At about 6.22pm on August 21, 2009, Pier 4 of the Malahide viaduct, along with the post-tensioned concrete beams of spans 4 and 5, collapsed into the Broadmeadow estuary, about 13km north of Dublin, just after an Iarnród Éireann (IÉ) train travelled across it.

All signals were immediately set to ‘Danger’ to prevent further trains from crossing the collapsed viaduct which is shown in Figure 1. At the time of the accident, it was reported that 82 passenger services travelled over the Malahide viaduct every weekday, with one additional train travelling on a Friday. In addition, six freight trains travelled over thecrossing every weekday. Based upon this traffic volume the potential for serious injuries was extremely high.

Figure 1: Image of collapsed viaduct from the RAIU report1.

Despite the relief that nobody was injured due to the collapse, the immediate challenges for IÉ were immense as the company, in parallel with repairing the physical damage to the viaduct, grappled with questions such as “What about all our other assets?”; “How can we reassure our stakeholders that our assets are safe and fit for purpose?”; “How can we ensure that this kind of event does not happen again?”; “……..”.

Responding to such questions has resulted in significant changes in relation to how critical infrastructural assets are managed within IÉ. These changes were informed by the investigation report produced by the Railway Accident Investigation Unit (RAIU), which was commissioned to study the collapse.

At that time the RAIU was an independent investigation unit within the Railway Safety Commission (RSC) and the investigation was initiated to improve railway safety by establishing the cause(s) of the accident with a view to avoiding such accidents in the future and making rail travel safer1.

This article provides some historical background and summarises the key findings of the investigation that studied the collapse. It goes on to explain key points of the approach taken by IÉ to maintain and protect the infrastructural railway assets both in the immediate aftermath of the accident and the intervening period and describes the systems that have been put in place to help ensure asset integrity going forward.

Summary of the investigation report

Historical data indicates that the Malahide viaduct has been susceptible to the defect of scouring since its original construction in 1844. Scouring is the engineering term for the erosion of soil or other bed-substrates surrounding the foundation, piers and abutments, of a bridge.

When fast-moving water erodes sediment from the bridge foundation, it results in scour holes. These holes can seriously compromise the bridge’s integrity and in the United States alone it is believed to have caused 600 bridges to collapse in the period 1984-20143.

Efforts to counteract scouring at the Malahide viaduct resulted in the construction of a stone weir in 1846 to protect the original timber piles and maintain a constant level of water in the estuary.

Despite subsequent changes to the viaduct such as installing masonry piers, post-tensioned concrete beams, and a grouting programme to inject concrete into the weir, the scouring problem remained. IÉ correspondence reviewed by the RAIU directly refers to this issue of scouring in 1957, 1972, 1992, 1995, 1996 and 1997. Periodic discharges of stones along the weir were made to protect it, with the last documented discharge of stones being in 1996.

According to the RAIU report, knowledge in relation to the scouring appears to have continued until 2002 when the person responsible for the Malahide viaduct left the division. The persons responsible for the Malahide viaduct after this time were not aware of the scouring issues or the necessity to continuously maintain the weir. However tidal scouring continued, and by 2009 this led to the collapse of a pier, resulting in the precast concrete beams it was supporting falling into the estuary below.

Figure 2: Progression of scouring and the undermining caused by piping under pier 4. Piping occurs when there is a sub-surface formation and progression of a continuous 'pipe-like' tunnel1.

Figure 3: Plan view of progression of undermining of pier 4 as a result of piping1.

The RAIU report documented that at certain points in the past IÉ did recognise the risk of scour prior to the Malahide viaduct collapse, but it appears that this had been forgotten, that the knowledge had effectively been lost, in the period of time leading up to the actual collapse. The report noted that the track patrolling standard called for a “check underneath for scour at least once a year” for all bridges over rivers or waterways.

The reality is that patrol gangers focus on inspecting the track and identifying points of change relating to the track condition using appropriate measurement equipment. Technical inspections were expected to identify scour. Special scour inspections did occur prior to the collapse, the problem was that they were not cyclic but ad-hoc with the the last scour inspection for Malahide (pre-collapse) believed to have been sometime between 1999-2002.

The RAIU reported finding no records to show that an annual check for scour of the Malahide viaduct was actually carried out. It is also noted that inspecting a bridge for scour requires specialist training/skills and there are no records to suggest that the patrol ganger(s) patrolling the Malahide viaduct track had received any training in the identification of scour1.

