Historical failures and the evolution of Fracture Mechanics

    1. INTRODUCTION

 

If one looks back at the undergraduate course on Strength of Materials, it may be recalled, that, from the beginning of the coarse through its end, the material (of which the structure is made up of) was idealized as being homogeneous (i.e. of uniform composition) and a continuum (i.e. made up of a continuous mass than of discrete particles). Thus, the subject of strength of materials never recognised or considered the inherent flaws in the structure.

It is interesting to note that many of the structural failures that have occurred since past few decades [some, not being very long ago] have been because of the flaws/ defects present in the structure since its inception and the growth of these flaws during the life time of the structure, thus, resulting in the collapse of the structure. In fact, if one goes back to the history of the subject of Fracture Mechanics, it may be appreciated that the subject has gradually evolved through the lessons learned through such structural failures which have been because of the inherent flaws in the present in the structure and their growth through different mechanisms during the life time of the structure.

Through this article, I have attempted to focus on the following structural failures and the lessons learned through these failures which have contributed to the evolvement of the subject of Fracture Mechanics over the past few decades. These failures being:

 

• Boston Molasses Tank Failure
• Liberty Ship Failure
• Comet Disaster
• Aloha Airlines Fuselage Failure

I have discussed the above failures through the following sections in chronological order and have finally (in the last section of this article) attempted to bring out the lessons learned through these failures.

        2.  BOSTON MOLASSES TANK FAILURE

 

The Collapse:

 

On January 15, 1919, Boston suffered one of history’s strangest disasters: a devastating flood of molasses…! A large molasses tank in Boston, Massachusetts in the United States exploded with a great force. The tank contained 2.3 million gallons (i.e. more than 8 million litres) of molasses at the time of collapse. The steel tank was enormous:  50 feet (15 metres) in height and 90 feet (27 metres) in diameter.

Figure: The Boston Post Headlines January, 16, 1919

 

The wave of molasses rushed through the streets of Boston at an estimated 35 miles per hour (i.e. 56 kilometres per hour) killing 21 people and injuring around 150. The power of the wave was sufficient to rip buildings off their foundations. A piece of the tank was blown into the elevated railway tracks, breaking girders (see picture below) and almost forcing a northbound train off its tracks. The property damage was estimated to a total of $100 million in today’s dollars.

Figure: A large piece of tank pushed against an elevated railway

 

Some structural aspects of the steel tank:

 

The tank was itself just 3 years old at the time of the disaster. The tank was constructed of large curved steel plates, seven vertical rows overlapping and held together with rows of rivets, the whole set into a reinforced concrete base.

The steel thickness varied from 0.67 inch for the first ring at the base to 0.31 inch for the seventh ring at the top. The figure below shows the vertical joint in the first ring at the base.

Figure: The geometry of the vertical riveted joint in the first ring at the base of the tank

Causes of failure:

 

Several factors that occurred on the day of the collapse and the previous day are believed to have contributed to the disaster:

 

1. Thermal shock: The tank was poorly constructed and insufficiently tested. Due to fermentation occurring within the tank, carbon dioxide production may have raised the internal pressure inside. The rise in the local temperatures that occurred over the previous day might have also assisted in the building up of the pressure.
    
   Records show that air temperature rose from -17°C to 5°C over that period thus resulting in a thermal shock.
    
2. Fatigue crack propagating from near the base of the tank: The failure occurred from a manhole cover near the base of the tank (the hoop stress being greatest near the base of a filled cylindrical tank), and it was believed that a fatigue crack grew around to criticality. The tank had been filled to its capacity eight times since it was constructed thus putting the walls under an intermittent cyclic load.
    
    
3. Poor structural health monitoring: It has been reported that the storage facility was never properly tested - by filling it with water - because a shipload of molasses was due only days after the completion of the tank in December 1915. From the beginning leaks had appeared and it is believed that the distillery company had painted the tank brown deliberately to hide the leaks (since molasses is brown in colour).
    
