AC 29-2C, Chg 4 12/2/2011 and steady and vibratory load levels in combination with flaws representative of those identified in the threat assessment.
(iii) The effect of the flaws is combined with the fatigue result analytically determined for the as-manufactured structure without flaws.
(C) Working Curve Determination. Reduction factors should be applied to the mean curve determined above to derive a working fatigue curve. As outlined in AC 27-1B MG 11, working curve reduction factors should include consideration of the number of specimens tested, variability (scatter), previous test data on the same materials or similar structures, as well as service experience. Different reduction factors from those used for conventional Safe-Life methodology may be employed if appropriately justified.
(D) Retirement Time Determination. The working fatigue curve, flight loads (paragraph f.(3)), and usage spectrum (paragraph f.(4)) are used with a cumulative damage analysis such as shown in AC 27-1B MG 11, to calculate a safe retirement time.
(iii) Safe-Life Retirements with Clearly Detectable Flaws.
(A) A retirement time may also be based on flaws larger than the BDF case, up to the clearly detectable size described in paragraph f.(5), if the applicant chooses. This could be the case, for example, if it was desired to allow a specific manufacturing-related flaw of detectable size to remain in place for the life of the structure without further inspection.
(B) The substantiation for this case can be the same as described in paragraph f.(7)(ii), except that the larger flaws selected for the replacement time substantiation are used instead of the BDFs.
(iv) Crack Growth Retirement.
(A) General.
(1) This approach depends on retirement rather than inspection to ensure the continued airworthiness of a PSE. The retirement time is established based on consideration of crack growth characteristics. Fatigue with damage is addressed by timely retirement and there are no explicit inspection requirements that are derived from this approach.
(2) This approach requires demonstration either by analysis, testing, or both, that the BEC (aBEC), the most severe crack consistent with manufacturing, maintenance, and service environment, will not grow or will not grow to critical size (aCRIT) under the service loading and environment before the structure is retired. The critical crack size (aCRIT) is established by limit load. The crack should be assumed at Page C - 118
the critical location, as defined by the largest stress intensity factor range under the expected service loading range including the ground–air–ground cycle. It is recommended that full scale fatigue testing be undertaken to provide an understanding of the fatigue behavior of the component in support of the chosen methodology. In particular it ensures hot spots are identified, which experience has shown analysis often fails to identify.
(3) A threat assessment (see paragraph f.(5)) should be performed to support establishing the BEC size to be used. It is intended that the BEC conservatively bounds the most severe defect resulting from manufacturing, maintenance, or the service environment. That is, there should be no probable defect, from any source, that would lead to failure of the part in less time than it would take the BEC to reach critical size. It should be noted that the resulting crack is a mathematical expedient that may not represent a true physical crack. If the BEC is defined by analytical back calculations from coupon or service fatigue life data, it will be highly dependent on the predictive tool used (i.e., growth algorithm, material data, etc.).
Therefore, the same predictive tool must be used to perform the fatigue tolerance evaluation. When the BEC is based upon test or service data, it must account for material variability in initiation and growth.
(4) To determine the retirement, the BEC should be assumed at the critical location and the crack growth characteristics should be determined for the expected load and environment spectrum. There are three different scenarios that could result from a crack growth assessment and be used for establishing a retirement time. These scenarios are illustrated in Figure AC 29.571B-2, Figure AC 29.571B-3, and Figure AC 29.571B-4.
(B) No Growth. The no crack growth scenario is illustrated in Figure AC 29.571B-2. Here the BEC does not grow when using top- of-scatter crack growth rate data. In this case the retirement time should not exceed the design service life (LDES).
Page C - 119
AC 29-2C, Chg 4 12/2/2011 CRA CK S IZ E a R
L TIME DES Figure AC 29.571B-2: No Growth (C) Slow Growth of Undetectable Crack. Figure AC 29.571B -3 illustrates the scenario where the BEC grows relatively slowly but becomes critical prior to becoming detectable (aDET). In this case, the retirement time should be set equal to the total crack growth life (LT) divided by a factor N.
CRA CK S IZ E a DE T L T a CRIT a R TIME Figure AC 29.571B-3: Slow Growth of Undetectable Crack Page C - 120
(D) Slow Growth of Detectable Crack. Figure AC 29.571B -4 illustrates the scenario where the BEC grows to a detectable size (at L1) before becoming critical (at L1+L2). In this case, the retirement time should be set equal to the total crack growth life (L1+L2) divided by a factor N.
CRA CK S IZ E a CRIT a DET a R L 1 L 2 TIME Figure AC 29.571B-4: Slow Growth of Detectable Crack (E) Life Factors for Crack Growth Retirement.
