It's been 50 years since a large steel pressure vessel fabricated in Wolverhampton exploded during hydro-test just before Christmas in 1965, throwing a chunk of steel the size of a car through the wall of the factory[1].
By a miracle no one was seriously injured, but TWI led the investigation into what caused the catastrophic brittle fracture that day. A large piece of that failure has been kept at TWI since the accident, serving as a memory of what can go wrong when lessons of fracture avoidance are not learned. You can see it still, standing in the atrium at TWI's Cambridge office.
The mechanism that sank the Titanic
Fifty years before the explosion of that pressure vessel, it had been another high-profile failure that had made the headlines: RMS Titanic. Only recently have engineers understood the role that the brittle fracture of materials played in the sinking of the Titanic on April 14 1912[2].
When the iceberg struck the hull of the Titanic, it was at high speed, and at low temperature (around 0°C). It is known now that the rivets and the hull steel had very low impact toughness, based on testing pieces of hull recovered from the seabed in the 1990s.
Low temperature, high loading rate and low toughness: a classic combination for brittle fracture. However, investigations at the time focused only on improvements to ship design and safety procedures, since, without an understanding of the fracture behaviour of metals, they could not seek to improve them.
Although the Charpy test to characterise material fracture behaviour was first described around 1900, it was not until the Second World War that the benefit of fracture testing was realised in earnest.
The birth of fracture mechanics
This time, it was the Liberty Ships, being built by the USA using welding instead of riveting to reduce ship production time from months to just 42 days. This was a significant contribution to the war effort and Allied victory, but a small number of the 2710 Liberty ships that were built between 1941 and 1945 suffered significant brittle fractures in the hulls; some even broke in half. Major effort in the USA and the UK was put into understanding the causes of these failures[3], bringing with it the true dawn of industrial fracture mechanics.
But by the time of the Wolverhampton incident in 1965, it had been over two decades since the Liberty Ship fractures had started the academic discussion about brittle fracture and fracture mechanics of steel structures.
At Cambridge University's engineering department, Constance Tipper, working with Prof (Lord) Baker, had written in the 1950s about the ductile-to-brittle transition of steels and the role that welds can play in making brittle fracture more likely[3].
Early research at TWI
In 1951 Baker had encouraged another of his research students, Alan Wells, to go and work for BWRA (as TWI was known then), and develop the understanding and testing of brittle fracture in thick section welded steel. Wells subsequently became well known for inventing the CTOD (crack tip opening displacement) concept of fracture, which he first described in 1961.
Therefore, at the time of the pressure vessel failure all those years ago, there was still no standard for fracture toughness tests, there were no fitness-for-service assessment procedures, and the understanding of fracture mechanics concepts such as CTOD was in its infancy.
Nonetheless, the fabricators should have known how to avoid hydrogen cracking in the welds; they could have performed Charpy testing, and ought to have ensured better control of the post-weld heat treatment in order to avoid catastrophic failure that day.
The BWRA report into the failure recommended using fracture mechanics principles to provide acceptance criteria, and to ensure hydro-testing was done above the ductile-to-brittle transition temperature[1]. Good advice to this day.
Lessons learned
In the decades since, failures like that pressure vessel have become extremely rare, certainly in the developed world, and not just because the litigation surrounding modern failures restricts the publicity around possible cases. No. Brittle fractures occur far less frequently because we have been learning the lessons over time about how to avoid them.
These lessons are now so firmly embedded in modern design codes and acceptance standards as to have become common knowledge. Post-weld heat-treat thick sections; meet certain Charpy requirements when low temperatures will be experienced; do not permit cracks in welds. Obvious.
Failures in recent decades seem more likely to be caused by corrosion damage, or by operator or procedural error (such as Piper Alpha or the Buncefield oil fire), rather than as a consequence of the behaviour of the welds and materials in service.
However, while materials from the 1960s remain in service they can still be susceptible to brittle fracture, as has been experienced in the USA with some fairly recent failures of vintage gas transport pipelines.
But as industry moves towards the rote-learning of these fracture-avoidance truths, are we at risk of becoming complacent that we have 'solved' brittle fracture?
Engineering critical assessment provides a technique to calculate the likelihood of fracture, even without the design code rules. Perhaps brittle fracture has gone away and we can stop worrying about it? But the more our understanding of fracture is seen as 'common knowledge', the less commonly understood that knowledge seems to become.
Risks from complacency
The Hatfield rail crash in October 2000 was due to pre-existing fatigue cracks which caused a brittle fracture of the rail[4]. The track operator had found the crack nearly two years before the crash and scheduled a replacement rail.
They should have known the risk that the cracks posed for causing brittle fracture of the rail, but they didn't understand the urgency and importance of replacing it, and failed to manage effectively the work of the track maintenance contractor. In addition, the track maintenance contractor failed to manage the inspection and maintenance of the rail at the site of the accident effectively and in accordance with industry standards.
The crash from the derailment of the train resulted in four passenger fatalities, over 70 injuries, and a manslaughter trial against six rail executives and their companies. The case highlighted the failings in the corporate knowledge of the condition of the railway tracks, and led to a huge programme of track repairs.
Two years after Hatfield, The Institute of Rail Welding was founded by The Welding Institute to help prevent further accidents through lack of understanding.
Today, if we become complacent about brittle fracture, we risk relearning the ignorance of last century. Today it is our responsibility to keep in mind the lessons of the past, not just for our business success, and technical reputations, but to save the lives of people who may needlessly die in horrendous accidents that brittle fractures can cause.
Continuing to inform industry
TWI has spent the last 50 years progressing the knowledge of fracture testing and assessment by the hard work of generations of its engineers. By sharing this knowledge and experience with our Industrial Members through projects, research and training, they too can keep ahead of current challenges in fracture. And this work continues to this day with TWI engineers continuing to lead the development of national and international standards for fracture testing and assessment.
Postgraduate research through the National Structural Integrity Research Centre[5] is investigating the validation of the CTOD concept to alloys other than structural steels, and the effect of a sour environment or high strain rate loading conditions on the fracture behaviour of modern steels, amongst numerous other integrity research topics.
Meanwhile TWI's Industrial Members also continue to seek out fracture knowledge, with fracture-related articles being amongst the most popular from TWI online. In this way it is hoped that no TWI Member will suffer the consequences of fracture complacency. Fracture is still as important today as it has been for the past half a century. And as for the next 50 years? I can't wait to find out!
Written by Dr Philippa Moore CEng FWeldI
Team Manager, Fracture Testing and Materials Characterisation
References:
- 'Investigation into the failure of an ammonia converter by brittle fracture', by F M Burdekin and T Boniszewski, BWRA Report, 1966.
- 'The Royal Mail Ship Titanic: Did a Metallurgical Failure Cause a Night to Remember?' by Katherine Felkins, H.P. Leighly, Jr., and A. Jankovic, JOM, 50 (1), pp. 12-18, 1998.
- 'The brittle fracture story' by Constance Fligg Elam Tipper, Cambridge University Press, 1962.
- 'Train derailment at Hatfield: A final report by the independent investigation board, Office of Rail Regulation, July 2006.
- The National Structural Integrity Research Centre, NSIRC, (http://www.nsirc.co.uk/) founded in 2013.