need to be evaluated at all appropriate levels of the building block tests to determine whether the manufacturing process change yields identical or alternate materials. Engineering judgment will be needed in determining the extent of testing based on the proposed manufacturing change.
f. Case A (alternative material) should always be considered as an important change, which requires structural substantiation. It is not recommended to try a sub-classification according to the basic constituents being changed, as material behavior (e.g., sensitivity to stress concentrations) may be governed by interfacial properties, which may be affected either by a fiber or a resin change.
7. Substantiation Method. Only the technical aspects of substantiation are addressed below.
a. Compliance Philosophy. Substantiation should be based on a comparability study between the structural performances of the material accepted for type certification, and the second material. Whatever the modification proposed for a certificated item, the revised margins of safety should remain adequate. Any reduction in the previously demonstrated margin should be investigated in detail.
(1) Alternative Material/Process: New design values for all relevant properties should be determined for any alternate material/process combination. Analytical models initially used to certify structure, including failure prediction models, should be reviewed and, if necessary, substantiated by tests. The procurement specifi cation should be modified (or a new specification suited to the selected material should be defined) to ensure key quality variations are adequately controlled and new acceptance criteria defined. For example, changing from first to second generation of carbon fibers may improve tensile strength properties by more than 20% and a new acceptability threshold will be needed in the specification of the alternate material to ensure the detection of quality variations.
(2) Identical Material: Data should be provided that demonstrates that the original design values (whatever the level of investigation, material or design) remain valid. Statistical methods need to be employed for data to ensure that key design properties come from the same populations as the original material/process combination. Calculation models including failure prediction should remain the same. The technical content of the procurement specification (Case B) should not need to be changed to properly control quality.
b. Testing.
(1) The extent of testing needed to substantiate a material change should address the inherent structural behavior of the composite and will be a function of the airworthiness significance of the part and the material change definition. For example, the investigation level might be restricted to the generic specimens at the test pyramid base (refer to figures in paragraph 7) for an identical material, but non-generic test articles from higher up the pyramid should be included for an alternative material. Care needs to be taken to ensure that the test methods used yield data compatible with data used to determine properties of the original structure.
Page A3-3
(2) The testing that may be required for a range of possible material and/or process changes should consider all levels of structural substantiation that may be affected. In some instances (e.g., a minor cure cycle change), possible consequences can be assessed by tests on generic specimens only. For other changes, like those involving tooling (e.g., from a full bag process to thermo-expansive cores), the assessment should include an evaluation of the component itself (sometimes called the “tool proof test”). In this case, an expanded NDI procedure should be required for the first items to be produced. This should be supplemented – if deemed necessary – by “cut up” specimens from a representative component, for physical or mechanical investigations.
c. Number of Batches.
(1) The purpose for testing a number of batches is the demonstration of an acceptable reproducibility of material characteristics. The number of batches required should take into account: material classification (identical or alternative), the investigation level (non-generic or generic specimen) the source of supply, and the property under investigation. Care should be taken to investigate the variation of both basic material and the manufacturing process.
(2) Existing references (e.g., The Composite Materials Handbook (CMH-17) Volumes 1 and 3, FAA Technical Report DOT/ FAA/AR-03/19), addressing composite qualification and equivalence and the building block approach, provide more detailed guidance regarding batch and test numbers and the appropriate statistical analysis up to laminate level. Changes at higher pyramid levels, or those associated with other material forms, e.g., braided VARTM (Vacuum- Assisted Resin Transfer Molding) structure, may require use of other statistical procedures or engineering methods.
d. Pass/Fail Criteria. Target pass/fail criteria should be established as part of the test program. For strength considerations for instance, a statistical analysis of test data should demonstrate that new design values derived for the second material provide an adequate margin of safety. Therefore, provision should be made for a sufficient number of test specimens to allow for such analysis. At the non-generic level, when only one test article is used to assess a structural feature, the pass criteria should be a result acceptable with respect to design ultimate loads. In the cases where test results show lower margins of safety, certification documentation will need to be revised.
e. Other Considerations. For characteristics other than static strength (all those listed in AC 20-107B, paragraphs 8, 9, 10, and 11), the substantiation should also ensure an equivalent level of safety.
Page A3-4