It is clear from the investigation report that no formal system was in place for capturing, recording, evaluating, and sharing details relating to inspection and maintenance within the company. For example, in terms of putting stone around the weir to reduce the likelihood of scouring no system was in place to capture the physical activity of doing this. In addition, there was no documented process of how to actually inspect the assets.

As a result, inspectors, or problem spotters, who were expected to identify problems which needed maintenance, could look at assets but were unaware of the asset’s prior history and the optimum process(es) of how to actually conduct an inspection, evaluate the potential risk and consider the most appropriate actions to take.

The RAIU report refers to this as “corporate memory”, which it describes as “knowledge and information from the company’s past which can be accessed and used for present and future company activities”. A total of 15 safety recommendations were made in the RAIU investigation report which clearly highlighted that IÉ did not actively engage in knowledge management, essentially the formal collection, storage and transfer of knowledge and information between generations of IÉ staff at the level needed prior to the accident.

In summary, prior to the accident deficiencies in three key areas within IÉ can be identified:

  1. Corporate: this can be evidenced by a lack of internal standards/processes (or at least a lack in them being adhered to), and a lack of management of data/information for the proper organisation and thus poor coordination of personnel in the field.
  2. Field personnel: this is evidenced by a lack of systematic recording of data, a lack of adequate training, lack of awareness and standards among the inspectors.
  3. Risk evaluation and decision support tools: prior to the accident documentation systems were rudimentary and basic with no systematic process or standard to follow or risk evaluation method to predict when a particular event might occur. In addition, no decision support tool, populated with appropriate data, was available to help employees quantify and respond to such risks.

What Iarnród Éireann-Irish Rail did after the collapse: Immediate/short term

Since the viaduct collapse, IÉ has been addressing the deficiencies outlined in section 2 by managing the company with a specific focus on ensuring the full use of, and taking the benefit from, the corporate memory within the company. In so doing the company has been managing and minimising potential risk(s) for all stakeholders.

The primary checks for overall bridge safety across the IÉ network are inspections by appropriately trained and qualified staff at a set frequency:

  1. A general engineering inspection is conducted every two years.
  2. A principal engineering inspection is conducted every six years.

A scour inspection, in addition to the principal engineering inspection, is scheduled on a risk-based frequency of every one, three or six years depending on the scour risk. This scour risk is a derived calculation from a series of metrics based on recordings from the underwater inspection as well as an assessment of the risk potential both upstream and downstream of the bridge.

An issue is that most bridges that were inspected did not exhibit scour on an ongoing or regular basis. The presence of scour problems and their progression to become a potentially significant problem was determined to be erratic, so a new approach had to be developed to fully understand how to deal with the risks surrounding this random scour behaviour.

Immediately after the accident IÉ, realising that the dearth of records could prevent it developing a new approach, began the process to collate, centralise and digitise all the documentation relating to its infrastructural assets.

The railway network was initiated in Ireland in the 1830s and much of the documentation was created in the subsequent decades as the railways rolled out across the country.

Multiple sources of information existed, and it is estimated that up to 90-95% of it was not digitised at the time the viaduct collapsed. Digitising all the information on bridges unearthed a treasure trove of documents, primarily of historic interest.

A significant number of these documents were of poor or questionable quality or use to the engineers faced with understanding the current state of the assets.

For example, drawings were mislabelled, the bridges were not built as designed and, in many cases, subsequent maintenance modifications or renewals were not adequately recorded. All this information was collated for up to 5,000-plus assets involving water, such as bridges, culverts and retaining walls. It was soon realised that although there is a value in digitising assets this alone was not enough to identify and understand the current status of each asset.

Instead, it became clear that there was a need to reset the baseline for every railway bridge asset in the country to include a root and branch study of the health, status, and construction of each. The objective always was to focus on what can’t be seen and to understand their true current condition.

Priority was given to the 452 bridges over water and specialist divers with knowledge of inspecting for scour damage were employed to check all those bridges using a preliminary prioritisation ranking based on factors such as utilisation, location, and age. The initial focus was on the 105 assets deemed to be of highest potential risk.

