Vortex-junction: the next-gen roundabout

[Stephane Dumas]

I spotted this interchange idea on Jalopnik http://jalopnik.com/5497618/vortex-junction-the-next+gen-roundabout

[rawmustard]

I spotted this interchange idea on Jalopnik http://jalopnik.com/5497618/vortex-junction-the-next+gen-roundabout

It seems like a different person has posted about this on all the different fora I read.

[agentsteel53]

rats, they patented the dang thing!  

I came up with that when I was 6 and saw my first rotary/roundabout/traffic-circle/whatever they call it in your jurisdiction, in Massachusetts it was the Cape Cod Rotary and it was awful and I thought "why doesn't traffic enter from the *inside* and leave at the *outside* and therefore there'd be a lot less weaving".  

for all I know, my dad still has the video in which I ask the question ... though likely he may have muted the audio track, as it was peppered with him also noting the trouble the M. F. Cape Cod Rotary caused!

[Chris]

That idea is imaginative, but completely useless. Nobody would want to build such a big interchange.

[froggie]

Less weaving, possibly, but MUCH LARGER right-of-way requirements.  Might as well build a regular interchange.

[agentsteel53]

Less weaving, possibly, but MUCH LARGER right-of-way requirements.  Might as well build a regular interchange.

could a regular interchange of that size handle six in/out paths (i.e. three roads intersecting, as in that diagram), or just the usual four?  More generally, how does that configuration scale as we add more paths, compared to a regular interchange?

[roadfro]

This "vortex" junction doesn't seem to cut the mustard to me.

The circulation would appear to work well for the one direction of travel indicated. But if a motorist were to enter at 4 and exit at 3, they'd be looking at cutting across 5 lanes of traffic or doing more than a full loop to get there.

They also mention putting something like a sports stadium in the middle, yet show no concept of how that might be accomplished. And if the middle of this thing could accommodate a sports arena (with required parking), it's probably looking at a half mile to mile radius from the center to the circulating roadway. Right of way alone would probably make this prohibitively expensive to build.

I'd guess a stack interchange, even with six entry/exit points, could be built on a smaller footprint. Whether a stack could be designed and built to be less expensive than the 'less complex but much larger' vortex would be another issue.

[Truvelo]

I also think it's a stupid idea. The number of bridges required will also add to the cost. It's also flawed in that if there are several freeways meeting at this point it means using the roundabout thingy to go straight through unless even more bridges are added to accommodate through traffic.

[Duke87]

Why the loops? The exiting roadways could just as easily make a 90 degree right after a slight left. It'd be smoother and allow higher speeds without costing any extra.
Still, there is some very harsh weaving here. The concept, as presented, is flawed.

Although, I see plenty of potential to play with this and modify it. That flaw may be able to be worked around.

[J N Winkler]

I am not real keen on the vortex interchange for the operational reasons Roadfro cites.  However, if there are more than two intersecting roads and it is possible to build complementary infrastructure to prevent sterilization of land within the interchange area, it can be much more economic than a fully semidirectional design with direct connectors.

Here is the mathematics:  for n intersecting freeways, you have 2n senses (directions of movement) total.  Each sense has to have a direct connector with all the other senses except (1) itself and (2) the opposite sense.  So the total number of direct connectors that has to be provided is 2n(2n - 2).  For a simple four-level stack interchange n = 2 and the number of direct connectors is 8 (this counts the link ramps handling right-turning movements as direct connectors).  For an intersection of three freeways n = 3 and the number of direct connectors is 24!

[xonhulu]

could a regular interchange of that size handle six in/out paths (i.e. three roads intersecting, as in that diagram), or just the usual four?  More generally, how does that configuration scale as we add more paths, compared to a regular interchange?

Especially if the 3 through routes (6 paths) don't need to meet at a single point, but can form sides of a triangle of roughly the same size.  Use some c/d lanes and the weaving wouldn't be too disruptive of through traffic.  You'd probably still have room for the stadium...

In a regular interchange, through traffic doesn't need to change lanes, but in the spiral model, wouldn't you have to move over two lanes to exit the opposite side?  Not sure that's any more efficient than a regular interchange.  But then, I don't know any examples of this kind of 6-path interchange, so there's nothing to compare it to.

[J N Winkler]

Why the loops? The exiting roadways could just as easily make a 90 degree right after a slight left. It'd be smoother and allow higher speeds without costing any extra.

That approach results in flyovers which cross the circulating roadway at an oblique angle and thus have greater skew and deck square footage, unless you go under and use culverted undercrossings.  (We don't use them much but they are very common in Spain, as are pergolas.)

[Alps]

Less weaving, possibly, but MUCH LARGER right-of-way requirements.  Might as well build a regular interchange.

could a regular interchange of that size handle six in/out paths (i.e. three roads intersecting, as in that diagram), or just the usual four?  More generally, how does that configuration scale as we add more paths, compared to a regular interchange?

I've modeled "regular" interchanges up to 8 paths.  With six paths there are three ways to go - a double stack modified to stay within 4 levels, a modified cloverleaf with enough loops and fingers to serve everyone, or a hybrid stack/double clo that's my favorite.  With eight paths, you pretty much end up with the double stack modified plus a hybrid cloverleaf laid on top of it.  The six path interchange would pretty much fit in a right of way this large - you lose the room in the middle but gain more maneuverability.  For less trafficked roads, you could just run a couple of the through roads on a second level and take away the through traffic from the conventional roundabout below.

