Airport capacity test

[mickael]

I wanted to do this test for a long time, and finnaly I found few hours. Im not sure if anybody did it before, but I hope it will be usefull, or at least a little bit interesting.

There may be many translating mistakes so sorry for that, and please someone correct me if it is annoying for him (under this post or by pm)...

At the beggining -major informations about the in-game airports:



Legend to the graphs (all indicators, except the first one, are made for the maximum load of the airports):
"time on apron" - its the time needed to complete unload and full upload of the plane (passangers+mail) - its always 10 seconds
"time on airport" - period of time which begins after landing (when the plane leave runway strip) and ends when it returns to the runway strip again (time spent on apron is included) - this is average value based on 100 observations.
"time on empty airport" - this is average value based on 5 observations. The testing plane is the only one on the airport.
"starts+landings/month" - sum of starting and landing planes per month. This is average value based on 24 observations.
"bandwidth" - sum of imported and taken passangers and bags of mail in a year. The averange capacity of the planes included in-game is 150 + 30. There is a plane which have such a load - Bakewell Luckett LB80 - so I used it to the tests.
"plains on airport" - average amount of planes which are at the same time on aircraft. Based on 100 observations.



It was hard to calculate the values for small, commuter and city airports. Despite of that the option "discable servising when breakdowns set to none" was on, the planes on this airports were using hangar. So the value for this airports on the "time on airport" graph may be not accurate.
P.S. for all time axis - the unit - seconds.



As I said: its for Bakewell Luckett LB80. Of course you can count it for any other: 12months*starts&landings*capacity.
And of course Im aware of that there is no way to rich such a bandwidth on small and commuter airport without cheating. That's only for comparison.



If someone want to check this indicators for other airports (f.e. from NewGRF Airports) - you can download .sav file- which will make it easier.

Final remarks:
-despite the fact that I used the cheat "jetplan will not crush" and option "number of plane crush: none" I counted many disasters.
-I know that no one will use as many planes as I did to the test - but probably smaller amount of planes will give the same - not good looking -effect (same plan route)



-its very suprising that international airport proved to be more efficient then the intercontinental.

[Lordmwa]

That is very interesting, thanks

I am surprised by the results and maybe a small change needs to be done to make it "correct"

[juval]

Hmm i find this very intresting, this maybe can help and fix the problems with the ineffective intercontinental airport,

I wonder have you made any observations on what might cause the delay on time at the airports?

I know 1 which is very disturbing, 1 of the intercontinental airports depot is turned the wrong way, which means that planes has to go around it to enter it! This is something i think must be fixed to get a more effektive airport, because i guess te purpose of bigger airports is to handle more planes efficiently

[Wasila]

So city carries less planes than small and intercontinental carries less than international? Or just less efficient?

[mickael]

I wonder have you made any observations on what might cause the delay on time at the airports?

I know 1 which is very disturbing, 1 of the intercontinental airports depot is turned the wrong way, which means that planes has to go around it to enter it! This is something i think must be fixed to get a more effektive airport, because i guess te purpose of bigger airports is to handle more planes efficiently

I cant say much about depots, because I tried not to use them. But in that case Im cuite sure youre right - it makes the aircraft taxilane much longer. The plane after landing goes to the depot - then back to apron - and to the starting strip: passing close to the depot.

The intercontinental airport can be cut into two pieces - left and right. Each side should be as efficent as metropolian airport. Both metropolian and 1/2 intercontinental airports have two runway strips, but the taxline by the first airport is much shorter (first reason)- not mentioning that the same amount of planes can be in move at the same moment. Longer taxline-> more planes should be at move(second reason). In results we see - planes spend too much time on intercontinental airport.

Edit: even the "time on empty airport" graph shows how much more time plane needs to go through intercontinental in comparison to other airports.

I found also another 'thing' (look screenshot). I think the citi airport would be more efficient if the plane 1 would start from the airport before plane 2 lands. But its only my supposition...

So city carries less planes than small and intercontinental carries less than international? Or just less efficient?

The airports are carring less plains - because they are poorly efficient. The international cope with plane faster - so it can handle more. With the city and commuter - truth is - we shouldnt compare them - the probabillity of a crash is much higher for commuter airport - thats 1 reason, and second: as I said the values for the three smallest airports may be not accurate.

[planetmaker]

Wow, very nice analysis and presentation you made there. Kudos! Something which I felt for long needed written down somewhere Can you maybe add it to the official OpenTTD wiki? It would find more reception there than in a forum thread which gets burried after some time.

[Wasila]

Still a little confused. So is an international airport actually better than an intercontinental to handle traffic? When I say efficient I mean traffic handled compared to footprint.

[juval]

i agree this should be posted there for more readers and this should be a projekt to make more efficiant as the example you gave in you screenshot

[mickael]

Still a little confused. So is an international airport actually better than an intercontinental to handle traffic? When I say efficient I mean traffic handled compared to footprint.

I can say that it is obvious. The only requirement is that i didnt made any mistake, what im pretty sure. To confirm that - you can test it by yourself.