In undertaking this work a new hybrid standard, based upon the literature and broadly similar to4, was created with the inspection frequency being established using a risk-based assessment number. This involves a series of steps, essentially forming a two-level assessment resulting in a risk rating for each individual asset:

  • Level 1 assessment involves an underwater examination and study of the flood plain information for each bridge site.
  • Level 2 assessment involves modelling using key elements such as site data, measured scour data, flood event data, and channel geometry data. A site-specific modelling of the highest potential risk bridges was performed by expert hydrologists to study how the scour risk might change in certain scenarios; for example, what would happen if new building developments occurred along riverbanks, up to 400m upstream or up to 100m downstream, to change the flow profile?

The outcome of the Level 1 and Level 2 assessment is a risk rating factor which determines the need for and nature of specific remediation or maintenance plans which might be needed for each site.

Specific changes that occur, for example, excessive flooding, can trigger specific actions and a need to review the assumptions underlying the models. International models for scoring scour risk (scour vulnerability rating) calls for various inspection regimes such as one, three, six and 10-year intervals.

Taking a conservative approach, IÉ used an initial scour vulnerability rating inspection frequency of one/three/six years. Tracking was performed and frequent inspections were used to profile/validate the model and the risk of scouring.

At all times the focus was on informed, fact-based decision making. During these inspections scour issues were found in a lot of bridges which necessitated remedial actions such as the addition of rock armour, stone mattresses and grout pumping. This highlights the importance of inspections to understand the true risks involved and formed part of the actions to design solutions to move all bridges to a six-year scour inspection frequency.

It is only in the last few years that the three-year inspection cycles have been changed to six-year cycles where engineering data supported such change. Even now, almost 15 years after the collapse the company still has in the order of 25 bridges on a one-year inspection frequency for scour.

Figure 4: A simplified flowchart of what IÉ actually did after the collapse.

Ongoing asset management plans: Using the IAMS system

Reports and report cards can be a useful tool to document corporate knowledge, but they need to be factual so that they represent the current situation and must be accessible to those who need such access.

At the time of the collapse the asset management system in use, the Infrastructure Asset Management System (IAMS) which was initially introduced in 2005, was rudimentary and poorly populated with some assets not even appearing on the register. This has now changed; all assets must be listed on the register. IAMS is not just a list of assets; it is a SAP-based system that facilitates decision making in addition to work management.

A key driver of the implementation of IAMS is that it facilitates the move from reactive (where the company responds to incidents) to predictive analysis of the network risk (where decisions can be made based on the actual risk).

In practice a local engineer will now get an email telling him/her what assets need to be inspected in the next two to four weeks, for example, and the system flags an error, or reminder, if the inspections are overheld and do not meet the timeline specified. In this way inspections are prioritised so that what really needs to be changed is correctly identified and clearly understood, instead of just focusing on spending the money which could be the case if the engineer prioritised both maintenance and renewals.

Due to the reorganisational restructuring as a result of the Malahide viaduct collapse there is now a split between the inspector (problem finder within technical services) and the maintainer (problem fixers). Splitting these responsibilities and having dedicated finders and fixers is an approach to try and avoid the need for local prioritisation decisions.

The finder creates a notification, and the fixers will agree to take on work as prioritised within a work-bank system. Every asset has three scores relating to design, condition and deterioration. A score of one is great (new), while four would indicate the need for a major retrofit. It is noted that in most cases points and crossings were a bigger issue than bridges.

It must also be noted that a lot of the assets involved are Victorian and there is likely to be always something 'wrong' in that the asset may have some rust, wear and tear, etc. If non-critical issues are flagged as a priority, then it will not be feasible to complete the work-bank or programme of work. The focus is to manage deterioration to ensure that it won’t pose a risk to human life.

To deal with such problems and reduce the sense of urgency the concept of a technical non-compliance was introduced. Examples would include the deterioration of masonry/steel/etc of a bridge. Any degradation observed is put on a two-year inspection profile and flagged to ensure that it is checked during the next inspection.

If it is likely to deteriorate before the next inspection, then it must be included in the current 'work bank'. It is noted that a significant level of professional judgment is required since faults might not be as binary as, say problems with signalling function infrastructure.

Developing this professional judgment highlighted the need for improved staff training. Developed by an external third-party consultant this internal training course within IÉ delivers a series of specialist modules focusing on key areas such as the inspection of masonry, steel, etc.

