Downloaded from https://www.everyspec.com APPENDIX C MIL-STD-31000A Many CAD systems now have capabilities for checking models in accordance with the recommendations of this handbook. Such tools may be acceptable for internal checking prior to model release. When models are released for analysis, manufacturing, or archival storage, it is preferable to employ validation software from independent, third parties. Third-party software is more likely to identify objective model quality errors than software written by the people who programmed the CAD system that created the models. Validation software may be operated in either interactive or automated modes.
When used interactively, each 3D model is read, visually checked by the operator, and (if applicable) corrected or sent for correction. Automated systems read batches of 3D models, check them, and then create records indicating which, if any, defects are present. The records may be stored as non-graphic attributes of the 3D model files or as records of a PDM database. In either interactive or automated mode, the process for checking 3D models proceeds as follows. First, the operator determines which quality criteria are to be checked and sets the appropriate threshold values in the software. The criteria definitions are in Exhibit A. Recommended threshold values are in Exhibit B.
Not all of the quality criteria listed in Exhibit A will be applicable to every data quality check. For example, if continuous curvature between curves, edges, or surfaces is not a requirement for the parts being checked, then these criteria should not be set. Likewise, for machinery parts, non- tangent surfaces may be acceptable or desirable. Threshold values may deviate from the recommendations of Exhibit B, if experience shows that higher thresholds avoid identification of defects that have no practical significance. Next the software reads each unique part model, checks it for errors, and generates a report. If no defects are found, the model is ready for release.
If defects are found, they may be corrected in one of two ways: with different parameter settings or by interactive modification in the originating CAD software. 1. Models are returned to the person who generated them to be corrected either by re-translation 2. Model defects are healed using automated interactive healing software. The first case, return to sender, is appropriate when parametric model features must be preserved or when defects are of such a gross nature (such as large parts of the model missing) that healing software can’t cope with them. Examples of use cases that require return-to-sender corrections are shown in the diagrams below. 78