World’s first FOD detection system at YVR

Posted: 3 April 2007 | Brett Patterson, Director, Operations Safety and Planning, Vancouver Airport Authority | No comments yet

In March 2000 an A330 departed YVR’s runway 08R shortly after 8pm. During its takeoff roll, and unknown to the flight crew, the port engine cowling fell off the aircraft and shattered into hundreds of pieces down one side of the runway. The flight crew of the tenth aircraft to use the runway reported seeing some debris on the runway, at which time airfield operations staff responded and reported the debris which initiated a clean-up procedure whilst simultaneously contacting the crew of the A330 to advise them of the situation.

In March 2000 an A330 departed YVR’s runway 08R shortly after 8pm. During its takeoff roll, and unknown to the flight crew, the port engine cowling fell off the aircraft and shattered into hundreds of pieces down one side of the runway. The flight crew of the tenth aircraft to use the runway reported seeing some debris on the runway, at which time airfield operations staff responded and reported the debris which initiated a clean-up procedure whilst simultaneously contacting the crew of the A330 to advise them of the situation.

Later that same year, the crew of an A340 landing on the same runway after 11pm reported running over something that resembled a piece of wood. On inspection of the runway following the incident, airfield staff found the pogo-stick from a DH8 aircraft, which was later found to have been dropped on the runway by the first aircraft to use the runway after a visual surface inspection by airfield staff found the runway to be ‘bare and dry’.

On the heels of these experiences, and under the leadership of Craig Richmond the airport’s Vice President Airport Operations, the airport’s Airside Safety Officer and Manager, Airside Operations were tasked to look for technology that would enhance runway safety with respect to Foreign Object Detection (FOD) management.

A review of current runway FOD management practices revealed that by following ICAO recommended practices for runway surface inspections we were aware of the status of these surfaces for less than 0.5% of the time they were operational. We wondered how many organisations, outside of airports, would accept knowing the status of their primary production line for such a small fraction of its operating time.

In 2001 we began a dialogue with QinetiQ about the use of millimetre wave radar as a means of detecting FOD on a runway surface, after reading of their technology in an industry journal. Our technology search neither began nor ended with QinetiQ, however, as we also met with a number of organisations proposing the use of other technologies including CCTV systems and laser radars. QinetiQ was the only company and millimetre wave radar the only technology that we were able to find that has proven its ability to detect small objects on a surface as large as a runway. Thus began the journey that led to the installation of the World’s first Automatic FOD Detection System at Vancouver International Airport in May 2006.

An overview of the system and how it works

YVR’s Tarsier System consists of a total of four radar sensors with two sensors scanning each of the airport’s parallel east-west runways (3350m x 61m and 3030m x 61m respectively). The radars are mounted on top of towers that are located between each runway and a parallel taxiway. As the radar works on a line-of-sight basis, each radar tower differs in height on the basis of the relationship between the height of the crown of the runway and the level of the field it is located in. In all cases the towers are located outside of the runway and taxiway strip and under the inner-transitional surface for the runways. In addition, the towers are located between 900 and 1,100m from each runway end which results in approximately 70 per cent of each runway having dual radar coverage.

All four radars are networked to a single user display which is located in the Airport Operations Centre where it can be monitored 24/7. The system operates much like the fire alarm system in an airport terminal in that it works silently until such time as it detects FOD on a runway surface or it diagnoses a fault in its own detection capabilities. In either of these events, the system emits an audible and visual alarm to alert users.

The user display screen is divided into three sections: a system status box, a map display and an event log. The system status box provides the user with information on the status of each radar sensor and of the communications link from the sensor to the display. The map is a graphical representation of the airport’s runways and taxiways with the locations of the radar sensors and, most importantly, an icon is displayed at the location of any FOD the system detects. The event log reports any system problems, but more importantly it displays information on each FOD incident including the date, time and GPS coordinates of the FOD to an accuracy of just three metres.

When the system detects FOD on a runway surface it issues an audible alert, displays an icon on the map at the precise location of the FOD and an event record is automatically created. The user then acknowledges the alarm and is able to silence the audible portion of the alarm for a prescribed time period whilst simultaneously alerting airfield personnel to respond and retrieve the FOD. Neither the event record nor the map icon will clear themselves until the radar system no longer detects the FOD on the runway surface. Airfield response vehicles are equipped with GPS moving map displays onto which responders can input the coordinates of the FOD to calculate the shortest route to retrieve the FOD. Once the FOD is retrieved, system users are able to input information on the event including identification of the responders and type of FOD found into the event record for later review and analysis.

How the system has performed to date

As the first system of its kind in the world, we understood at the outset that there would be a lengthy commissioning period during which time we would need to be working in close cooperation with QinetiQ to attain our desired system performance standards.

