Diverging Diamond Interchanges

[vtk]

The DDI at I-270 and Roberts Rd in Columbus works very well. It moves traffic much better than the conventional diamond interchange it replaced. I do not believe a SPUI would offer any significant operational advantages over a DDI here. I have not heard of any wrong-way drivers at this location. Local drivers call it weird, but they seem to have accepted it more quickly than when roundabouts made their debut here.

The primary hypothetical operational weakness of a DDI is a suboptimal experience for through traffic on the arterial roadway. When the volume of that traffic flow is insignificant compared to the traffic entering and exiting the freeway, as is the case at Roberts Rd and I-270, this concern is of minimal importance.

[lordsutch]

A new DDI at University & I-75 in Michigan is signed for 25 mph.  Directly upstream and downstream of the DDI the arterial is signed for 45 mph.  Subdivision streets are signed for 25 mph.  Good thing drivers travel nearly 2500 feet through this suburban DDI at 25 mph.  Maybe the designers just want drivers to slow down so they can admire their fabulous DDI.  We are living in bubble-boy world where high speed suburban arterials are signed for 25 mph. 

It takes 1.14 minutes to drive 2500 feet at 25 mph. It takes 0.63 minutes to drive 2500 feet at 45 mph. So, basically you're adding 30 seconds to traverse the intersection completely, assuming neither light is red and traffic can maintain the posted speed limit.

On the plus side, any T-bone crashes at the signalized conflict points will be much less likely to lead to deaths and injuries at 25 mph than 45 mph.

[vtk]

A new DDI at University & I-75 in Michigan is signed for 25 mph.  Directly upstream and downstream of the DDI the arterial is signed for 45 mph.  Subdivision streets are signed for 25 mph.  Good thing drivers travel nearly 2500 feet through this suburban DDI at 25 mph.  Maybe the designers just want drivers to slow down so they can admire their fabulous DDI.  We are living in bubble-boy world where high speed suburban arterials are signed for 25 mph. 

It takes 1.14 minutes to drive 2500 feet at 25 mph. It takes 0.63 minutes to drive 2500 feet at 45 mph. So, basically you're adding 30 seconds to traverse the intersection completely, assuming neither light is red and traffic can maintain the posted speed limit.

On the plus side, any T-bone crashes at the signalized conflict points will be much less likely to lead to deaths and injuries at 25 mph than 45 mph.

And what's the traffic volume of drivers traveling that full 2500 foot distance on University?  What are the traffic volumes of drivers who make a right turn to enter I-75, drivers who make a left turn to enter I-75, drivers exiting I-75 who turn right, and drivers exiting I-75 who turn left?  I suspect (because this is the most sensible reason to select a DDI over other designs) that the through traffic on University is the smallest-volume traffic movement at the interchange.  Fussing about adding half a minute of travel time for a small fraction of the users of an interchange is missing the point.

[tradephoric]

I suspect (because this is the most sensible reason to select a DDI over other designs) that the through traffic on University is the smallest-volume traffic movement at the interchange.  Fussing about adding half a minute of travel time for a small fraction of the users of an interchange is missing the point.

I couldn't find an environmental assessment for the I-75/University DDI but ironically there is another DDI being constructed at I-75/University about 1000 miles to the south.  Take a look at this model FDOT put together.  Can you honestly tell me that the through traffic on University Parkway isn't a significant movement?  It looks to be just as heavy if not heavier than the traffic entering I-75. 

https://www.youtube.com/watch?v=24jMVZszPTY

UDOT put together an evaluation of the first DDI constructed in Utah at American Fork and I-15.  Detailed turning movement counts are included in that study.  If you do some simple arithmetic you will find that 37% of drivers traveling EB American Fork continue straight through the DDI and 51% of drivers traveling WB American Fork continue straight through during peak hour.  Contrary to what you suspect VTK, a lot of drivers continue on the arterial at a DDI and don't enter the freeway.


https://www.udot.utah.gov/main/uconowner.gf?n=10172614219775523

It's difficult to provide good signal progression at closely spaced traffic signals.  Providing good signal progression in one direction often leads to green-to-red signal progression in the other.  Here is a video of the "arrival performance"  of traffic at the Bangerter DDI in Salt Lake City, Utah.  Throughout the video you see drivers on Bangerter Hwy experience green-to-red coordination while traveling straight through the DDI:

https://www.youtube.com/watch?v=RhazZu4xMeA

Another thing to consider is that DDI's often require short cycle lengths (to prevent backups on the bridge deck).  This means that DDI often aren't coordinated with the surrounding signals which lead to another potential green-to-red scenario for drivers.

UDOT DIVERGING DIAMOND INTERCHANGE (DDI) OBSERVATIONS AND EXPERIENCE:
Coordination of the DDI with adjacent signals is not easily done. Most DDIs need a lower cycle length than the adjacent signals. This may result in a vehicle having to stop at both the off ramp terminal and the next adjacent signal.