(1) In determining the factor of N to be used for determining the retirement time, consideration should be given to the crack growth data used (e.g., top of scatter data versus average data, number of specimens used to generate data, etc.).
(2) The minimum suggested N value should be N=2 in the case where the conservative top-of-scatter crack growth data are used in the crack growth analysis, or N=4 when the average crack growth data are used in the crack growth analysis, or N=4 when the crack growth life is obtained from the crack growth test of one specimen (for two or more full scale specimens, N=3 of the shortest crack growth life can be used).
(3) It should also be noted that with this approach, the validity of the crack growth threshold, Δ Kth, is especially important since there is no element of inspection to ensure continued airworthiness. Consistent with this, additional attention may be required for validating the crack growth threshold value(s) used in the analyses.
Consideration should be given to the influence of the test procedure used to develop values, microstructure, heat treatment, crack size, loading conditions, environment, grain size and orientation, etc. In general, a coupon- testing program may be necessary Page C - 121
AC 29-2C, Chg 4 12/2/2011 to develop a consistent ΔKth database and the use of published data may require additional conservatism.
(8) Inspection Intervals. Each of the following three methods provides a means to establish inspection intervals for detectable damage or detectable damage growth.
The time of the first inspection should coincide with the repetitive interval established unless the applicant can substantiate an alternate time.
(i) Safe-Life Inspection for a CDF provides a safe interval of operation between repetitive inspections for the presence of probable detectable flaws. The substantiation is accomplished by testing and analysis employing conventional Safe-Life methodology except that intrinsic and discrete critical flaws are considered.
The size of flaws considered should be “clearly detectable”, which is intended to be a conservative representation of detectable flaws that could remain in place for the entire interval in spite of routine inspections for general condition. It should be noted that this method, since it is a Safe-Life (crack initiation) method, is not appropriate for use when the flaw being considered is already a crack.
(A) The method described in paragraph f.(7)(iii), Safe-Life Retirements with Clearly Detectable Flaws, may be employed for this case, except that the calculated retirement time is used as a repetitive inspection interval.
(B) The repetitive inspection consists of examination of the structure for the presence of the flaw using the substantiated inspection method. If no flaw is found, the structure may be returned to service for another inspection interval period, up to the established retirement time. If the flaw is found, the structure is retired; or, if a repair procedure for the specific flaw type has been substantiated, the structure is repaired and returned to service for another inspection interval period, up to the established retirement time for the structure.
(C) Substantiation of repairs should include careful consideration as to whether undetectable cracks may now exist and whether the original certification approach is still applicable.
(ii) Crack Growth Inspection. This approach depends on detection of cracks before they become critical to ensure the continued airworthiness of a PSE. While any inspections that are capable of detecting cracks with high reliability may be used with this approach, the criteria stated in paragraph f.(6), Inspectability and Inspection Methods, should be considered in making the selection. The inspection method chosen will define the initial inspectable crack that will be used to perform the fatigue tolerance evaluation. Once the initial inspectable crack is defined, crack growth, and residual strength assessments must be performed to determine the time for the initial inspectable crack (aDET) to grow to a size (aCRIT) that would result in a catastrophic failure of the rotorcraft if limit loads were applied. This assessment could be theoretically done analytically or by test; however, in most cases it is performed analytically using fracture mechanics methods. The resulting life for a DET to grow to Page C - 122
aCRIT is used to set the inspection interval. This general process applies to both single and multiple load path structure regardless of the level of inspection (e.g., for complete load path failure or less than load path failure in a multiple load path structure). The details of defining the interval once the crack growth life has been determined are discussed later.
(A) Single Load Path Structure. The time for a detectable crack (a DET) to grow to critical size (aCRIT) in a structure is denoted as L2 in Figure AC 29.571B-4. If this were a single load path structure, the inspection interval would be established as L2 divided by N. (See paragraph f.(8)(ii)(C) for guidance on values of N.) This interval is valid until the part is retired.
(B) Multiple Element Structure.
(1) Depending on inspectability considerations and residual life characteristics of the structure following a load path failure, it may be beneficial to take advantage of the redundancy of a multiple load path structure. On the other hand, the safety of a multiple load path structure can be managed without taking advantage of its redundancy. In this case, each load path would be considered independently and inspection intervals established for each load path consistent with paragraph f.(8)(ii)(A).
This may be necessary for similarly stressed load paths when damage according to the threat assessment could occur in each element at the same time.
