Measuring Runway Overrun Risk
Dave Jesse on February 21, 2013

This blog carries on from the previous touchdown measurement and looks at how we can relate touchdown conditions to the likelihood of a runway overrun.

When looking to measure the risk of a possible runway excursion on landing, safety managers have used parameters such as the speed of the aircraft at touchdown, the position of the touchdown point and the distance between the touchdown point and the end of the runway.
These direct measurements all characterise the landing, but are not directly related to the likelihood of overrunning the end of the runway. If the pilot overruns the end of the runway, it must be because he could not achieve the level of deceleration required to stop on the runway.

KPV “Deceleration To Stop On Runway”.

At the touchdown point, the distance from the aircraft to the end of the runway is calculated. With the distance to go (S) and the groundspeed of the aircraft (U) known, the average deceleration required to stop in the available distance (a) is calculated from a=U2/2.g.S.

Features of this KPV.

This is a single value that will increase for:

  • Higher landing speed
  • Landing further past the touchdown zone
  • Shorter runways

Furthermore, the units of the value can be related directly to the available runway friction, µ.
If the available friction is less than the Deceleration To Stop On Runway value, the aircraft will not be able to stop on that runway using brakes alone. Of course reverse thrust or propeller drag will help in stopping the aircraft, so high values do not mean an overrun occurred, but if these devices had not been used correctly or failed to operate, an overrun probably would have occurred.
This KPV will allow safety managers to use a histogram from many flights into a given runway to estimate the probability of a runway overrun given medium or poor braking action and stopping on brakes alone.

“Deceleration To Stop On Runway”.Example


This KPV reflected well the difference between landing on long runways (typical value 0.06) and short ones (typical 0.13), and highlighted a few landings on short runways where the aircraft had landed deep or fast.

After this promising start, we “tested” the KPV using data from an actual runway overrun, and while the deceleration required to stop was higher than in some cases (0.105), it was by no means outstanding when compared to other operations.

The key issue in the overrun case was that the deceleration was not maintained. For too long the aircraft was allowed to roll down the runway without adequate deceleration and when the end of the runway come into view, there was not enough friction to stop, so the aircraft rolled off onto the grass – fortunately without injuring anyone.

A similar characteristic is seen in cases where the crew allow the aircraft to roll down the runway at high speed to use turnoffs close to the end of the runway, thereby reducing the margins for error.

It was therefore decided that a second KPV was needed to reflect the progress of the deceleration, that is, how fast the aircraft was approaching the end of the runway.

KPV “Runway Overrun Without Slowing Duration”

Throughout the landing rollout, the groundspeed and distance from the aircraft to the end of the runway are computed. The time taken to overrun the end of the runway if the aircraft does not slow down further is computed along the runway. The KPV is the minimum time value on that landing.

Features of this KPV.

This is a single parameter that will reduce for:

  • Higher rollout speed
  • Delayed braking
  • Shorter runways

It can be described as the shortest time to reach the grass if the brakes fail.

“Runway Overrun Without Slowing Duration” Example

Note: The original name of this KPV was as shown on the chart, but as the measurement assumes no braking, the new name “Runway Overrun Without Slowing Duration” has been adopted which conforms to the POLARIS KPV naming convention.

Values at the left end of this chart indicate cases where the speed was high close to the end of the runway, and there was (in two cases) less than 15 seconds before the aircraft would reach the end of the runway.

There are about 20 KPVs relating to landing performance in the POLARIS open source library, including these two KPVs. We continue to work with our colleagues in the aviation safety industry to improve our measurements of this important issue.