Pre kick off meeting 1 1 NCHRP_22_24 January Time Frame CM/E finalised the vehicle and test object modelling documents (Parts 2&3) –These documents should be revised at regular intervals Validation procedures still require research –Further analysis of the Round Robin data in ROBUST is a necessary input –Difficult to predict the delivery of a draft procedure – possibly 2007
Pre kick off meeting 2 2 NCHRP_22_24 January Validation. If we want to use CM for the certification of a system which are the requirements for a “validated” model? Validation should be based on a comparison between test and simulation using: –Severity indices. –Barrier performance: Deformation. Failures. –Vehicle trajectories. –Vehicle time histories (acceleration yaw ratio …) –….. Validation methods MUST be able to validate also tests repetition.
Pre kick off meeting 3 3 NCHRP_22_24 January Validation The problem is how to define an objective procedure to compare time histories. Direct comparison between avcceleration has been already demonstrated to be not applicable. Is not possible to find an objective way to compare such different curves. Other approaches: –Statistical approach (Chalmers software). –Velocity. Chalmers software: –Statistical comparison between a master and a signal to be tested. –The comparison is based on 8 statistical indices.
Pre kick off meeting 4 4 NCHRP_22_24 January CFC 60 3 over 8. We could relax the limits. But with no statistical meaning.
Pre kick off meeting 5 5 NCHRP_22_24 January Filtered at 12.5 hz Always 3 over 8
Pre kick off meeting 6 6 NCHRP_22_24 January Different experimental test Worse condition. Different requirements in the second phase?
Pre kick off meeting 7 7 NCHRP_22_24 January Only the first phase. Always the same problem. Minimum peak fails even if is less important.
Pre kick off meeting 8 8 NCHRP_22_24 January First phase. Only way to pass all the tests. First part of the impact and relaxation of limits.
Pre kick off meeting 9 9 NCHRP_22_24 January Two experimental tests 4 over 8!
Pre kick off meeting NCHRP_22_24 January Other possibility Window obtained from test where simulation must be contained. Not easy to justify a window (delta time, delta g, filtering)
Pre kick off meeting NCHRP_22_24 January Other possibilty 3. Final consideration Compare two velocity time histories (experimental and numerical). If the difference between these two time histories become greather than X(t)% your model is validated until this time. After this time your model is not validated. We are applying this approach also to the deformable barrier and to other impacts
Pre kick off meeting NCHRP_22_24 January Validation. Comparison based on local velocity components. –Reference frame is not inertial. –Accelerations evaluated on a reference system “mounted” on the vehicle. –To proper evaluation of velocity relative mechanics should be used. Comparison based on global velocity components. Planar motion. Need of yaw rotation (not filtered). Evaluation of global components of velocity (interpretation less immediate)
Pre kick off meeting NCHRP_22_24 January Vehicle rotation Vy Global Vy Local x y x y X Y α x y α X X α x y X
Pre kick off meeting NCHRP_22_24 January Round Robin. Rigida barrier h=800 mm. Tb11 –900 kg –20°
Pre kick off meeting NCHRP_22_24 January Round Robin 1. Same new vehicle. Only exp. results
Pre kick off meeting NCHRP_22_24 January Round Robin 2 Different not new vehicles
Pre kick off meeting NCHRP_22_24 January Robust 4.3 –Two more tests on RR1
Pre kick off meeting NCHRP_22_24 January All the tests. Same rigid barrier. Different vehicles. 12 nominally identical.
Pre kick off meeting NCHRP_22_24 January Round Robin. Test and simulation –Not so bad. –Friction influence. (Not understood from acceleration comparison).
Pre kick off meeting NCHRP_22_24 January Rigid barrier results Components –Vy global: good agreement between tests –Vx global: scatter between tests (exit velocity is different) –Vz global: less significative. Seems to work.
Pre kick off meeting NCHRP_22_24 January Deformable barrier. Task 4.1 Robust. N2 barrier (max 1500 kg 110 km/h 20°). Different vehicles. Gorund differences.
Pre kick off meeting NCHRP_22_24 January Yaw angle x y x y X Y α x y α X X α x y X Angle
Pre kick off meeting NCHRP_22_24 January Yaw angle problems during tests
Pre kick off meeting NCHRP_22_24 January Comparison.
Pre kick off meeting NCHRP_22_24 January Comparison 1 Dynamic deflection (ground) –Curve 1 (blue)=0.9m –Curve 3 (red)=0.7m –Difference 25%
Pre kick off meeting NCHRP_22_24 January Tests and simulation
Pre kick off meeting NCHRP_22_24 January Conclusion From Round Robin encouraging results. From deformable barrier results not acceptable. Problem: –With these results (deformable) window to accept simulation is too wide. –According to 1317 these tests are equivalent. Means that a simulation with these differences is validated?