This article is not meant to be a scientific dissertation about the subject of metal fatigue but just scratching the surface for understanding of (young) engineers with a different expertise but interest in the subject. Just my view.
Note; during research for- and writing of this article, I came upon an excellent high level paper by R.J.H. Wanhill titled "Fatigue requirements for Aircraft Structures", that covers pretty much the scope of this article. It's free downloadable from Researchgate.net and also from below:
Metal fatigue related accidents
Boeing 737 Aloha 243
The above picture shows the fuselage of a Aloha Boeing 737 that blew apart in flight by cabin pressure loads, due to widespread undetected fatigue cracking in the fuselage skin that grew to a point that the residual strength was unable to sustain normal cabin pressure loads and subsequently failed. The accident triggered a intensified drive to better understand and manage metal fatigue progress and the contributing factors to crack initiation and crack growth. The accident happened in 1988 but was certainly not the first metal fatigue related fatal accident in aviation.
The failure mode was tension in the skin lap joints caused by cabin pressure cycles.
Below the accident investigation report.
DeHaviland Comet
Other high profile accidents were the subsequent in-flight breakups of two DeHavilland Comets within a year. The Comet was the first commercial jet and had a pressurised cabin. The failure mode was similar (but not the same) to the aforementioned Aloha accident.
Just as after the Aloha accident, studies started to better understand the phenomenon of metal fatigue.
Failure mode was fuselage skin cracking originating from window corners caused by cabin pressure loads
The year was 1954. Below the enquiry report
Boeing 707 Dan Air Lusaka
Another less known accident was the Dan Air Boeing 707 that crashed not far from Lusaka due to an in flight separation of the horizontal stabilizer. Also caused by metal fatigue.
The accident occurred in 1977.
Failure mode was fatigue cracking in the horizontal stabilizer spar caused by aerodynamic loads. Personal note; it says here aerodynamic loads and not flight loads; anyone that has carried out static high power engine runs on commercial transport aircraft, knows that it can be violent on the structure of the aircraft, especially on the empennage.
Below the enquiry report
Boeing 747 JAL 123
The last accident I will mention here was the Japan Airlines Boeing 747 that had a post tail scrape repair on the aft pressure bulkhead that was erratically installed, causing much higher fastener loads than intended in the repair design and consequently fatigue damage progressed much faster than anticipated with the design, resulting in the inflight failure of the aft pressure bulkhead. The accident occurred in 1985. The explosive decompression was so powerful that it blew the vertical stabilizer off the aircraft, resulting in secondary damage to the flight control hydraulics that eventually rendered the aircraft uncontrollable and it tragically ended in the side of a mountain.
Failure mode; web cracking in the aft pressure bulkhead, caused by cabin pressure loads
Report below (very thorough and detailed)
Metal fatigue in gas turbine engines
Since the above accidents there have been more accidents and incidents related to metal fatigue, however the most recent years mostly in gas turbine engines. This is a totally different failure mode than aircraft structural failures with much different conditions like material properties, vibration spectra, design geometry, operating temperatures, existing (impact, overheat) damage, material flaws, etc.
Metal fatigue in power plants have become more prominent because of the evolution in engine architecture; broken and shed blades are of all times, but now the emphasis is increasingly being put on fuel efficiency, fan diameters have increased progressively as they produce up to 90% of thrust and consequently contain a much higher energy level than in the low bypass engines of a bygone era.
Blade-out certification tests have become a spectacular and extremely expensive event. Unfortunately (fan) blade out events have become reality.
Metal fatigue in Helicopters
Helicopters are machines that have lots of sources of vibration originating from the power train that can cause fatigue failures within these components under certain conditions.
Structural fatigue progresses similar to fixed wing aircraft structures but fatigue in power drive train components are subject to different loads.
As the inspectability is limited in power drive train internal components, commercial helicopters are required to be equipped with health monitoring systems that are basically a myriad of vibration sensors at critical positions that sense vibrations levels at frequencies that are expected to occur at these locations. Vibration levels can not only indicate impending failure but also wear and possibly fatigue progress.
As this article is limited to aircraft structure, a review of the above show that the vast majority (but not exclusively) of fatigue cracking starts at pinned fastener holes. Over time types of fasteners have been developed to reduce fatigue initiation.
History of fatigue analysis
Metal fatigue analysis is not a new science.
The first known researcher of the phenomenon was August Wöhler, (22 June 1819 – 21 March 1914) a German Engineer from the Hanover Area. Wöhler was empoyed by the Prussian Railways and investigated the failure of the axles of rail cars that continued to break despite the fact that they were loaded well below the material stress limits. He discovered that repetitive loading and unloading fatigied the metal which caused cracking, which increased material stresses, which accelerated crack growth and so on, up to failure.
The relation between maximum stress level, stress amplitude and stress cycles was graphically displayed and was called the "Wöhler curve". Example below;
Over the years the understanding of this complex pehnomenon has improved.
A world renowned standard book was written by then Delft University Professor Fatigue and Fracture Mechanics, Jaap Schijve (passed away Jan 2023), titled "Fatigue of Structures and Materials".
In general, simplified, it can be said that fatigue damage is a function of material stress amplitude and a number of loading cycles for a given material.
