AUTOMATIC ITERATION TO IMPROVE EVIDENCE MATCH AND SIMPLIFY THE MSMAC3D PROCESS
As any SMAC user is aware, many iterations of the SMAC program may be required to go from an initial approximation to an acceptable match of the measured trajectory and damage targets. Throughout the iterative process, the impact speeds and impact speed change results may not change significantly. Also, what constitutes an acceptable match can vary widely among users. Sometimes the focus is on a detailed match of the positions of rest; sometimes the focus is on a match of damage locations and extents on the vehicles. There is currently no standardized measure of the correlation of simulation results with the accident evidence.
Since the initial development of the SMAC program, there has existed a need to simplify the application process. The ultimate simplification would entail an automatic iteration procedure. Optimization techniques applied to the evidence match can serve to reduce the required time and effort and, also, can achieve greater uniformity of the evidence interpretation.
The working hypothesis of the automatic iteration of SMAC is that a unique set of impact conditions is required to achieve an acceptable match of all of the documented evidence (both damage and trajectory). The use of quantitative measures of the overall “fit” to the documented evidence and applications to experimental crash tests provided a means of proving the hypothesis, as well as demonstrating reconstruction accuracy and convergence rates. Note that successful efforts on automatic iteration in the msmac2D case were reported in Ref 13 (Fig 20 -23). That research proved the feasibility of an automated procedure for achieving a “best match” of measured evidence, starting with CRASH-type linear momentum/damage analysis initial approximations. Small deviations from a perfect “match” are, of course, imposed by the existing limitations of the SMAC computer program and by any inaccuracies in the reported evidence.
In the reported results the automatic iteration/optimization of msmac2D was demonstrated to successfully converge toward evidence matches in a variety of impact configurations. With measured evidence from full-scale tests on flat surfaces the automatic iteration procedure of msmac2D produced correlation of impact speeds and impact speed changes with deviations in individual results of less than approximately ± 10%.
The research confirmed that the initial approximation of the CRASH program, or any proper combination of linear momentum and damage analysis solution procedures, should provide an adequate initial approximation from which an automatic iterative/optimization procedure of msmac2D can be used to determine the impact speeds and impact speed changes within ± 10%.
For 3D simulations the only change required was to add automatic iteration/optimization of msmac3D results is to simply have the control algorithms use the msmac3D program instead of msmac2D program for iterating/optimizing towards a unique solution. The msmac3D automatic iteration/optimization program is currently in testing and will be included in the msmac3D release.
The msmac3D setup automatically generates all required 3D vehicle input definitions based on the vehicle make/model and any other supplemental information and the user can easily include 3D terrain information if applicable. Upcoming extensions of the automatic iteration/optimization routine will be to include comparisons with measured tire mark information and an extension to iterate/optimize on rollover and other single vehicle accidents.
Automatic Iteration solutions of a SMAC Reconstruction of a sample RICSAC Test.
Automatic Iteration Results with Crash Test results on Impact Velocity (first) and Impact Speed Change (second).
Comparison of msmac automatic iteration correlation factor “score” with maximum error.
