ATC (approach) question: stack handling

[Jeroen Hoppenbrouwers]

Busy with London Control again, and pushing the limit.

If you have a few aircraft in a stack, do you religiously exit only the lowest one, or can you (when under radar surveillance, of course) manipulate the stack a bit and have an aircraft that happens to be at the right spot in the holding pattern for your string of pearls exit early, while the one below is still running the outbound leg?

Of course taking all vertical separation into account as usual, and not leaving somebody stuck in the stack while everybody else passes straight over him.

The alternative is to put everybody exiting the stack on headings like the usual fan-out for formation breakup. But this is a lot messier. Often, eyeballing the actual stack patterns and picking the aircraft that is accidentally in the best position to make up a string of pearls is a lot easier.

I also prefer to always have an aircraft exit the stack over the fix, so the stack leaving heading and position are the same for everybody.

Hoppie

[cdpohl]

In my days it was 'First come, first serve', unless a problem/emergency dictated otherwise. The aircraft lowest in the stack would be next for the approach. Solely vectoring aircraft for an approach in a very busy environment increases the controller's workload immensely, and often there is not sufficient room to do so, even though speed control is employed.

[tango4]

Well,
I would say that the general rule is indeed first in first out.
But as with most rules, they are not to be followed blindly all the time.
At LFPG, we work a bit differntly from EGLL.
Stacking is not our "normal" way of slowing the traffic. We mostly use what we call "linear waiting" which is slowing down aircraft in advance (using a sequencing tool to help us that communicates with PARIS ACC). But sometimes, this method is just not enough (too much traffic, unexpected CBs, unexpected fog, etc). and we have to add the traditional stacking method.
So basically, I think when it comes to stack management, we are clearly less proficient than our british colleagues.


But now, when you find yourself in a holding situation, it basically means you have much traffic. So the most important rule here would not be "first in, first out", but "making sure you do not waste a slot".
And in some cases, it is much better to have an aircraft exit the stack because he is on the inbound leg, rather than taking the one below in a less favourable position.


At least that is how I do it IRL.


Charles

[Cbf]

Hi tango4,

At LFPG, do you often use RNAV STAR BIBAX 8V and LUKIP 8V?

https://www.sia.aviation-civile.gouv.fr/dvd/eAIP_08_NOV_2018/FRANCE/AIRAC-2018-11-08/pdf/FR-AD-2.LFPG-fr-FR.pdf

Page 205

[tango4]

No, they are not used frequently.
The STARs are the responsibility of PARIS ACC who sends us the aircraft enroute to the IAF for the initial approach.
In the examples you mention the sector TP is responsible for sending us the aircraft sequenced for MOPAR IAF. Normally they use the 8W or 8E Stars (LFPG facing West or East).
The STARS you mention are called "merge point" and are used to theoretically absorb more delay without opening the stacks.
I am not really convinced by the concept personally.
If you want to know how it works, tell me and I'll try to elaborate a bit.


Charles

[Cbf]

Hi Charles,
Thank you for your answer. I am interested by an explanation.
Cbf

[tango4]

Okay, so here goes.
In order to understand the "merge point" you need to understand the way it works normally first.
At LFPG traffic is organized in the following way. You can imagine the LFPG terminal area as a Big rectangle with the airport in the middle.
Departures will flow through the sides (so 4 major flows: North, West, South, East) and arrivals come through the corners (NW, NE, SE, SW) so you have 4 Feeder fixes which are the IAFs for the initial approaches (I simplify things here omitting other fixes used for LFPB/LFPO or Prop aircraft).
So if you come to LFPG in a jet aircraft you will be assigned 3 things:
*A STAR that ends at one of these 4 fixes (NW=MOPAR, NE=LORNI, SE=OKIPA,SW=BANOX)
*Then an Initial Approach (Approach transition for our american friends) that starts at this fix and ends by "Radar Vectors" (except for some night approaches but that is another thing).
*A final approach: most of the time an ILS


One of the jobs of "PARIS CONTROLE" (Paris ACC located in Athis Mons close to Orly) is to deliver aircraft on those 4 fixes. One sector manages arrivals for one fix.
Between two control centers, you have what we call and LOA (Letter Of Agreement) that defines the silent transfer conditions. It means that Paris Control is going to transfer aircraft to us (deGaulle Approach) without any coordination, provided they follow the conditions described in this LOA.


