Virtual Testing And Research On The Behavior Of A Passenger Car In Quasi-static Roof Crush And Dynamic Rollover

Vehicle rollover crashes are complex events which result in relatively higher fa-tality rate than any other crash scenarios. In the United States, it’s indicated that therewere approximately10,000people killed as occupants involved in light vehicle roll-over crashes each year. This accounted for one third of occupant fatalities in crashes.In China, rollover crashes also critically threaten the occupant safety. It has been re-vealed that roof crush is a predominant factor accounting for fatalities and seriousinjuries of belted occupants in rollover crashes. Therefore the current test proceduresconcerning rollover have been particularly emphasizing on the roof mechanical beha-viors. As a unique compulsory legislation of rollover in the US, FMVSS216incurs alot of arguments due to the poor representative and its quasi-static loading manner.Jordan rollover system (JRS) is considered as a versatile test procedure that can wellrepresent the real world scenarios. It is also meaningful to develop some new me-chanisms to reduce the risks in rollover crashes.A study of double-sided roof crush was carried out by virtual testing in accor-dance with FMVSS216. A Ford Fiesta FE model was refined and validated against theroof crush test. A full factorial DOE (Design of Experiment) was then made based onvarying pitch and roll angles. Thereafter parametric analyses were conducted to in-vestigate the effects of body structure, windscreen, and the dynamic variables of rollangle and pitch angle on roof strength. Subsequently, the response functions con-cerning the resistance force of the double-sided roof were hence constructed with theresults from a series of virtual tests based on9levels of roll angle and4levels ofpitch angle. Finally, the worst loading conditions were evaluated using an optimiztionanalysis. The results show that roof strength is a function of the roll and pitch angles.In general, the roof on the near side performs stronger than the far side and greatlyinfluences the overall collapse behaviour. Within the127mm (5inch) crush, thestrength on both sides of the roof decrease as the roll angle varies from10°to45°.The variation of the pitch angle influences the resistance force. In addition, the con-tribution of the windscreen on the far side roof varies from18%to37%with differentloading angle combinations. From the virtual parametric analysis and optimisationcarried out, it further shows that the recommended stringent loading angle combina-tions in the newly updated FMVSS216roof crush test should be roll angle of45°and pitch angle of10°for both near and far side roof Load application.The dynamic behaviour of vehicle rollover was investigated based on JRS testprocedure. The Ford Fiesta FE model for rollover virtual testing was established andfurther validated against the roof impact test in dynamic condition. The mechanicalbehaviours including the roof intrusions and relevant resistance forces in the rolloverevents were investigated at varying combinations of initial pitch and roll angle. Af-terwards, the behaviour discrepancies on the roof were evaluated while the vehiclesubjected to quasi-static and dynamic test with fixed angle combinations. Then para-metric studies were carried out to assess the effect of JRS test setup involving dropheight, roll rate of the vehicle as well as the velocity of the moving roadbed. The re-sults reveal that the intrusion level on the far side is extremely greater than that on thenear side in the same test. As the initial pitch or roll angle rises, the intrusion on thefar side tremendously goes up, while it drops down sharply on the near side. Form thecomparisons between the quasi-static and dynamic test, it could be revealed that thecrush distance in FMVSS216is overestimated; in addition, the contribution of thedynamic inertia effect to the resistance forces on the far side varies from17%~72%.The initial drop height extensively results in an additional intrusion amount and in-trusion velocity by210%and92%on the near side respectively, whereas there is nosignificant difference to be found on the far side. With the varying velocity of movingroadbed, the maximum discrepancies are correspondingly31%and63%in terms ofroof intrusion and intrusion speed on the far side, while trivial differences could benoticed on the near side.Finite element models of road surface were established with regard to differenttypes of materials including bitumen, concrete as well as sandy soil in LS-DYNA. Asensitive study on the roof structural behaviour subjected to JRS test was then carriedout with different road types. Thereafter the effect of varying stiffness and the frictionfactors on the roof behaviour were conducted. The results indicate that roof intrusionis a function of the friction factor as shown in logistical regression equations; whilethe stiffness could hardly make any significant effect. With different stiffness of theroad surface, the discrepancies in intrusion amplitude and intruding speed vary from2.2%~8.5%and0.8%~4.8%correspondingly.Two approaches were proposed to enhance the crashworthiness of the vehicle inrollover crashes in this thesis. One is a strengthen roof and another an egg-box styleroof panel. A finite element model of egg-box panel was established and validatedagainst the relevant quasi-static crush tests which particularly focus on the collapse behavior of the egg-box. The strengthen roof and the egg-box panels were then re-spectively employed in the JRS test. As depicted in the results, the intrusion on the farside could be reduced by31.4%and33%respectively with the strength roof and theegg-box roof panel; while the intruding speed could be decreases by21.8%and26.6%correspondingly with the strength roof and the egg-box roof panel. Ultimately, theroof that incorporates the two approaches can considerably reduce the far side intru-sion amplitude and intrusion speed by57%and37.5%respectively.