(2) When considering redundancy in a multiple load path structure, two scenarios might be possible; one where the required inspection is for a completely failed load path and one where the inspection is for less than a load path failure. In either case, the remaining life of the secondary load path after primary load path failure is used to determine the inspection interval. Consistent with this, the resulting intervals are only valid until the cumulative fatigue damage or crack growth in the intact structure is taken into account. This issue is illustrated in a crack growth context in Figure AC 29.571B-5. Crack growth in the secondary load path from an initial crack as detailed in paragraph f.(8)(ii)(B)(3)(i) will proceed along curve A-B as long as the primary load path remains intact and load redistribution is negligible. However, at the time of primary load path failure, loading on the secondary load path will increase due to load redistribution and crack growth will be accelerated (e.g., subsequent growth from point 1, 2, or 3 depending on if the failure occurs at time t 1, t2 or t3). Note that the residual life, Lr, in the secondary load path is inversely proportional to the time at which primary load path failure occurs. This should be considered whenever L r is used in establishing repeat inspection intervals.
Page C - 123
........ ---·-· _ ..... -----·-·-· .... -•-· ·----------------------·-------------------------- _____ .... ---· ------- ----- )t_ --- ---------- .... -·
.---
_________ .,.
AC 29-2C, Chg 4 12/2/2011 A B 1 2 3 Lr1 Lr2 Lr3
CRACK SIZE Primary Load Path Crack Growth Primary Load Path Failure Secondary Load Path Crack Growth - Intact Structure t1 t2 t3
CRACK GROWTH LIFE Figure AC 29.571B-5: Decreasing Residual Life in Secondary Load Path for Multiple Element Crack Growth with Inspections.
(3) Inspect for Load Path Failure. If a failed load path is easily detectable and the residual life and strength of the remaining structure is sufficient, this approach may be optimum. Analysis or tests as described in the following paragraphs can determine the inspection interval.
(i) Evaluation by analysis. Figure AC 29.571B -6 illustrates an example of multiple load path structure for which a completely failed load path is easily detectable. The inspection interval is based on the life of the secondary load path (L r) after primary load path failure at time NF. Consistent with this, damage accumulated in the secondary load path prior to primary load path failure must be accounted for in the analysis. In order to do this within the context of a crack growth analysis, it is necessary to assume some initial crack, of size ai, exists in the secondary load path at time zero.
This initial crack size should be representative of a normal manufacturing quality unless the threat assessment indicates that larger damage could exist. Crack growth accumulated prior to a load path failure is accounted for by calculating the amount of growth, (Δai), between time zero and NF. Load redistribution that may occur prior to N F should be considered. The residual life, (L r), then becomes the time for a crack of size ai + Δai to grow to critical size, assuming a complete load path failure has occurred (i.e., “failed” condition loads used). It should be noted that the assumed time of load path failure would also represent an upper limit of validity for any repeat inspection period based on Lr. It is therefore recommended that N F be assumed equal to the retirement Page C - 124
..
Primacy Load Failure
CRACK GROWIB ~ Secondacy Path Failure Limit 12/2/2011 AC 29-2C, Chg 4 time for the structure being inspected or the rotorcraft design life if the structure has no declared retirement time. Based on the above, (A) Inspection Interval = Lr/N [For N refer to paragraph
f.(8)(ii)(C)].
(B) Limit of validity = NF (i.e., repetitive inspection time would not be valid for operation beyond NF).
Figure AC 29.571B-6: Multiple Load Path Structure Analytical Evaluation to Support Inspection for a Failed Load Path.
(ii) Evaluation by Test. Figure AC 29.571B -7 illustrates some key points if an inspection for a complete load path failure is to be developed based on testing. The inspection interval is based on the test demonstrated residual life (L r) subsequent to load path failure. Because the residual life decreases with the time accumulated prior to a load path failure, there will be a limit of validity to the L r and it will be dependent on the time at which a load path failure is simulated, (N D).
(A) The test article should consist of as-manufactured production parts. Representative “well” condition loading should be applied for some Page C - 125
AC 29-2C, Chg 4 12/2/2011 predetermined period of time, (ND). It is recommended that the “well” condition loading be of sufficient duration so that ND/LSF is not less than the retirement time minus one inspection interval for the structure being inspected or the rotorcraft design life if the structure has no declared retirement time. At the end of this period, the load path that is to be inspected for complete failure should be disabled (e.g., saw cutting, attachment(s) removal, member removal) to simulate its failure. The test should then be restarted with a representative “failed” condition loading. (Note that the external loads may be the same as for the “well” condition if the member failure simulation results in the correct “failed” condition internal load redistribution.) The test should continue until the desired residual life has been achieved or to the time at which the secondary load path can no longer support limit loads without failure, whichever is less, (N0).
(B) In developing the test spectrum, consideration should be given to proper use of representative loads, truncation of non-damaging loads, inclusion of ground-air-ground cycles, clipping of high magnitude loads, and load sequence.
(C) Based on the above,
(a) Demonstrated residual life = Lr = N0-ND.
(b) Repetitive inspection time = Lr/N [For N refer to paragraph f.(8)(ii)(C)].
