Taxiway deviation study
Introduction
1. British Airports Public Ltd.
carried out a taxiway deviation study at
2. Over 77 000 aircraft taxiing
movements in all weather conditions were recorded. The purpose of the study was
to show that pilots do not deviate significantly from the centre line when
taxiing. The analysis of the data had two objectives. The primary objective was
to estimate the probability of two Boeing 747‑400 aircraft colliding wing
tips when passing each other on parallel taxiways. The secondary objective was
to estimate the expected number of years that would elapse before this
collision occurred. The study also attempted to assess the adequacy of the
separation distances between parallel taxiway runways and between taxiways and
objects, as recommended in Annex 14, Table 3‑1 ECAR 139 TABEL 3-1.
Straight sections of taxiway
3. Based on some 2 000 observations of B747 deviations on the straight sections of taxiways, the study concluded that the probability of two B747‑400 aircraft colliding when passing each other on parallel taxiways is around 10-8, i.e. I in 100 million. This assumes that the taxiway centre lines are 76.5 m apart and that the aircraft have a wing span of 65 m. Upon analysing the data in the large data base which had been established, there was ample evidence to indicate that taxiing aircraft do not deviate from taxiway centre lines to any great degree. The data also provided an estimate of the number of occasions per year that two standard B747s pass each other on Heathrow's parallel taxiways. This would be about 80 occasions per year out of around 34 000 B747 air transport movements. This figure was considered low because, in the majority of cases, the aircraft would be moving in the same direction through the taxiway system. Departing aircraft use one runway and arriving aircraft use another; their paths rarely overlap. Figure A4‑1 shows the distribution of deviations upon which the analysis is based.
Curved sections of taxiways
4. It was considered that the probability of collision on curved sections on taxiways is of the same order of magnitude as in straight sections, i.e. 10-8. There were not enough observations of B747 deviations on the curved sections of taxiways to repeat the analysis detailed for straight sections. Data were collected at two locations. Data from the inner curve proved not to be useful because there was a large paved area to the inside of it which pilots tended to cut across. The number of observations on the outer curve was low because several months of maintenance work had closed that part of the taxiway. Figure A4‑2 shows the distribution of the 185 B747 deviations observed on the outer curve. The negative numbers are deviations from the centre line towards the inside of the curve. This is in a different form to Figure A4‑1 which shows only absolute deviations. Figure A4‑2 shows that almost all aircraft main wheels cut across the corner. For a collision to occur at a curved part of the parallel taxiway, the outer aircraft must deviate to the inside and the inner aircraft must deviate to the outside. Figure A4‑2 shows that the latter is very unlikely to happen. The over‑all distribution suggests that the collision probability would be similar to that of straight sections, i.e. 10 -8. Curved sections of taxiways were regarded as less of a problem than straight sections of taxiways since there will always be fewer curved elements in a given taxiway layout. Thus, the chance that two aircraft pass on a curved section is much less than on a straight section.
Speed effect
5. The analysis showed that an aircraft's spe does not effect its lateral deviation.
Bad weather effect
6. No link could be established between bad weather and large taxiway deviations. During the data collection period, most weather conditions were encountered including snow fall, heavy rain, strong winds and visibility reduced to 1 000 m.
