Accident Investigation Analysis
[In collaboration with Sam Leslie, Eric Sikorski and Michael Waltrip]
Challenge:
Accidents occur when hazards escape detection during preventive measures, such as a job or process safety analysis, when hazards are not obvious, or as the result of combinations of circumstances that were difficult to foresee. Choose a relatively well-documented accident/incident and apply one or more of human error taxonomies to describe and classify the antecedent-contributing factors. What happened? What was the extent of the resulting damage or loss? In the remainder of the paper, document the known antecedent conditions. What type of human error was involved? Identify errors at as many levels of the organization or system as you can (i.e., operator, supervisory, organizational climate, policy level, etc..). In your analysis, consider factors such the physical, environmental, administrative, or process hazards, the need for new or more extensive safety training, or unsafe work practices. Additional factors to consider are the level of workload, personnel selection and training. Was the operator a novice or expert? Finally, describe the lessons that can be learned from this incident? What changes (if any) can or should be implemented to reduce the likelihood that a similar event will occur in the future?
Solution:
Background
The accident was between a vessel called the Windoc and Bridge 11 on the Welland Canal in Allanburg, Ontario in Canada at 2054 EST. The vessel was traveling downbound on the canal and was going under Bridge 11 when the bridge was prematurely lowered making contact with various parts of the vessel. Both the bridge and the vessel saw significant structural damage. The vessel then drifted downstream and caught fire approximately 800 meters away from the bridge, however this review will primarily focus on the collision and not the fire. There were no injuries or casualties reported from either the collision or the fire.
On his scheduled day off after two 12 hour shifts, the Bridge 11 Operator woke up at 0800 EST and took two pain pills (Darvon-N) for back pain. Around lunch time the Bridge Operator consumed two to four glasses of wine. He received a call between 1300 and 1400 EST from the St. Lawrence Seaway Management Corporation (SLSMC) team leader asking him to work an overtime shift that evening, to which he agreed. He did not report, nor was not asked, about whether he was fit to work. He rested at home before driving to work and arriving around 1820 EST.
Traditionally, there is a Limit of Approach buffer of 60 meters where the vessel waits until the span of the bridge is entirely raised. Once the span is raised, the lights switch from red to green, indicating to the boat captain that it is safe to proceed. Between 1847 and 1941 EST, the operator had successfully maneuvered the bridge on two occasions allowing four vessels to travel underneath. The operator has high visibility from the control room window and is able to see the highest point of and the back of the vessels that go under the bridge.
At 2050 EST, the operator called the Traffic Control Center to report the vessel Windoc going under the bridge. At the time conditions were good; visibility was high, there was no precipitation, winds were mild, and the water speed under Bridge 11 was less than one knot. After the call, the Bridge 11 Operator lowered the span after reportedly seeing the vessel clear the bridge. The bridge has a range of motion of 32.9 meters between fully-raised and fully-lowered and it generally takes 1.5 to 2 minutes to switch between states. In this situation, the bridge had been lowered 18 meters, taking between 49s to 65s, before the collision occurred. The vessel was about halfway through going under the bridge at five knots when it made contact with the lowering span. The collision resulted in damage to the wheelhouse, main mast, engine room vents, and funnel for the vessel. The bridge was also damaged, both in the center of the vertical lift span directly from contact with the vessel and on the extremities of the span resulting from the bowing of the span. There was no damage to the environment as a result of the accident. Further, there were no injuries to personnel resulting from the collision or subsequent fire.
Analysis
The Transportation Safety Board of Canada’s official Marine Investigation Report (2001) provides a comprehensive review of the Windoc bridge striking and subsequent fire on board the ship. This report will focus on the primary incident of the vessel striking the prematurely lowered bridge. The human error of lowering the bridge too soon was an error of commission on the part of the Bridge 11 Operator. Based on the report and the circumstances surrounding the Bridge Operator’s actions, we can infer that he made a rule-based mistake according to Reason’s error taxonomy (Wickens, Hollands, Banbury, & Parasuraman, 2013). The choice to lower the bridge was based on the same “if-then” logic that was applied by the Bridge Operator on two occasions preceding the accident during his shift. The logic was, ‘if the vessel has passed under Bridge 11, then lower the bridge’. In this case, however, the perceived ‘if’ conditions that triggered the ‘then’ action were not met by the environment; the Windoc had not passed under Bridge 11 despite the Bridge Operator’s perception that it had.
In terms of the Five Operational Levels of Human Error presented in lecture, there are a number of factors at the working level (i.e. Bridge Operator) cited in the report that resulted in his active error. He was working on his scheduled day off, he had previously worked two 12-hour day shifts, he took pain medication the morning of his shift, and he consumed wine with lunch on the day of the accident. The Bridge Operator’s lack of recall of the accident make it difficult to determine if he perceived that the vessel had cleared the bridge or whether he made the wrong decision to lower the bridge even after seeing the vessel had not cleared then failed to correct his action. However, based on the Bridge 11 Operator’s phone conversation with the Traffic Control Center (TCC) stating the vessel had cleared (i.e. “was under”) the bridge, it seems that he perceived the vessel as being clear of the bridge. Though the working level Bridge Operator error is prevalent, is important that we look beyond the bridge operator to the latent factors at other operational levels that resulted in him being in the position to make such a critical mistake.
