Model reduction of finite element models applied to nonlinear squeak and rattle simulation
| dc.contributor.author | Godborg, Filip | |
| dc.contributor.author | Valiyakath Basheer, Anoob | |
| dc.contributor.department | Chalmers tekniska högskola / Institutionen för mekanik och maritima vetenskaper | sv |
| dc.contributor.examiner | Abrahamsson, Thomas | |
| dc.contributor.supervisor | Bayani, Mohsen | |
| dc.contributor.supervisor | Gibanica, Mladen | |
| dc.contributor.supervisor | Abrahamsson, Thomas | |
| dc.date.accessioned | 2019-09-18T12:50:11Z | |
| dc.date.available | 2019-09-18T12:50:11Z | |
| dc.date.issued | 2019 | sv |
| dc.date.submitted | 2019 | |
| dc.description.abstract | High level of modelling details in finite element models contributes to an increase in computational cost. To deal with this, the Component Mode Synthesis can be used, which combines the principle of dynamic substructuring with techniques of model reduction. The most common Component Mode Synthesis method is the Craig-Bampton method that has been labeled as the standard method for dynamic substructuring of linear systems. However, for dynamic substructuring of nonlinear systems, no standard method has emerged as the research within the area is still at an early stage. Hence, it would be beneficial to investigate if the Craig-Bampton method can be used for squeak and rattle simulations including nonlinearities in the form of contact. This thesis aims to increase the understanding of how the Craig-Bampton method can be used for simulations including nonlinearities in the form of contact. The study is carried out on a passenger car door which is excited using three different load cases, to see the influence of different levels of contact. The substructured model is compared to the original model by using the Modal Assurance Criterion, frequency response comparison and by comparing contact force with respect to time and amplitude. Further, by gradually excluding more elements from the substructure, the thesis aims to see if the accuracy of squeak and rattle prediction can be improved by having a higher number of elements closer to the contact region. The results show that the substructured system had 97.9 % lower computational time than the original FE-model. Although it was accurate at predicting contact with respect to time, the calculated contact force amplitudes were generally higher. Neither the level of contact nor the number of elements close to the contact region had any significant effect for squeak and rattle prediction of the substructured model. The computational time is found to be dominated by the number of interface DOFs. In this case, an additional 32.5 % reduction in computational time was achieved by locally coarsening the mesh around the contact interface. | sv |
| dc.identifier.coursecode | MMSX30 | sv |
| dc.identifier.uri | https://hdl.handle.net/20.500.12380/300325 | |
| dc.language.iso | eng | sv |
| dc.relation.ispartofseries | 2019:21 | sv |
| dc.setspec.uppsok | Technology | |
| dc.subject | Component Mode Synthesis | sv |
| dc.subject | Craig-Bampton | sv |
| dc.subject | model reduction | sv |
| dc.subject | Guyan reduction | sv |
| dc.subject | nonlinearity | sv |
| dc.subject | Modal Assurance Criterion | sv |
| dc.title | Model reduction of finite element models applied to nonlinear squeak and rattle simulation | sv |
| dc.type.degree | Examensarbete för masterexamen | sv |
| dc.type.uppsok | H | |
| local.programme | Applied mechanics (MPAME), MSc |