Airlines' 'padding' of schedules

[Melburnian1]

I think airline schedules are padded but necessarily so. It would be impossible to arrive on time to an arrival slot and gate, if some contingencies are not built in to a timetable to account for the multiple variables that beset an aircraft operation....

And yet many flights, consistently manage to pick up time on a gate-to-gate basis.

As the erudite article explains, building in 'contingency' time makes the airlines even less efficient, and costs you and I when travelling more wasted time.

 

[Quickstatus]

And yet many flights, consistently manage to pick up time on a gate-to-gate basis.

As the erudite article explains, building in 'contingency' time makes the airlines even less efficient, and costs you and I when travelling more wasted time.

But without contingency time (which is a better term than"padded") flights would be arriving late. Im not sure which is better.
I should correct myself that "Padded" for me would be unnecessary extra added over and above contingency time - so in a way its not really padded if a flight only has "contingency time". To work out the contingency time I think @jb747 is correct, the airline uses some metric to calculate how much of that time to build into a schedule. But apart from that I don't think unnecessary time is padded in.

The gate to gate time does not really tell the full story either. Ive been on flights were we were early at the gate, but there was no one to operate the air bridge. So did we arrive early but we disembarked late?.
Shorter gate to gate can also be explained by more favourable weather, and traffic and ATC directions , or is just a manifestation of statistics
I think this requires an analysis at the statistical level as @jb747 alluded to:

Pick a flight - say QFxx and monitor the actual gate to gate (GtG) performance over a period of a year.
And you observe that over the 365 occurences the shortest GtG time is 13 hours and the longest is 13 hours 30 min (excluding the outliers due the diversions, and airport delays) and assume that the scheduled is 13hrs 15min

However you then observe that 50% of the flights have a GtG that is within 5 minutes of scheduled (either faster or slower). and a further 25% have a GtG which is within 10 minutes of schedule (either faster or slower) and the other 25% within 15 minutes of schedule (either faster or slower).

So you also observe that 50% of flights would have a GtG faster than 13hr15min and 50% with a slower GtG.

Statistically I would say that a flight is padded (with unnecessary extra time) when all or most of the actual GtG is faster than scheduled.

Statistically some flights will be faster on a GtG basis than scheduled. That is just the variability of flight operation becoming visible.

Should all flights be scheduled at the shortest actual GtG (based on actual historical performance)?. If so then most flights would be late depending on the definition of "On time". And what irks business travellers more is being late.

Lets say we do schedule at the shortest actual GtG. That means that to improve the performance so all the flights stack up at at schedule time, the aircraft has to fly at a cost index which is fuel inefficient, possibly displacing freight and passengers weight to accomodate the greater fuel uplift - especially on those long range flights. It may actually be impossible for every flight to meet that GtG time due to fuel:weight constraints. Im not sure that is good by any analysis.

I would rather each flight operate at the most fuel efficient settings and at the least cost (least cost and most fuel efficient may be competing outcomes) and build into the schedule an appropriate contingency time based on actual past performance.

 

[jb747]

Some aircraft have flight computers which targets the flight according to selected outcomes. Some of those outcomes might be minimal flght duration or lowest cost. Minimising flight duration may be possible up to a point but is limited by fuel. Lowest cost target is (simplistically) flying at the most fuel efficient speed and altitude, but ATC might prevent the most efficient altitude. So to a point pilots can save a bit of time here or there, but I understand it is not much. And even if departure or arrival is early, other factors as discussed will throw it all out the window.

I expect that all of the FMCs have this capability. Cost indexes go back at least as far as 1985.

When you load a flight plan, it will have a CI upon which it was based. That affects the fuel burn, programmed speed, and climb points. The crew can overwrite this target if they want to go a bit faster, or save a bit of fuel. CI 0 is roughly max range cruise. CI 500, is get there quickly, but at quite high fuel expense. Most plans were around 60-100. We used to play with the times to see what sort of outcome it gave. I don’t know exactly how the companies define their CIs. It’s a trade off between the cost of running the aircraft, and the cost of feeding it.

Alternatively, you could program a constant mach cruise. This was my preferred method if I wanted to make up some time. Using a CI the speed will decrease as the aircraft gets lighter. That’s more fuel, but not time, efficient. A constant mach will use more fuel, but save more time.

The FMCs are also capable of targeting a time. For instance, I might target the a point if I was going to an airport with a curfew. It would slow me if I was going to be early, or push the speed up if late. The longer the leg, the better it worked, but it could be quite fuel inefficient. ATC also don’t like what appear to them to be random speed changes.

 

[Franky]

I just had a good example if it matters whether schedules are padded, or not, on Good Friday night. To me it means bugger all when Mother Nature intervenes. My wife was due to leave HKG on CX for PER at 1450. I received a text from her saying they had begun boarding on time. At 1600 I received another text saying they were still on the tarmac due to bad weather, which was affecting both incoming and outgoing flights. Take off time was unknown. It eventually left three hours late.

Using what little erudite and scholarly skills I have, I guessed that I would have to manage my expectations about her new arrival time, and decided that as the pilots would not be able to put on a few ‘homeward bounders’ to the throttles, I should expect her flight to arrive at 0130. I then padded my arrival time at the airport to pick her up at the door at the optimal time of 0155, which worked a treat.

To me, there are obviously too many variables to expect all flights to depart and arrive on time all day, every day, and I’m happy that what we are offered in the scheduling is about as good as it can be, given these variables.

 

[QFflyermelbourne]

Airlines have sufficient data accumulated from years of flying to know typical flight times for a month, or seasons.

I really don't see the problem here. I depart Melbourne a lot and if all goes well you can depart the gate and be in the air within 5 mins. But I've also regularly had it where long taxi or holding has meant it is 30 mins or more before we've even got airborne.

Now - in your long haul where a plane might fit only 1 or 2 flights in a 24 hour period you can live with getting hit with a surprise delay such as that. But assume it's a work horse 737 that has six or seven flights to complete in a single day. If the conditions prevail and the same aircraft is passing through Melbourne (for example) a couple of times it could easily end up running 90 mins or more late by the end of it. Allowing padding keeps the whole operation within a fairly close range of the day's timetable. If you are booked on any of the later flights you should be very pleased about this.

If this wasn't necessary you can be sure airlines wouldn't be doing it. Saving 20-30mins off every domestic flight could allow them to squeeze an extra sector per day out of an aircraft on a domestic pattern. You think they wouldn't prefer to do that if they could.

I for one am very happy with the planning and published times. I just wish our regulators and airport owners would build decent infrastructure for the 21st century that would allow much more efficient operations.

 

[Melburnian1]

I really don't see the problem here....Now - in your long haul where a plane might fit only 1 or 2 flights in a 24 hour period you can live with getting hit with a surprise delay such as that. But assume it's a work horse 737 that has six or seven flights to complete in a single day. If the conditions prevail and the same aircraft is passing through Melbourne (for example) a couple of times it could easily end up running 90 mins or more late by the end of it. Allowing padding keeps the whole operation within a fairly close range of the day's timetable.

If this wasn't necessary you can be sure airlines wouldn't be doing it. Saving 20-30mins off every domestic flight could allow them to squeeze an extra sector per day out of an aircraft on a domestic pattern. You think they wouldn't prefer to do that if they could....

I've not suggested that 'every domestic flight' could have 20-30 minutes shaved off it. You're exaggerating.

Read the article. Having recovery time - padding - leads to further inefficiencies. This is costly for passengers and airlines alike, even though you may believe that the latter is counter-intuitive.