In real world aircraft structures, of course there is no given stress amplitude, this would depend of design geometry, operating environment, way of operating aircraft, used materials, existing corrosion damage ( big contributor!) and mechanical damage.
NASA conducted a lot of research into metal fatigue and provided software and load spectra in order to use as tools for Damage Tolerance Analysis. This software is still available under the commercial name AFGROW
But in order to understand metal fatigue, there must be a basic understanding of Fracture Mechanics. Without going into the weeds of this interesting but complicated subject I will briefly touch on it here.
Fracture mechanics basics
In order to understand the background of damage stolerance, a basic understanding of fracture mechanics is required.
The following items are abbreviated:
Basic failure modes of fastened joints
Crack tip plasticity and material properties
Accumulated damage and crack growth
Crack instability
Basic failure modes of fastened joints
As a substantial number of the aircraft metal fatigue related accidents have failure of fastened joints as major contributing factor, is it important for engineers and designers to have a good grasp of the fatigue characteristics of fasened joints.
Below grapic shows the failure modes.
Every experinced (repair) design engineer knows that fastener selection, edge margin and fastener pitch are important parameters to mitigate fatigue related failure.
Fastener selection must always be such that the joint is bearing critical and never fastener shear critical
Sufficent fastener pitch (4D) mitigates net-tension failure, sufficient edge margin (>2D) mitigates shear-out and fastener shear strength being greater that material bearing strength mitigates fastener failure.
2. Crack tip plasticity and material properties
In general, on a stressed element, any discontinuities cause higher material stresses than the nominal stress applied on that element. The radius and taper are to a high degree driving the stress rise.
On an element containing a crack, the crack tip has a extremely reduced radius and consequently, the material stress at the crack tip is much higher than the nominal stress on that element. The material stress at the crack tip very easily exceeds the yield stress in ductile materials of failure stress in less ductile materials.
Example of ductule materials (you can bend without break) are:
Mild steel
Low Aluminium Alloys like 6061 and 2024 (very commonly used)
Examples of less ductile (no or little bend without break) materials:
High strength Aluminium Alloys like 7000 series, 7075 being the most commonly used
High stremgth steels such as used in Airliner landing gears
Cast Iron
Below figure shown the theoretical material stress around the crack tip.
As shown, the stress level approaches infinity, which in the real world not occur
Below figure shows the real world material stress level around the crack tip for a material with some ductility.
Ductility can be described as yield to failure. Less yield to failure is less ductile.
Generally, high strength materials have less ductility than low strenth materials. Consequently high strength materials (if loaded to their high stress capabilities!!!) have higher peak stresses at the crack tip and higher crack growth rates.
3. Accumulated damage and crack growth
With alternating loading materials accumulate damage as is scientifically proven. In undamaged homogenous materials, this damage is on molecular level and not detectable.
Once the damage starts to concentrate in one spot, as indicated above local stress levels start to rise and crack initiation takes place.
Once a crack has formed, it will grow under alternating loads as the peak stress is now located at the crack tip with every load cycle. Subsequentl load cycles cause separation of material as it exceeds the failure stress at every load cycle..
This is called crack growth. The crack growth rate is a very important parameter in damage tolerance management.
4. Crack instability
Crack instability is the point where the reduction in load bearing material surface is reduced to where it is no longer capable to sustain the load applied and the material fails in static overload.
This is usually expressed in critical crack length.
There is however a sliding scale in crack growth. See graph of typical crack growth below. Mind it is a log scale.
K is the stress intensity the crack tip peak stress divided by nominal stress
a is crack length.
One can see that in region III there is a progressive increase in crack growth rate until failure in overload.
Damage tolerant structure and the concept of damage tolerance
In the past there were two philosophies of material fatigue mitigation design:
Fail safe design, where a backup structural element is available in case of a failure of one. This is still being used but in lesser extent due to increased effectiveness of damage tolerance designs. In many cases the backup structure was seen to be as subject to fatigue damage as the main structural element.
Safe life design; this is still very much applied in cases where there is very high crack growth rates or limited inspectability or repairability. Commonly engine rotors and landing gear parts which are usually high strength materials, hence high crack growth rates. The philosophy is to limit the functional life of the component to the cycle count where there is an extreme low probability of crack initiation.
The concept of damage tolerance is a structural design that relies on inspections of various methods and intervals to detect crack growth and associated strength reduction before the residual strength decreases below certification loads, known as limit load.
Every structural element is designed to limit load, times a safety factor of 1.5 which is called ultimate load.
Limit load being the maximum load on a structural element to occur in certified operation.
The philosophy is that cracks and other strength reducing damage (such as corrosion) are detected before residual strength dips below limit load.
If cracks are detected, repairs should be applied to restore strength to ultimate load capability of the structural element.
Inspection intervals and thresholds are initially developed by theoretical analysis carried out by the manufacturer, based on service loads, load spectrums and materials and geometry.
A publicly available tool is software developed by NASA in the past and still active and commercially available. The name of the software is AFGROW and can be used by independent Design organisations to substantiate Damage Tolerance Analysis.
Graphic representation of the damage tolerance cycle shown below;
FAA published good comprehensive practical guidance material in Advisory Circular AC 120-93.
Downloadable below.