The first phase of operations, the commissioning phase, saw the system being operated ‘behind-the-scenes’ in an engineering mode only following the completion of its installation in May 2006. It was during this period that we learned more about the detection capabilities of the radars in different operational and environmental conditions – including rain and snow. Some of the technical issues we identified and began to address were sensitivity, flashes and scintillation while operationally we developed an initial concept of operation and made changes to system operating parameters to deal with weather conditions – namely snow.

Sensitivity was an issue, but definitely not a problem, as we originally found that the radars were alarming on very small objects which posed no threat to aircraft safety – in particular the system was alarming on debris as small as a 4″ stalk of grass. Once this issue was identified it was simply a matter for the engineers to de-sensitise the radars, which they were able to do successfully. The system’s required performance specification, which it has met, is to detect an object the size of a 2″ bolt from a distance of 1,000 meters (for those that are radar savvy the required detection capability is a -20 dBm2 sphere at the noted distance).

Flashing is in reference to very large signal returns attributable to the radar beam reflecting off of buildings and other large stationary objects on the airfield. If uncorrected for in software, these could lead to false alarms. Over time and with operating experience the software developers and radar engineers at QinetiQ have been able to identify and distinguish the signal returns of these buildings from the FOD signals we are aiming to detect and they have progressively built the necessary algorithms to reduce these false alarms.

We’ve experienced the phenomenon of a false alarm occurring when the radars scan through a moving propeller blade (caused by range-Doppler coupling). Again, through operating experience the software developers and engineers have been able to build algorithms to successfully address this phenomenon.

During the commissioning phase we were able to observe the radar system performing during a major snow event for the very first time. Those familiar with Vancouver will appreciate that we don’t often get snow, especially 34cm in one storm. As a result of this experience, and the fact that the system will detect and alarm on accumulated snow and ice on the runway, we requested a change to the detection area for snow events. Specifically, with the click of a button we can now reduce the radar detection zone on the runways by 10m on the sides and ends. The goal of this modification is to eliminate the alarms caused by the residue left from ploughs, brooms and blowers conducting snow removal operations. Since this late November snow event we have not had a significant snowfall to fully test the solution, however we have enough information to realise that we will also need to reduce the sensitivity of the radars during snow events to eliminate alarms caused by snow residue. This is something that we are currently working with QinetiQ on and will have in place before the next winter season.

The operational integration of the radar system is an evolving process. Our current practice is to immediately close a runway once we are alerted to the presence of any FOD on it. From a practical standpoint this means that aircraft departures are halted while arriving aircraft inside of approximately 2-miles from touchdown are permitted to continue to land while those outside of this range are sidestepped to the parallel runway, when practicable, or vectored around.

Recognising that we had a great deal to learn about integrating this system into our operation and about the true FOD environment of our runways we have developed an initial concept of operations in which we do not close a runway based on receiving a detection alarm from the radar system, although we continue to close the runway if the FOD is reported from conventional sources. In this initial concept of operations we interact with the system, as described earlier, and respond to FOD alarms on a first priority basis between aircraft movements. In this manner, we have the enhanced safety benefit of continuous surveillance of the runway by the radar system while we continue to learn more about our runway FOD environment and the radar systems performance capabilities without placing undue strains on runway capacity.

On January 25th we placed the system into operational use and to date our list of FOD items detected includes dead birds, an aircraft fuel cap, an aircraft static vein, tire remains, runway sealant and a large bolt and washer. With less than sixty-days of operational experience we expect we have a lot more to learn about the performance capabilities of the system and how it can best be integrated into our operation. We have made one recommendation to enhance the system and that is to add a high-performance camera to each radar tower to support the development of a more risk-based approach to FOD detected on the runway. Development of this camera system upgrade, slaved to the standard radar system (i.e. it would automatically zoom in on FOD detected by the radar system), is well underway by QinetiQ.

While the original goal of acquiring this technology was to enhance runway safety, we now know that there are significant efficiency gains from operating the system as well. Specifically, once the system proves its capability and reliability over an extended period of time there should be an opportunity for airport operators to work with regulators on the elimination of some of the daily visual runway surface inspections. With an average runway visual inspection taking approximately three minutes to complete, this could equate to gaining between three and five runway slots back in a peak operating period. In addition, we have calculated the average time it takes to respond and retrieve FOD reported by flight crews at between eight and nine minutes, with much of that time taken in searching for the FOD. Compare this with our experience with Tarsier, whereby responders are given the precise coordinates for the FOD and can proceed straight to the location, retrieve the FOD and exit the runway in times that average between four and five minutes. This four to five minute time saving represents a significant reduction in delay, particularly during peak operating periods.

Over the coming weeks and months we look forward to sharing experiences amongst other airports currently installing QinetiQ’s Tarsier system and to accelerating the integration of this system into airport operations.

Send this to a friend