After completion of the training the participants submit a number of reports and, if acceptable they are proposed to progress to the practical inspection assessment. This involves the trainee being taken to a random bridge where he/she is expected to inspect the bridge under real-world conditions, while being monitored by external assessors.

This training and the IAMS systems which have been put in place facilitate the determination of risk values as part of an overall risk model development. The risk value is determined by the product of hazard and vulnerability assessments of elements at risk6. A three-stage process has been implemented to develop this risk model:

Stage 1: (a) Data requirements and (b) Initial hazard model.

Stage 2: (a) Model refinement (hazard analysis) and (b) Vulnerability assessment and consequence analysis.

Stage 3: (a) Risk Model and (b) Decision Support Tool.

Figure 5: Overview of risk model (based on6).

It is this risk model, shown in Figure 5, which underpins all decisions made in relation to appropriately responding to this risk.  

Lessons to be learnt

1) The importance of knowledge management and appropriate training

It is difficult to predict failures of infrastructural assets. Since the accident there is an increased awareness and focus on the management of knowledge with a significant investment in the IAMS and the provision of specialised training to ensure its effective use. This helps to make tacit knowledge more explicit and ensures that the relevant skillsets and knowledge are developed and retained even if/when people are promoted within, retire or resign from the organisation or when work practices are perhaps changed as a result of cost-cutting exercises.

This can be of particular importance during periods of high employment when many people tend to change jobs, such as the 'great resignation' which occurred after Covid. Such events can cause discontinuities of knowledge management in terms of anecdotes being told between people to make knowledge explicit and underline the importance of having an accessible, fully populated, and current IAMS. A similar loss of knowledge can occur due to cost-cutting; an anecdote from another jurisdiction describes the situation7:

 Drainage assets lineside used to have dedicated maintenance crews; these were demobilised, with crews being amalgamated into permanent-way teams, primarily as a cost-cutting exercise. As a result, over a period of years, drainage quality declined resulting in degraded track-quality and subsequently train-delays.”

2) Align inspections and maintenance to the real risk

From a resource perspective it is important to align the actual inspections (and any maintenance needed) to the actual risk profile of the asset. This means that the inspection frequency might have to change if specific events occur to increase the risk. Being able to understand the risk and having the flexibility to adapt the approach is critical to the overall safety of the infrastructural assets.

3) Scouring is not the only potential problem

The work described here is focused on the activities initiated as a result of the Malahide incident which had a particular focus on scouring related issues. The engineering group is faced with many other potential challenges.

For example, bridge strikes are a big problem, and it can be difficult to fully understand the potential damage which could have been caused to bridges. The group is now working on an initiative to proactively deal with potential asset integrity issues by instrumenting thousands of assets to get a clearer understanding of any damage caused due to issues such as strikes or overloading.

4) Extreme weather events

Changing climatic patterns are likely to influence and test the protocols put in place by IÉ following the Malahide viaduct collapse. An example of this was seen in October 2011 when the average monthly rain for October fell in just a few hours causing flash floods resulting in damage around the Dodder bridge to the south end of Lansdowne Road station8. This damage was found by divers and resulted in the bridge being closed to traffic while the structure was being repaired.

A more recent example relates to the consequence of unusually heavy rainfall, such as that which fell in the Midleton region in October 20239. This excessive rainfall in a karst region caused sinkholes to appear at a rate much higher than normal; such behaviour will need to be carefully monitored in the medium to longer term to ensure that any damage can be repaired prior to becoming a potential safety risk.

Other less visible examples relate to representations which must be made to secure finance to deal with storm related issues, such as increasing the capacity of culverts, modelling embankments to highlight where strengthening might be deemed necessary and adding rock-armour to provide protection at high-risk locations.

Conclusion

The overall IÉ approach to solve the problem might be best described as hybrid, incorporating aspects of both a 'top-down' and 'bottom-up' approach.

From a top-down perspective, IÉ management increased the resources (money/manpower) available, put new processes/standards in place and provided a new approach to training as well as augmenting existing systems (IAMS as a key element in the management of data, info aiming to reduce risk).

On the other hand, the complete re-baselining of all assets with a detailed root and branch inspection was very much bottom-up and was really crucial in giving IÉ knowledge of their assets to provide the required confidence in the safety of their infrastructure.