   It should be noted that the failure occurred due to a fatigue crack growing near the manhole cover to its critical length, thus, throwing light on the crack growth mechanism due to fatigue.
    
   Metallurgical aspects of the tank that led to a brittle collapse:
    
   Two metallurgical aspects of the molasses tank steel are of note: the likely brittle behaviour during the failure temperature (4°C) and the observation of a microstructural feature called Neumann bands.
    
   Low carbon and manganese in steel leading to explosive / brittle behaviour:
    

Most of the structural steel production in 1915 was by open hearth process (https://en.wikipedia.org/wiki/Open_hearth_furnace). The material exhibited very good strength with high ductility. At that time, fracture toughness was not specified in material procurements.  The chemistry for steel in the plates containing the manhole (where the fatigue crack grew to its critical length) is shown the table below and compared with 2 modern steels:

Table: A typical chemistry for the molasses tank steel shell plate and current steels

Note the particularly low value of manganese in the steel used to construct the molasses tank. Chemistry, thermal processing and quality of production all affect the fracture properties of steel.

For the class of steels corresponding to the molasses tank, carbon has the greatest effect on the propensity (tendency) for brittle behaviour. The higher the carbon, the higher the temperature at which there is a transition from brittle to ductile behaviour. Some research data (of 1951 i.e. 32 years after the collapse!) suggests that the transition temperature for the molasses tank steel could have been as high as 15°C which is significantly above the operating temperature at the time of failure (which was 4°C).

Neumann bands (the contribution of the failure to Fracture Mechanics)

Neumann bands are narrow bands, a few micrometres wide, usually within a grain. Substantial research even before the molasses tank failure indicated that this feature was only produced in low carbon steel by explosive loading or at extremely low temperatures. Expert reports showed micrographs of Neumann bands were found near the primary fracture. It is now known that a rapidly propagating crack in brittle low carbon steel can produce Neumann bands without explosive loading. Crack velocities in brittle steels can reach as high as 3,000 feet per second.

1   4. LIBERTY SHIP FAILURE: BRITTLE FRACTURES AND THE BIRTH OF FRACTURE MECHANICS

Figure: One of the ‘survived’ Liberty ship

History behind the construction of Liberty ships:

As part of the war effort during the World War II, the materials, munitions and supplies were vital to the United States in order to sustain the battle against the Germans. A very successful bombing operation by Germany on the ships had inflicted significant damage to the munitions of the United States.

As a result, in 1941, President Roosevelt announced that $350 million would be spent to provide a ship building programme, the objective being to build ships faster than the enemy could sink them!

In pursuit of this objective, the Americans decided to use the method of “welding” than the conventionally followed method of “riveting” for ship building. The ships were called as “Liberty ships”. In addition to increasing the speed of construction, the use of welding also decreases construction costs. The number of skilled labourers required carrying out welding on the ship’s hull and the deck were thought to be significantly lesser than the numbers required to carry out using riveting.

The sinking, cracking and the damage:

2708 Liberty ships were constructed between 1939 to 1945. 1038 damages or accidents were reported by April, 1^st 1946!

More than 200 Liberty ships sank or were damaged beyond repair. “Schenectady” (as named) was one of those ships which broke into 2 pieces with a loud sound (see picture below). The accidents were caused due to the lack of fracture toughness of the welded joint. The accident highlighted the importance the importance of fracture toughness and marked the “birth of fracture mechanics”.

Figure: “Schenectady”- the ship that broke into two

Causes of collapse:

It was clear from the nature of the failures of the ships that the collapse occurred through a brittle fracture. It was observed that nearly all the failures occurred in the cold waters of the North Atlantic whereas the ships stationed in the South Pacific remained intact! The mystery was later resolved by Constance Tipper of Cambridge University.