Can you maybe add it to the official OpenTTD wiki?

First - someone have to confirm what I posted, maybe improve this.
Then someone have to correct this text, and maybe rewrite it, couse my language knowledge is below zero, as I suppose.

[planetmaker]

Can you maybe add it to the official OpenTTD wiki?

First - someone have to confirm what I posted, maybe improve this.
Then someone have to correct this text, and maybe rewrite it, couse my language knowledge is below zero, as I suppose.

Well, your results "feel" correct from my experience without having done the explicit statistics myself. Also, your English seems quite well and a wiki allows correction once the text is there.

[oklmernok]

I make my own test scenario. I use the same LB80 air plane however my scenario is not so crowded as your one.
my results are (landing+take off):

commuter: 26 /3 min.
city: 17 /3min
metropolitan: 24 /3min
international: 47 /3min
intercontinental: 51 /3min

The bandwith is not so simple. it depends on the planes (small - large, slow - fast)!
The city airport use it's depot as a parking area? If the airport is crowded the planes will stay there for along time.
Is it a bug? (service is OFF)

[Zuu]

When calculating the incoming queue to an airport in the PAXLink AI I include non-stopped aircrafts in the hangar. The queue length is used to determine if an airport can take more traffic or not. If I would not include the hangar in the queue length, eventually the hangar would become over-flooded with aircrafts.

In PAXLink I have to in some comparisons include and in some exclude aircrafts that head to the hangar to be sold. This is probably not an issue for you. The non-stopped in hangar I would include as a measure of the entry queue length.

[mickael]

I make my own test scenario. I use the same LB80 air plane however my scenario is not so crowded as your one.
my results are (landing+take off):

commuter: 26 /3 min.
city: 17 /3min
metropolitan: 24 /3min
international: 47 /3min
intercontinental: 51 /3min

The bandwith is not so simple. it depends on the planes (small - large, slow - fast)!
The city airport use it's depot as a parking area? If the airport is crowded the planes will stay there for along time.
Is it a bug? (service is OFF)

Im not sure - but the game speed probablly depends also on the CPU performance, so in the different situations 1minute can be 'longer' or 'shorter' if you know what i mean. Thats why I used game time - not real time.

If the airport is crowded the planes will stay in a hangar or spend there a lot of time- but its true only for city airport -i mean that by my test it happened only on the city airport and less on commuter.
I dont know if its a bug or not, but it dosnt work properly on havy load.

Btw. Perhaps there is a volunteer, who will post it on wikipedia?

[oklmernok]

Yes this is not an exact measure, however it seems to be the same as your one, except the intercontinental airport. (You asked someone to confirm)
I dont know how you made your count: I use a stopper, I heard the landing planes and watched the take off planes. Counting the intercontinental was funny :)

My "Hangar problem" is not a bug, I can make it all types of airport if I stop some planes just before the take-off.
The city airport has a special landing and hangar preference because of the taxiway crossing. The other airports have better taxiways design so they wont collect planes (take off pref.).

small: 15 /3min -- NO JET! planes (Coleman Count)
city: 17 /3min
metropolitan: 24 /3min
commuter: 26 /3 min.
international: 47 /3min
intercontinental: 51 /3min
where: 1 month = 31 days = 1 min+12 sec

the result:
when you can build any other airport than CITY (for jets) it is better to change (but check your government before demolish)!
The commuter could be better then the metropolitan because it is smaller and has 2 additional separate heliport.

The count could be different if we try it with only small (80) and only with large (400) capacity and a fast plane... who will make it?

[oklmernok]

So, I used these planes:
slow: Coleman Count (biggest no-jet) 65pass+8
small: YAa46 (80pass+10) smallest jet-plane
large: Dinger 200 (400pass+80) largest plane
fast: Dinger 1000 2,346 km/h (130pass+45)
the slow plane was tested only small airport.
option "number of plane crush: NONE"
option "disable servising when breakdowns set to none: ON"
airports were full with passang. and mail

if a hangar is collecting plane then I used the "c"-code (more landing than take-off)
the value is the number of take off+landing /3 min:

_______________slow__80____400___fast
small airport:_____15_
city airport:_______x___21c___17c___17c
metropolitan:______x___29____19c___25
commuter:________x___29____25c___25
international:______x___49____48____55
intercontinental:___x___51____39c___46
where: 1 month = 31 days = 1 min+12 sec

It shows that the best airport for me (and for Mickael) is the international.
You have to know that the city airport count is: 8 takoff+13 landing (80) and 6 takoff +11 landing (400) and 7/10 (fast) / 3 minutes. So it is really collect the air-planes. The metropolitan rate: 9/10 (400), the commuter rate: 10/12 (400) and the intercontinental rate: 23/25 (400).
If an airport collect planes then there will be more traffic on the ground, so that's the reason for the less value.
And as you know: this is only one counting.. please confirm it. If we have more measure we can make better statistic.
If your plane often crashed during the landing there is only one good airport: the intercontinental (the international has 2 runaway but one is dedicated to take-off and the other one is for landing)