At the organizational level, there was no system in place to ensure that bridge operators were competent and fit for duty despite the position being classified as safety-sensitive by the SLSMC. The Bridge Operator had limited interaction with other organizational personnel on the night of the accident and those he did interact with were not trained to recognize or ask about his fitness for duty. With bridge operators working alone and the lack of a specific procedure for shift handover, “there was little opportunity for coordinators and managers to observe specific employees for competence and fitness for duty”, and there were “few formal procedures for monitoring performance and safety between peers” (Transportation Safety Board of Canada, 2001, p.29). Further, the SLSMC had no record of the Bridge Operator taking prescribed narcotics for back pain due to a previous on-the-job injury as organizational protocol did not require employee self-reporting of prescription drug use.
At the design level, the main issue is related to communication. According to the Transportation Safety Board of Canada (2001) report, noise levels in the bridge prevented the Bridge Operator from effectively monitoring VHF radio communications. Though the Vessel Master did not properly identify himself or the bridge in question as it was being lowered, the assessment was that such communication would not likely have been heard by the Bridge Operator, and even less understood, due to the machine noise level being higher than radio volume. The Vessel Master did blow the ship’s whistle several times and though it should have been heard by the bridge operator (Transportation Safety Board of Canada, 2001), that is not certain due to the high noise level of the bridge.
The TCC was able to see vessels passing under Bridge 11 through a fixed camera mounted on top of the bridge on a monitor from a remote location The TCC Operator also had direct communication with the Bridge 11 Operator. In phone communication with the TCC, the Bridge Operator stated the Windoc was “under the bridge” meaning that the vessel had cleared the bridge though this is a potentially confusing term especially under conditions of impairment or high stress. The system did not to allow the TCC to override the action (i.e. cessation of bridge lowering) of the Bridge Operator despite the TCC having a view of the bridge remotely and direct communication that provided evidence he had taken incorrect action and was incoherent. Finally, there were no sensors in place to indicate the vessel had not cleared the bridge which could have been tied to warning signals in the bridge operator’s station or, potentially even more effective, an automated system with an override feature to stop the bridge from lowering.
The SLSMC is a not-for-profit corporation of Seaway users and other interested parties under agreement with the Government of Canada. The SLSMC is responsible for Canadian operations and Seaway structure maintenance. All regulatory authority of the Seaway falls under Transport Canada. At this regulatory level, the two main issues were with medical/fitness for duty assessments and personnel certification. At the time of the accident, there was no regulation for SLSMC personnel in safety-sensitive positions (i.e. the bridge operators) to undergo routine medical assessment nor was it required for personnel to self-disclose the use of prescription medication (Transportation Safety Board of Canada, 2001). At the time of the report, “there (was) no regulatory requirement for the operator of a lift bridge to be certified”. Though the Bridge 11 Operator had received two weeks on-the-job training, it was 12 years prior to the accident, four of which were spent in a different role as a Lock Operator. No mandatory re-training occurred following his return to bridge operations.
Recommendations
Following from the above-mentioned errors at the operational levels, it is important to analyze ways to reduce the probability or nullify the possibility of such situations repeating again. At the working level, one of the recommendations is to restrict the number of hours an operator can work within a set period of time. This prevents exhausted employees from being on the job at such a safety-sensitive position. To complement this recommendation, the organizational level leaders should implement a requirement of at least two operators in the operating room, similar to air traffic controllers. This change would reduce the operational reliance on one person’s fitness to perform on a regular basis.
The design level requires significant attention while analyzing this accident. The glaring ineffectiveness of the whistle to convey the emergency to the bridge operator has to be addressed. One recommendation is to use an alternative method, such as an electronic signal, to convey an emergency while also reducing the noise inside the operating room by padding the room walls and/or reducing the functional noise of the bridge gears. Also, the terminology used in communication in this setting needs to change. The term “under the bridge” should not be used to convey that the vessel has cleared the bridge. In an emergency, such as this, it could cause confusion. To add on to the above recommendations, on the regulatory level more safety training and assessments need to be implemented to assess performance and fitness for work at such a critical safety position.
References
Transportation Safety Board of Canada. (2001). Marine investigation report: Striking and subsequent fire on board Bridge 11, Welland Canal and Bulk Carrier Windoc (Marine Investigation Report No. M01C0054). Transportation Safety Board of Canada.
Wickens, C. D., Hollands, J. G., Banbury, S., & Parasuraman, R. (2013). Engineering Psychology and Human Performance (4th ed.). New York: Routledge.