She demonstrated that there is a critical temperature that there is a critical temperature below which the fracture mode in steels shifts from ductile to brittle. That is: the fracture toughness of steel changes drastically over a small temperature range. The low temperature steel is brittle and fractures by cleavage. At high temperature, it is ductile and fails by plastic collapse. In transition between ductile and brittle, both mechanisms of fracture can occur. Thus, there is a region where the material is 100% ductile and a region where the material is 100 % brittle and a transition zone as shown in the figure below.

Figure: The ductile-brittle transition (note: the low fracture toughness in the zone corresponding to low temperatures)

The ships stationed in the North Atlantic were susceptible to brittle fracture as the temperature was in the corresponding zone as marked above.

Quality of the welds:

 

The welding techniques used in ship production caused a controversy. There were issues with inexperienced labourers who had been drafted in to increase ship production for the World War II efforts. It was believed that the unskilled welding caused micro-cracks in the weld itself, thus, resulting stress concentrations which contributed to the brittle fracture of the Liberty ships.

Crack growth, fracture: Locations

It should be noted that the most common type of cracks (for most of the ships) was one which began at the square corner of the hatch (opening) which coincided with a weld. Thus, both the weld and stress concentrations acted as localised areas of high stresses.

One remedial measure adopted was to use rivet steel arrestor plates in areas of higher stress concentration thus arresting crack growth. In fact, Victory ship was an upgrade in ship

design had arrestor plates to maintain a less stiff and stronger ship design that was better able to deal with fatigue.   

Thus, the reasons of the failure of the Liberty ships could be summarized as;

• The material used did not have sufficient fracture toughness especially at lower temperatures.
   
• The standard of the welded joints was in general poor due to inexperienced welders; which meant there were micro cracks in the welds.
   
• The all-welded construction eliminated crack arresting plate boundaries which are present in riveted joints.

1   4. THE COMET DISASTER

de Havilland Comet was the world’s first production commercial jet liner. Developed and manufactured by de Havilland (https://en.wikipedia.org/wiki/De_Havilland), the Comet’s first prototype first flew on 27 July 1949. It featured an aerodynamically clean design with 4 turbojet engines buried in the wings, a pressurised fuselage and “large square windows”. For that era, it offered a relatively quiet, comfortable passenger cabin and showed signs of being a commercial success at its 1952 debut.

Initial success and the sensation:

During its first year of operation, the Comet carried a total of 28,000 passengers covering a total of 104 million miles. By 1953, de Havilland Comet had firm orders of 50 Comets from the world’s several airlines and was negotiating for 100 more. The Comet was looked upon as a great success throughout the world.

And, then, after a year of service, the accidents began (1953)….!

Accidents in 1953:

The first 2 accidents in 1953 occurred during take-off.

In the first, the plane failed to become airborne. The accident was blamed as a pilot error and no one was killed.

The second accident left no survivors and appeared to be a design flaw and modifications were done to the wings that allowed a greater lift at low speed.

The third accident occurred as the Comet crashed on take-off from Calcutta in India. In spite of these 3 accidents, the public confidence was in place until January, 10 1954…!

What happened on January 10, 1954?

On the 10^th of January, 1954 a Comet after departing from Ciampino Airport in Rome to London, had climbed 26,000 feet en-route to its assigned altitude of 36,000 feet when it plunged into the sea. Witnesses in the island of Elba in Italy saw the aircraft fall into the sea in flames. All 29 passengers and 6 crew members were killed.

The Elba investigation and the Comet water tank test:

While a crash investigation is normally conducted by the government or aviation authority, it was decided that the British authorities would head the Elba investigation. Around 70% of the Elba wreckage was recovered following thorough investigation which for the first time included using under water television cameras.

The authorities salvaged the wreckage from the ocean and reconstructed the aircraft. Once the debris was put together, it was found that the cabin itself had failed and the conclusion was the disaster was due to extreme decompression of the cabin.

The Ministry of Civil Aviation decided upon a unique test to figure out the exact cause of the accidents. They built a large tank, large enough to hold one of the grounded comets. The wings protruded from the water tight slots from the sides of the tank (see picture below):

Figure: The Comet in the water tank for pressure tests

The wings protruded from the water tight slots from the sides of the tank. Then the tank and the cabin were flooded with water.