Generally speacking, the idea is that Paris Terminal Sectors will send us aircraft en route to the IAF (actually, one waypoint before for stabilization purposes), with no less than 8Nm between aircraft, no catchup, at a prescibed speed.


So in the example of the sector in charge of MOPAR arrivals, their main job is to merge flows coming from BIBAX and LUKIP STARs.
Now, as I mentioned in my previous post, in rush hour, you cannot just let aircraft come in fast and flow into the TMA. Approach is just not going to be able to absorb this traffic. To cope with that, you have two main solutions (which are not completely exclusive to each other of course):
*SOLUTION1: Holding patterns.
This is the oldest solution, and traditionally the one used in London TMA. You basically let the aircraft come in to IAF at High speed, then enter the stack, and have the approach controller get them out one at a time in sequence.
Main drawback: not very fuel efficient.
Main advantage: you are more precise in feeding the runway, so this avoids losing a slot because you slowed traffic too much for example. In the case of EGLL, that makes perfect sense as this airport is extremely capacity constrained with only two runways. Each slot is precious.
*SOLUTION 2: Linear Waiting
The idea is to slow traffic in advance using a sequencing tool.
This is a tool that we have in the approach room on which we see all incoming traffic in columns (one runway, one column). This tool helps compute the delay for each aircraft. Each terminal sector has the same tool, but only with the view of his aircraft and the associated delay. So you mix this tool and some phone coordinations in order to pass instructions to Paris ACC to slow of accelerate traffic flows.


As I mentioned these methods are not completely exclusive. I guess controllers in London do not let all traffic come in at 340kt in rush hour (and i heard that in the recent years they started using sequencing: but I am not certain of anything here, I'll let the Brtitish PSX team talk here !).
And in LFPG where we use the linear waiting method, we sometimes reach the limit. Past a given point, it does not make any sense to slow an aircraft to 180kt while 80Nm from the field. In those situations, you will definitely open the stacks.


Now, here come the question: how does Paris ACC mixes the flows from BIBAX and LUKIP ?
For example, if LFPG is facing East, you see that BIBAX8E and LUKIP8E are straight lines from BIBAX/LUKIP direct to KOLIV->MOPAR. So their job is to vector the airplanes out of the STARS to have them eat up some miles and in the end get the 8Nm separation. The biggest advantage of that method is that with some practice you become rather precise, but it is quite time consuming (many messages).
The idea behind those "merge point" STARs is to automate that process. You let the aircraft join the arc. When it is 8Nm further from KOLIV than the previous one in the sequence (which at this point in enroute to KOLIV), you just have to say "direct Koliv" and there you are, the aircraft is 8Nm behind the previous one. This is much easier than vectoring.


But that is the theory. In pratice it is very difficult to precisely time your message and when we see Paris use that method, you frequently find yourself with too much separation. And as the flows from LUKIP and BIBAX are flying the Arc in opposed directions wind effects are also harder to integrate.
They talked about using that method to sequence aircraft onto the ILS but I fear this would be even worse. As soon as you have some CBs for example, you cannot do that. And as with manual flying, vectoring requires practice.
The argument I often hear is "you'll be more efficient 90% ot the time. We accept bigger capacity constraint  in case of CB." First of all, that assumes weather forecast are accurate (insert polite laugh...). Many times we have to face aircraft avoiding all over the place without any warning when CBs are developing in the area.
And I really don't think we would be actually more precise with that.
Because another thing you have to take into account is the human factor: when you tell a pilot "FLY HDG xxx", most pilots will do it pretty fast. When you tell "Fly direct to", you will see huge variations between pilots (depending on proficiency/familiarity with the area/fatigue).
I do think that concepts like TBS (Time Based Separation) already in place in London are much better concepts to help you gain capacity (you get aircraft closer to each other thanks to a tool while still retaining radar vectoring proficiency).


Hope that was not too obscure/long !
And keep in my that those are my personal views only of course.


Charles

[Cbf]

Hi Charles,
Thank you for this interesting explanation.
Cbf