[kphoger]

It takes 1.14 minutes to drive 2500 feet at 25 mph. It takes 0.63 minutes to drive 2500 feet at 45 mph. So, basically you're adding 30 seconds to traverse the intersection completely, assuming neither light is red and traffic can maintain the posted speed limit.

What is the likelihood of the highlighted statement being reality at a DDI?

[lordsutch]

It takes 1.14 minutes to drive 2500 feet at 25 mph. It takes 0.63 minutes to drive 2500 feet at 45 mph. So, basically you're adding 30 seconds to traverse the intersection completely, assuming neither light is red and traffic can maintain the posted speed limit.

What is the likelihood of the highlighted statement being reality at a DDI?

Pretty low to begin with, of course.

[vtk]

Tradephoric, my my defense of the DDI applies only to those situations where the through arterial volume is insignificant, in which case your whole signal progression rant is insignificant. Don't rant at me for DOTs building them where they are not appropriate.

[jeffandnicole]

I missed signal progression talk.

[tradephoric]

The primary hypothetical operational weakness of a DDI is a suboptimal experience for through traffic on the arterial roadway. When the volume of that traffic flow is insignificant compared to the traffic entering and exiting the freeway, as is the case at Roberts Rd and I-270, this concern is of minimal importance.

Here are 2030 turning movement counts for the Roberts Rd and I-270 interchange.  During the AM rush, 55% of EB Roberts Rd drivers travel straight through at least one DDI signal and 24% travel straight through both DDI signals.  Going WB in the AM, 53% travel straight through one DDI signal and 45% travel through both DDI signals.  If you have any questions how I came up with these percentages I'll be more than happy to step you through it (I sincerely mean that). 


http://www.morpc.org/trans/nw33270ex%20fut%20conditionssecondhalf.pdf

The percentage of through traffic at the Roberts Rd and I-270 DDI is not "insignificant" .  I agree that a primary operational weakness of a DDI is a "˜suboptimal experience for through traffic on the arterial roadway' but I disagree with VTK that it's "˜hypothetical'.

[vtk]

http://www.morpc.org/trans/nw33270ex%20fut%20conditionssecondhalf.pdf

To better understand the data tradephoric dug up, I visualized it by drawing each movement as a lane whose width is proportional to the traffic volume.

2003 AM/PM peak hour volumes:


Forecast 2030 AM/PM peak hour volumes:


I noted that the Roberts Rd through movements in 2003 are pretty darn slender compared to some of the turning movements.  Then I scratched my head for a bit.  Why was the through Roberts Rd traffic forecast to increase so much?  Roberts Rd doesn't really go anywhere to the east.  Did the consultant assume the long-abandoned Roberts Rd / Lane Ave bridge might be built?  And why does the PM NB to WB movement actually decrease from 2003 to 2030?  Then I remembered traffic volume projections don't have a great track record of turning out accurate.  Then I remembered this study was done in like 2004, before the Great Recession.  Remember how a few years ago a lot of folks were proclaiming we'd reached "peak VMT", as though humans will never again do as much road driving as we recently did?  Then, as the economic recovery continued, people started driving more again.  Traffic volumes today are probably still a bit lower than what would have been predicted for this year a decade ago, and traffic volumes in 2030 will probably maybe wind up a bit lower than what this study suggested.  So, basically, I'm going to say those 2030 projections should be interpreted with a big giant asterisk.

After thinking a bit more about it, I decided those figures still don't quite illustrate my point as clearly as I'd like.  Then I realized, all I really need to do is show the relative volumes of Roberts Rd through traffic and the entering & exiting traffic, with those broad categories in aggregate.  So here are the pie charts:



The Roberts Rd through traffic accounted for roughly 10% of all the traffic using that interchange in 2003.  The forecast for 2030 says that slice of the pie grows to 20%, but again, big giant asterisk.  I'll ballpark the 2016 figure at 10—15%.

I'm really not worried about a slight inconvenience to less than 20% of the users of an interchange if it means a better experience for the majority.  And it really is only a slight inconvenience, as the two signals at a DDI operate on simple 2-phase cycles, so the green time at each is nearly half.  Also, my perspective on this whole thing is by comparing the current DDI to the conventional diamond interchange it replaced, which had significant delays (especially from EB Roberts to NB I-270 in the afternoon).  Other interchange designs may operate comparably well, but would have been significantly more costly to build, and it would be difficult to convince me that cost is justified given how adequately the DDI here operates.

Perhaps it's worth having a discussion about signal progression: its advantages, its requirements, factors that reduce its effectiveness, and situations where it's really not even relevant.  But that is a highly technical discussion that belongs in its own thread, because its fine details tend to overwhelm threads that are supposed to be about other topics.

[tradephoric]

After thinking a bit more about it, I decided those figures still don't quite illustrate my point as clearly as I'd like. 