By using water instead of air, water being a much less compressible fluid the test would be much safer and fuselage would be able to be repaired and re-tested as necessary. Had air been used, the results would have resembled the catastrophic inflight break up like the 2 flights. The wings of the aircraft were moved up and down by hydraulic jacks to simulate the flexing of the wings that occurs due to the air currents during flight.

The Comet that was being tested had undergone 1230 pressurized flights before testing and 1830 “tank flights” before the fuselage failed at the corner of a square forward escape hatch window.

Figure: Fuselage failure at the corner of a “square” escape window

Several months later the results of the tests were confirmed when a large section of the cabin roof was recovered from the sea. A crack had started in the corner of the navigation window at the top of the fuselage. Even through it was a small crack of a few mm at the corner of a square window, structurally, total rack length = length of the actual crack + length of window diagonal.

The stress concentrations were high specifically because of the squarish shape of the windows and window frames which is very different from the round/oval shapes of modern airplane window.

And this was intuitive reason for which the US Administration did not allow the Comet to fly because the size of windows was larger than the normal windows. This was purely out of engineering judgement, then…!

 

Causes of the disaster besides the square windows:

Apart from the square windows, de Havillands testing of the new plane was inadequate. They had limited the new cabin to static testing alone meaning that during the tests, they had subjected the cabin to pressurization alone but had neglected the effects of motion such as flexing of the wings. Thus, the actual service loads were not simulated properly and effect of flexing was neglected. This provided an explanation that the structure was not fully design proof.

If one looks at the history of fatigue development, this was the time that the concept of S-N curve was being researched.

1   5. ALOHA AIRLINES FLIGHT 243 FUSELAGE FAILURE

 

Aloha Airlines Flight 243 was a flight between Hilo and Honolulu in Hawaii. On April 28, 1988, a Boeing 737 serving flight suffered extensive damage after an explosive decompression in flight, but was able to land safely in spite of fuselage failure mid-air! There was one fatality (of one of the flight attendant) and another 65 passengers and crew were injured.

Figure: Aloha Airlines flight 243 fuselage failure

It may be noted that the other three failures discussed in this article: Boston Molasses tank failure, the Liberty Ships failure, Comet disaster, all occurred after the structure had been in service for the first few years. But, in the case of the Aloha airlines fuselage failure, the failure occurred after the structure had been in service for 19 years whereas the intended design life was 20 years. Following the investigation, it was concluded that widespread corrosion was the main cause.

 

The disaster marked a turning point in the history of aircraft corrosion. At the start of the jet age (1950 to 1960’s), little or no attention was paid to corrosion and corrosion control. Only the more recent designs like Boeing 777 and later version of 737 have incorporated significant improvements in corrosion prevention and control in design and manufacturing.

16.  LESSONS LEARNED THROUGH THESE FAILURES

 

In summary, it could be concluded that the following lessons have been learned through these failures;

Boston molasses tank failure:

• The failure brought out the importance of structural health monitoring. The tank was painted brown and hence no visual inspection was possible on any leak of molasses which is also brown.
   
• The failure reinforced the concept of ‘Neumann bands’ which are known to occur in in a rapidly propagating crack in brittle steels.
   
• The failure threw light on the composition of steel necessary for ductile behaviour.

Liberty ship failure:

• The disaster brought out the importance of temperature effect on fracture toughness and marked the birth of fracture mechanics.

Comet disaster:

• The failure highlighted that cracks could occur in stress concentration zones and grow in service due to fatigue loading causing failure
   
• The failure also highlighted the importance of proper simulation of service loads during testing. It may be noted that flexing of wings had been neglected during the testing.
   
   

Aloha Airlines fuselage failure:

• The failure brought out the importance of stress corrosion cracking
   
• It gave impetus to improved design approaches for corrosion prevention and control

     

A