(c) Limit of validity = ND/LSF.
(d) LSF = 2, Life safety factor.
Page C - 126
Primary Load Path Intentionally Disabled ~ u ''Failed" Condition Loading < ~ U "Well" Condition --------------------------.
~ Loading IN TEST Secondary Load Path Not Capable of Limit Load 12/2/2011 AC 29-2C, Chg 4 Figure AC 29.571B-7: Multiple Load Path Structure Evaluation by Test to Support Inspection for a Failed Load Path.
(4) Inspect for Less Than a Load Path Failure. Inspection for less than a load path failure may require special non-destructive Inspection (NDI) procedures but will result in longer inspection intervals. Figure AC 29.571B -8 illustrates how inspection intervals could be established on the basis of crack growth and residual strength evaluation.
(i) In this case, the inspection interval is based on the life of the secondary load path (Lr) subsequent to primary load path failure at NF plus the time (LP) for a detectable crack (aDET) in the primary load path to grow to critical size under in-service loads. The determination of L r is the same as discussed in paragraph
f.(8)(ii)(B)(3)(i).
(ii) Based on the above, (A) Repetitive Inspection = (LP + Lr)/N [For N refer to paragraph
f.(8)(ii)(C)].
(B) Limit of validity = NF.
Page C - 127
AC 29-2C, Chg 4 12/2/2011 NF Lr
CRACK SIZE ai aDET LP Crack Growth in Primary Load Path Crack Growth in Secondary Load Path after Primary Load Path Failure Failure of Secondary Load Path at Limit Load Failure of Primary Load Path Under Inservice Loads Crack Growth in Secondary Load Path with Primary Load Path Intact Dai
CRACK GROWTH LIFE Figure AC 29.571B-8: Multiple Load Path Structure Analytical Evaluation to Support Inspection for Less than a Failed Load Path.
(C) Safety Factors.
(1) In determining the factor of N to be used for determining the inspection time, consideration should be given to the crack growth data used (e.g., top of scatter data versus average data, number of specimens used to generate data, etc.)
and the capability of the inspection procedure.
(2) The minimum suggested N value should be N=2 in the case where the conservative top-of-scatter crack growth data are used in the crack growth analysis, or N=4 when the average crack growth data are used in the crack growth analysis, or when the crack growth life is obtained from the crack growth test of one specimen (for two or more full scale specimens, N=3 of the shortest crack growth life can be used).
(iii) Safe-Life Inspection for a Failed Element.
(A) A Safe-Life Inspection substantiation for a Failed Load Path provides a safe interval of operation between repetitive inspections for the failed load path. The substantiation is accomplished by testing and analysis employing conventional Safe-Life methodology except that the configuration of the structure substantiated is with the Page C - 128
critical load path inoperative and appropriate flaws imposed on the remainder of the structure, as determined by the threat assessment.
(B) The method described in paragraph f.(8)(i) can be employed for this case with the following differences:
(1) The principal “flaw” considered is failure or loss of the most critical load path. The load path failure can be the result of fatigue cracking, static failure, or a fractured or missing fastener, as determined by the threat assessment, paragraph f.(5).
(2) The remainder of the structure may be representative of normal manufacturing quality unless the threat assessment indicates that larger damage should exist.
(3) The mean strength for the substantiation should be based on the number of cycles from the first load path failure to the first initiation of cracking at any other point in the remaining structure. Any applied load changes or load distribution changes that occur as a consequence of the load path failure should also be included (bending due to increased deflection, for example).
(4) When the remaining structure may have some pre-existing fatigue damage at the time the first load path fails (due to both load paths being highly loaded, for example), this should be factored into the analysis.
(5) The remaining structure after first load path failure must be shown to have limit load capability, considered as the ultimate loading, except in some cases where no retirement life is provided and fatigue damage is expected (see paragraph
f.(10).
(6) The inspection conducted is for the failed or missing load path.
(9) Retirement Time and Inspection Interval Schedules.
(i) Based on the evaluations required by § 29.571, inspections, retirement times, combinations thereof, or other procedures have been established as necessary to avoid catastrophic failure. These inspections, retirement times, or approved equivalent means must be included in the Airworthiness Limitations Section (ALS) of the Instructions for Continued Airworthiness (ICA) as required by § 29.1529 and Appendix A29.4 of the regulatory requirements. These inspections, retirement times, or a combination of both are normally stated in hours time-in-service, but may be stated in other terms, such as engine starts, landings, external lifts, etc.
(ii) The design service life should be specified in the fatigue evaluation methodology that must be approved by the FAA. In any case, routine inspections for wear, fretting, corrosion, cracking, and service damage are appropriate. These routine inspections should be noted in the ICAs (maintenance manual) but are not required to Page C - 129