounts would
exceeded the given limit. Then he compared which one occurred first and which one was the largest to identify the root cause of the noise arising during the engine stop. Ligier notes that he created and ran this model in a single day – a time-saver in comparison to other software, which took over a week for the same task. As a supplier of soundproofing components
for the auto industry, Autoneum has a great interest in this class of simulation software. The company mainly uses software for acoustics, thermal management and CFD simulation, as well as vibro-acoustic testing of materials, components and systems, including up to a full vehicle. In addition to using commercial codes, the firm needed to develop specialised ones to analyse the acoustic and thermal characteristics of multi-layered components, including poroelastic materials. Autoneum Gold for the genetic-algorithm optimisation of body dampers and local stiffeners, combined with Nastran FE, is one example. The company also uses its VisualSISAB tool for detailed simulation of insulation and absorption of acoustic 3D components, combined with VA One SEA or Autoneum Revamp SEA for vehicle noise calculations. Autoneum software is mainly for in-house use, however, in several cases it has been sold to OEM customers. Treasuri, for the simulation of acoustic and
structural effects of insulation components in a full vehicle, will be integrated into MSC Nastran in 2012.
www.scientific-computing.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48