Aligning these changes, personnel working in the field took a more proactive approach to recording details in IAMS, to implementing new standards and they were given more responsibility in highlighting possible issues up the corporate chain of command.

The sense of moral duty to the public to ensure that the rail network is safe for all passenger helped to overcome some initial resistance of employees to implement the required changes and their overall effectiveness has become a source of pride for, and credit to, all employees involved.

In summary, IÉ now has a formal competence management system in place to ensure all structures are only inspected by competent engineers. The cyclic undertaking of these inspections is ensured through its asset management system where an alert is received if a bridge exceeds its inspection timescale and dedicated underwater scour inspections are cyclically in place for all relevant bridges, again with the asset management system tracking compliance.

The rail network is of great strategic importance to Ireland and likely to become even more so in future as we strive to meet our sustainability obligations by facilitating an increased and more efficient use of public transport to reduce our carbon emissions while simultaneously servicing a growing population.

Bridge failures in other jurisdictions such as the US10 and Norway11 in 2023 highlight the importance of rail safety and the need for asset management in the context of public infrastructure. The approach adopted by IÉ since the collapse of the Malahide viaduct will help to ensure that the people who need to know really know what the company knows.     

Authors: Kevin Delaney, TU Dublin and Diarmuid Cowhie, CIÉ. This article was written in conversation with Cathal Mangan, chief civil engineer of the company, and who was heavily involved in the response to the collapse. Dr Kevin D Delaney lectures and researches in the areas of design; innovation; knowledge management; and engineering education. He is an educational developer for the Enterprise Academy, a Human Capital Initiative Convene Project at TU Dublin with a particular focus on the areas of creativity, human centred design and innovation for enterprise. Previously he spent 10 years in advanced development and design roles in global world class companies in the automotive and connector industries. He is a Chartered Engineer and Fellow of Engineers Ireland and is the chairperson of the Mechanical & Manufacturing Division Committee in 2024-25. Diarmaid Cowhie works with the CIÉ Group, of which Iarnród Éireann-Irish Rail is a member, and previously worked with Network Rail. He spent 10 years working in industrial automation before pursuing a career in finance. He is a Chartered Engineer and is a member of the Mechanical and Manufacturing Division’s Committee.                                                          

References

1) Railway Accident Investigation Unit (2010) Investigation Report No 2010-R004: Malahide viaduct collapse on the Dublin to Belfast Line on August 21, 2009 [Online].  Available at: www.raiu.ie/download/pdf/accident_malahide.pdf (Last Accessed October 2023)

2) https://www.crr.ie/about/ (Last Accessed March 2024)

3) Prendergast and Gavin (2014), A review of bridge scour monitoring techniques, Journal of Rock Mechanics and Geotechnical Engineering 6, 138-149.   http://dx.doi.org/10.1016/j.jrmge.2014.01.007 (Last accessed March 2024)

4) https://onlinepubs.trb.org/onlinepubs/nchrp/docs/NCHRP20-68A_15-02.pdf

5) Infrastructure Asset Management System (IAMS) refers to the internal GIS system used by Iarnród Éireann-Irish Rail in the mapping and management of assets such as points, bridges, embankments and culverts. https://www.crr.ie/assets/files/pdf/safety_audit_irish_rail_2007.pdf

6) https://www.tii.ie/tii-library/conferences_and_seminars/nrc/nra-nrc-2016/3_3-Paul-Doherty-Catherine-Joyce-GDG-Irish-Rail-TII-Conference.pdf (Last accessed March 2024)

7) Personal communication D. Cowhie (2024).

8) https://www.independent.ie/irish-news/flood-damage-forces-dart-services-in-dublin-south-to-close/26786477.html (Last accessed October 2023)

9) Midleton flooding news report reference: https://www.independent.ie/videos/flooding-on-cork-rail-line/a1173307656.html (Last accessed March 2024)

10) https://apnews.com/article/montana-bridge-collapse-train-dbd905bac1b35a63107213db7983bafc (Last accessed March 2024)

11) https://www.newcivilengineer.com/latest/norwegian-steel-truss-rail-bridge-as-storm-hans-wreaks-havoc-for-nation-16-08-2023/ (Last accessed March 2024)

*It is noted that the RSC became the Commission for Railway Regulation (CRR) on February 29, 2016, following its designation as a regulatory body under EU law.