Appendix 1 Crack growth analysis and tests

Crack growth characteristics should be determined for each detail design point identified in accordance with 7(f) above. This information, when combined with the results from the residual strength analyses and tests, will be the basis for establishing the inspection requirements as discussed in Section 8. Crack growth characteristics can be determined by analysis or test. However, due to the large number of detail design points that are typically evaluated, and the practical limitations involved with testing, analyses are generally relied on to determine crack growth at the detail design point.

(a)Analyses. In order to perform a crack-growth analysis a number of key elements are needed. These include:

(1)a load/stress spectrum applicable to the detail design point;

(2)an initial crack size and shape to be assumed;

(3)a cracking scenario to be followed;

(4)applicable stress intensity solution(s);

(5)a crack growth algorithm; and

(6)material crack growth rate properties. A loading spectrum must be developed for each detail design point. It is derived from the overall aircraft usage spectrum that is discussed in paragraph 6(b). The spectra at each detail design point may be modified for various reasons. The most common modification for metallic structure involves the deletion of high infrequent loads that may have an unrepresentative beneficial effect on crack growth if retardation is considered. Also, local load events that are not part of the overall aircraft spectrum should be included (e.g. flutter damper loads during pre-flight control surface checks). The initial crack size and shape and subsequent cracking scenario to be followed are problem-dependent. Applicable stress intensity solutions may be available in the public domain or may need to be developed. Many references exist which provide technical guidance for the application and development of stress intensity solutions. Care should be taken to ensure that the reference stress used for the spectrum load and stress intensity solution are compatible. Crack-growth algorithms used in predicting crack extension range from simple linear models to complex ones that can account for crack growth retardation and acceleration. It is generally accepted that the use of a linear model will result in conservative results. A non-linear model, on the other hand, can be conservative or non-conservative and generally requires a higher level of validation and analysis/test correlation to adequately validate the accuracy of the algorithm. Coupon testing should be performed using representative materials and spectra types (e.g. wing lower cover, pylon support lug, horizontal-stabiliser upper cover) that will be encountered in the course of the overall aircraft crack-growth evaluation. Crack growth rate data (e.g. da/dN vs ∆K vs R, da/dN vs ∆Keff) for many common aerospace materials is available in the public domain. Additionally, testing standards (e.g. ASTM) exist for performing tests to gather this data. The generally accepted practice is to use typical or average representation of this data for performing crack growth evaluations.

(b)Tests. Crack-growth testing using coupons is typically performed to generate crack growth rate data and to validate crack growth algorithms used for analyses. Simple specimens are generally used that have well-established stress intensity solutions for the characteristic cracking that can be expected. The primary issue for these tests is the pre-cracking required to achieve a well-behaved fatigue crack before data is collected. Effective pre-cracking procedures (e.g. ‘load shedding’) have been established and are described in the public domain. Care must be taken to ensure that subsequent crack growth is not affected by the prior pre-cracking. In order to minimise the test time for actual structural components and/or full-scale test articles, the test loading spectrum may be modified by eliminating small magnitude load events or by replacing them with a fewer number of larger load events that give equivalent crack growth. Crack-growth behaviour may be obtained from actual structural components and/or full-scale test articles. However, inducing active fatigue cracks of the desired initial size and at the desired locations can be extremely difficult. Past success in obtaining useful data has been achieved on an opportunistic basis when natural fatigue cracks have developed in the course of normal cyclic testing. Naturally occurring and artificially induced fatigue cracks may be monitored and data collected for at least a portion of the overall crack-growth period to be used for setting inspection requirements. This data can be extremely useful in supplementing and validating the analytical predictions, in some cases it may be the sole basis for the establishment of inspection requirements. Where fatigue test crack growth data is used, the results should be corrected to address expected operational conditions.

[Amdt 25/19]