Directory Layout

The following model directory layout is recommended:

SISAME Input File Naming

SISAME Input Files

SISAMEM Input Files

SISAME Input File Usage

Path Names

For cross-platform input files use relative path names: the applications can translate relative paths to the host platform format automatically.

Standard Model ID Codes and File Extensions

Model Documentation

Create and maintain documentation files for each vehicle and each event as per those in the SISAME Model Library.

Vehicle Instrument Signal Preparation

The NHTSA Tools applications such as UDSMod, UDSTool, and Scale can be used to prepare UDS vehicle signal files for SISAME modeling use. For example:

• Set file name extensions based on standard component ID codes

• Clear nonblank SENLOC and OCCTYP fields

• Change AXIS from XL to XG

• Check/fix forces mis-scaled by 10³ or 10⁶ or -1 (barrier contact forces should be positive)

The instrument signals should be visually examined:

• Instrument failures: eliminate signals with large nonsystematic errors

• Acceleration/force signal tail bias

• Acceleration signal first and second integrals

Film Data Collection/Preparation

If test film analysis is available, collect displacements for the full instrument signal time span for the:

• Occupant Compartment

• Engine - Top

• Front Wheel(s) (both for offset tests)

• Vehicle contact interface, if possible

Convert the film displacements to UDS files using Film2UDS or another procedure using UDSMod. An instrument signal UDS file can be used to set the baseline specifications and then modified as follows:

• Set DEL to the film time step value

• Set SENATT to the film target location

• Set INIVEL and CLSSPD

• Set AXIS=XG

• Check for film time step and displacement scaling errors (look at velocity before t=0)

Name the film displacement UDS files using the form v1234daV.Ext where 1234 stands for the test number, V is the vehicle model letter code, and Ext is the standard file extension for the target component.

Collect film displacement target values for use in VeCor at 100, 200, and 300 ms (but at least 10 ms before final tail to allow for time shifting). A 30 Hz zero-phase-shift filter is recommended prior to sampling film data.

Modeling

General

• Use an output time step of at most DelTOut=.0001 sec

• Use relative paths to input files for cross-platform portability

Weight Extraction

• Run weight extraction if VTB test with barrier force or VTV test

• Specify known vehicle weight(s)

• First run with no weight estimates/bounds to assess inertial completeness of the signal set

• Then run using weight estimate/bound information: if fit degrades significantly try looser estimates/bounds

Model Extraction

• First run with all possible load-paths present to get best-case fit

• Start with StaType=SI for deforming load-paths with unknown behavior and change those with simple extracted behavior to simpler types

• Use Symmetric=True load-paths when the primary deflection direction may vary between matching/referenced structures in single or multiple events

• Enforce known structural symmetries using references

Dynamic extraction

Static and dynamic contributions may not be clearly distinguishable from the instrument data alone, particularly in single-event extractions, so dynamic parameter estimate/bound specifiers are recommended. The dynamic parameter specifiers shown below provide a good starting point.

Single Event Extractions

Metric

DynType=LM  MSlp=?( ~0[2.5] )

DynType=AM  MSlp=?( ~0[5] )  MMax=?( ~1[100] )

English

DynType=LM  MSlp=?( ~0[4] )

DynType=AM  MSlp=?( ~0[8] )  MMax=?( ~1[100] )

Multiple Event Extractions

Metric

DynType=LM  MSlp=?( ~0[5] )

DynType=AM  MSlp=?( ~0[10] )  MMax=?( ~1[200] )

English

DynType=LM  MSlp=?( ~0[8] )

DynType=AM  MSlp=?( ~0[16] )  MMax=?( ~1[200] )

Refining Dynamic Specifiers

• LM magnifiers provide a simpler and more efficient extraction process

• AM magnifiers can represent a broader range of behavior but at some cost in efficiency

• Adjust AM specifiers to avoid both MSlp and MMax having sufficiently large values to give unrealistic magnification

• AM magnifiers with sufficiently large (MMax-1)/MSlp ratios can be replaced by more efficient LM magnifiers