[Alps]

For an intersection of three freeways n = 3 and the number of direct connectors is 24!

And this is why for something that large you start having combined ramps (think volleyball interchange with ramp weaves).  Look at Garden State Parkway / US 9 / I-287 / NJ 440 / CR 514 for an example of a multi-path interchange that uses combined ramps (even combined with freeway mainlines) to accomplish the necessary moves.

[Duke87]

After playing with it, this is what I've come up with (pardon the quick crude drawing):

Only four spokes shown here for the sake of simplicity, but the same concept applies with more (would demand a larger radius, though).

Movement to the next consecutive spoke involves separate ramps that don't even enter the rotary proper. Entering the rotary itself and then following your lane will take you directly off three spokes ahead. Going only two spokes involves shifting left. Going four spokes (a U-turn in this case), involves shifting right. Going five would involve three lane shifts right, but not in the hugest of hurries.

Thoughts?

[Scott5114]

Left exits make the MUTCD cry...

[Michael]

Maybe you could add a right exit opposite the loops?  It's too complex IMO.

[codyg1985]

From the looks of the modified interchange Duke87 created, going through the interchange without existing would require traffic to weave over to the left in order to stay on the same road. That thing would be a weaving nightmare.

The original interchange just seems like a waste of space to me.

[J N Winkler]

From the looks of the modified interchange Duke87 created, going through the interchange without existing would require traffic to weave over to the left in order to stay on the same road. That thing would be a weaving nightmare.

Yes, that is a problem.  It could be solved by running the intersecting freeways through the interchange core, and I am pretty sure this could be done while keeping the number of levels crossing at any point down to two.

The original interchange just seems like a waste of space to me.

It does have the potential to sterilize land, but I think this could be mitigated by building accesses to allow parcels within the interchange to be developed.  The more serious problem, as I see it, is that it requires a TOTSO for all traffic passing through the interchange, even if it is continuing in the same direction.  This could be eliminated as I suggested above (by running the freeways through the interchange core) but the tradeoff is that a crossing structure has to be provided in the middle with one level for every intersecting freeway.

Because of the TOTSO requirement, the interchange design as proposed really works well only when the majority of the traffic is stopping in the middle of the interchange (e.g., at the sports stadium).  If bridges have to be provided to avoid inconveniencing through traffic, the design problem becomes different and there are probably other solutions which are more efficient in terms of construction cost, use of land, etc.

[roadfro]

I'm fairly convinced that any roundabout-esque interchange design with left exits/entrances just isn't going to work, without some mitigating factor that will start to make the design either too complex or too expensive.

Left exits make the MUTCD cry...

Actually, left exits make the AASHTO Green Book cry...

The MUTCD doesn't include much text about the physical design of interchanges, but only how to sign and stripe them. The "Green Book" (formally, A Policy on Geometric Design of Highways and Streets) is the design standard many agencies use, or at least what they base their own design standards off of.

[Scott5114]

I was referring to all the new stuff with regards to left exits that was added in the 2009 MUTCD. Left exits now have to have an exit tab that is twice as big and with the word "LEFT" in black on yellow above "EXIT". For non-numbered exits, a yellow "LEFT" tab is required.

[realjd]

While I agree with the consensus that this is a bad idea for a freeway interchange, I think the original intent was to replace a regular roundabout at a location where low-speed surface streets intersect.

Of course, his basis for this was that the inside lanes for a roundabout are underutilized. If people use a roundabout properly, all lanes are used, even the inside ones, since the inside lane has the right of way to exit the roundabout at any given street.

[CL]

Speaking of complex interchanges, does anyone see the usefulness of this monstrosity (click on exit 265 - Provo Center Street and then "design on" and zoom in)? Why in the world would they build that?? A simple diamond or SPUI interchange would suffice here, would it not? (As a side note that same project is building two diverging diamonds)

[Truvelo]

I think the Provo Center Street proposal is a wonderful design. It means all movements are freeflow with no traffic lights or stop signs. The only downside is southbound I-15 to westbound Center Street means doing an almost complete lap of the roundabout.

[roadfro]

While I agree with the consensus that this is a bad idea for a freeway interchange, I think the original intent was to replace a regular roundabout at a location where low-speed surface streets intersect.

If that's the case, it's even worse. If a regular surface streets intersect at something like this, it makes it even more prohibitively expensive to implement. I could *maybe* see it if some kind of large regional park would be put in the middle. But the whole interchange would still take up a large chunk of real estate in any kind of populated area.

Of course, his basis for this was that the inside lanes for a roundabout are underutilized. If people use a roundabout properly, all lanes are used, even the inside ones, since the inside lane has the right of way to exit the roundabout at any given street.

The inner lane of a roundabout does not always have the right of way to exit at any exit of a roundabout. Generally the only time a driver in the inner lane should have a guaranteed right of way to exit is at the last exit before reaching the leg they started at.

It all depends on the design, though. A properly designed roundabout will have effective utilization of the inner lanes(s) based on the signs and markings that dictate the lane assignments. Of course, this assumes the roundabout has adequate advance signage, proper lane markings, and carries drivers intelligent enough to interpret these...which is not always the case.

The 2009 MUTCD shows typical lane markings at a variety of roundabout applications, far more guidance than the two or three figures of the previous manual. While not a design guide, the figures do give one a good idea of the way modern roundabouts should be designed.