Hacker News Reader

Comment by esemor | Hacker News Reader

As an airline pilot flying Boeing 757 and 767 I think this concept is really cool but it does overlook some fundamental principles of the unreliability of weather in order to be practical.

This week I have landed in variable winds gusting up to 45 knots and then every inch of wingtip clerance counts, a drawback of this design.

What if the aircraft makes a long landing (floats) and suddely faces a different wind component that subjects it to a tailwind.

Also, the full circle would only be useable on zero wind days which makes me wonder about the economical reality of having a large part of the pavement be unusable most days.

View on HN

I am also a pilot. I'll add to this by explaining a couple of basic principles in flying aircraft.

The #1 reason this is a very bad idea is it is not compatible with flying what's called a stabilized approach, which is fundamental to safe landings. What this means simply is you fly your last (final) leg of approach on a straight line aligned with the centerline at a shallow glide angle. The moment before touch down you reduce engine power and flare (bring the nose up) and stall the wings just above the runway. If you are flying into a crosswind, you crab (fly with the nose angled into the wind to stay on centerline), then kick out of the crab angle just prior to touchdown (you control the rudder with foot pedals). A banked circular runway is totally incompatible with this. If you misjudge your approach a bit and land long, you miss the runway. You have to go from wings level to a banked turn at exactly the right moment. Lots of potential for things to go wrong. It's just a bad, unsafe idea.

Another reason this is bad, higher landing speeds. If you are flying in a banked turn, your wing will stall at a higher airspeed. Heavy aircraft already land fast, and anything that adds to that creates problems, wear on tires and brakes, etc.

These people also seem to be unfamiliar with basic geometry. A "circle-ish" runway configuration would do the job while retaining long straight runways. For example, just arrange eight runways in an octagon configuration, with the airport terminal etc in the center. If you have enough land area to work with, this is easy. The problem, of course, is that land in large metro areas is expensive, so you end up with compromises such as intersecting runways as you see at airports like San Francisco (SFO).

So while this may be fun as a flight simulator challenge, it is a bad, unsafe idea for the real world.

I don't think you need to "go from wings level to a banked turn at exactly the right moment" for this. There are two ways to make it work.

As proposed, the runway is banked. You bank your aircraft well above the runway surface. You fly above the runway, possibly following it as your holding pattern prior to landing. You can keep going around, banked already. When it is your time to land, you continue around in that bank and descend to the runway.

Alternately, it could be unbanked with straight landings. This makes the circle considerably thicker. Landing is quite normal, aside from the runway markings.

BTW, your "kick out of the crab angle just prior to touchdown" method may be standard, but it is pretty bad. The B-52 gets this right, with 4 pairs of wheels that touch the ground at the same time and are all capable of being rotated.

The whole point of a stabilized approach is to avoid unnecessary maneuvers in the final moments of flight, and to simplify decision making. Your approach is a straight in powered glide. If you like where you end up, you cut power, flare and land. If you don't like the situation, you apply full power and climb straight ahead, then turn according to the rejected landing procedure for the runway or ATC instructions.

What you describe needlessly adds complication to the approach, which adds risk, which will certainly result in accidents and fatalities. It's just a bad idea, and there is a reason airport designers never considered this (it is not as if the idea of banked curved roadways is new).

Sorry as a pilot, I can say this is not just a bad idea, it is a lethally bad idea. A fun simulator challenge, but not something that will work in the real world. If you are doubtful, I would suggest you try circling around a point in IFR conditions within 50' lateral tolerance above a circular runway. Be sure to add zero visibility and a 10-20 knot cross wind to simulate doing so in the clouds above the runway as you descend.

> Be sure to add zero visibility and a 10-20 knot cross wind to simulate doing so in the clouds above the runway as you descend.

With a circular approach to a circular runway, aren't all winds cross winds (and, also, headwinds and tailwinds)?

That was a trick suggestion. Also while dealing with flying the airplane, also be sure to keep track of whether the landing pattern is clockwise, or counter-clockwise, or was it clockwise, and which heading you are currently on and how that relates to your rejected landing turn in case you need it, which you probably will. I don't see what could possibly go wrong.

All I'll say is if I was the head of the programming committee for the air transport association conference, I would invite the designer to keynote, as comic relief.

I don't see why a human has to do this.

Even if they do, I don't see why it can't be done with an awareness system like the F-35 helmet. That would let you see the outline of the runway through arbitrary fog.

There is no special "rejected landing turn". You rise up just a bit, keep following the runway, and retry when you wish.

> I don't see why a human has to do this.

Because automated systems fail.

> Even if they do, I don't see why it can't be done with an awareness system like the F-35 helmet.

Because most aircraft aren't F-35s and expenses and fallible systems that are appropriate and sensible for combat aircraft that may have to operate in conditions in which civilian aviation would not aren't necessarily something you want to make civilian aviation dependent on in conditions in which, with sensible things like straight runways, it would not be.

> I don't see why a human has to do this.

Because a human driving the plane has a vested interest in surviving the landing.

A person who writes buggy landing software, or an operator "piloting" it from the ground can get another job if they fail.

The Romans would put civil engineers underneath newly-built bridges, then march the army over. Very few low-quality bridges were built, and nobody built more than one low-quality bridge.

This sort of thing works for planes too.

No. Wind can be circular. For example, right above the runway could be the eyewall of a hurricane.

Good point.

That might produce other problems for landing, though.

Correct me if I'm wrong, but I thought the whole idea behind the circular part was so that doing a crosswind landing would be "easier", because you could setup the landing so the crosswind pushes your towards the center...

If the wind cooperates and goes in a single direction during landing, sure, but why does nature have to be so kind?

There can be pretty severe shear at different altitudes on the way down, and the wind can flip direction any time it wants.

thanks, good explanation

and conventional runways are ok with all wind? because the angle stays constant?

You never land with a tailwind and if the wind shifts the airport simply change the runway to the opposite one. If you use a circular runway and landing with a cross wind you would during your landing roll be subjected to a tailwind.

Sure, you could stop using that part of the runway but then we're back at the practical economical argument for this design.

"Sure, you could stop using that part of the runway"

Would you have to stop using that part of the runway, or stop using it within a certain height?

That is to say, while landing, if at that part of the runway the aircraft was already solidly on the ground would it still be a problem? Likewise taking off; if the take off point was simply well before that part of the runway, would that be fine (have to think it would be, because after taking off the plane has no reason to follow the runway - it's already gone)?

If that's fine, then would you actually have to stop using that part of the runway? Could you just rotate the landing/launch points so that nobody was landing/launching at that part?

Good thought, savety margins and taxi way to the next exit add even more space.

Would there be taxiways to the next exit? I would have thought that a level interior ring of the runway would basically be one big circular taxiway; entrances and exits would basically be lights and markings, and planes would get directed to the appropriate entrance/exit depending on traffic, weather and the plane's own characteristics.

Reply on news.ycombinator.com