I believe your point was the DDI at I-270 and Roberts Rd works fine because the amount of through traffic on Roberts Rd is "˜insignificant'.  After reading the study you have backtracked on that a little and now describe the Roberts Rd through traffic as "less than 20%" .   But if you consider the Roberts Rd traffic that travels through at least one of the DDI signals the percentages exceed 50% (whether you look at 2003 or 2030 volumes).
 
So far we have only been considering the potential delays on Roberts Road but consider the delay drivers exiting I-270 experience.  Every driver exiting the freeway encounters a "NO TURN ON RED"  sign regardless of what off-ramp they take.  Drivers have to wait for a green light to proceed and are not allowed to make a simple right turn on red in gaps.  There are plenty of interchange designs that allow drivers to turn right on red exiting the freeway but this DDI isn't one of them.

Also, my perspective on this whole thing is by comparing the current DDI to the conventional diamond interchange it replaced, which had significant delays (especially from EB Roberts to NB I-270 in the afternoon).  Other interchange designs may operate comparably well, but would have been significantly more costly to build, and it would be difficult to convince me that cost is justified given how adequately the DDI here operates.

I could agree with that bold statement if you replaced "comparably well"  with "better" .   Maybe you are so biased towards DDIs that you won't even admit that other interchange designs would perform better.  There's no doubt that the DDI leads to unnecessary driver delays when compared to other interchange designs. Regardless if another interchange would have been practical at I-270 and Roberts Road, you should at least be able to admit that a DDI isn't the most efficient interchange design out there.  Anybody with a shred of common sense would be able to admit that.

[lordsutch]

There are certainly examples of DDIs that permit right turns on red and/or use a yield instead. I'm not sure off-hand why the 270/Roberts DDI isn't one of them (some DOTs do have a general policy not to permit right on red for multi-lane right turns).

Parclos have their own issues, usually involving short acceleration or deceleration distances and merge lanes and unfriendliness to pedestrians and cyclists, in addition to greater land use, so they're certainly not better than a DDI under all circumstances.

[jeffandnicole]

...Regardless if another interchange would have been practical at I-270 and Roberts Road, you should at least be able to admit that a DDI isn't the most efficient interchange design out there.  Anybody with a shred of common sense would be able to admit that.

The most efficient interchange designs are going to be full, non-stop movement designs that don't have any crossing traffic.  That's limited to cloverleafs and flyovers.  And both of those can have limitations, especially in terms to the room needed, weaving and cost factors (especially overpasses).   So then you take a look at the area, including the surrounding area, and figure out which designs will work best for the area based on current and future traffic flows.

[vtk]

I could agree with that bold statement if you replaced "comparably well"  with "better" .   Maybe you are so biased towards DDIs that you won't even admit that other interchange designs would perform better. 

No, I meant "comparably well" because other designs have other drawbacks. A SPUI has a three-phase signal instead of the two-phase signals of a DDI, so wait times for traffic turning left to enter the freeway may be longer. A parclo has loops that take more time to drive around than direct ramps. Anything that doesn't fit in the original ROW would probably still not be built yet because of the additional paperwork. In the absence of a perfect solution, there's very little room for improvement over "adequate". Or do you have a perfect solution you would like to share?

[jakeroot]

A parclo has loops that take more time to drive around than direct ramps.

I think a better argument against Parclos is the sheer amount of redundant land within the loops. DDIs, SPUIs, and regular diamonds are much better fits for urban areas. Even out in the sticks, you still need to buy extra ROW for a Parclo A4 or B4 (though as long as that's not an issue, I will agree with Tradephoric that Parclo B4s are the best interchange you can build, short of full grade-separation).

[tradephoric]

I could agree with that bold statement if you replaced "comparably well"  with "better" .   Maybe you are so biased towards DDIs that you won't even admit that other interchange designs would perform better. 

No, I meant "comparably well" because other designs have other drawbacks. A SPUI has a three-phase signal instead of the two-phase signals of a DDI, so wait times for traffic turning left to enter the freeway may be longer. A parclo has loops that take more time to drive around than direct ramps. Anything that doesn't fit in the original ROW would probably still not be built yet because of the additional paperwork. In the absence of a perfect solution, there's very little room for improvement over "adequate". Or do you have a perfect solution you would like to share?

To determine which interchange has the best operational performance you got to compare common measures of effectiveness.  What design has the lowest network delay, lowest average driver stops, and shortest queue distances?  You can't answer those questions by simply knowing that it takes more time for drivers to travel around a loop ramp.  Your argument lacks any meaningful facts.

[vtk]

What design has the lowest network delay, lowest average driver stops, and shortest queue distances?  You can't answer those questions by simply knowing that it takes more time for drivers to travel around a loop ramp.  Your argument lacks any meaningful facts.

Taking extra time to traverse a loop is network delay, just like waiting at a red light. Besides, your arguments – a lack of "through-roadway experience" for the very small fraction of traffic that might expect it, and a signal that's difficult to coordinate for two-way progression in a corridor that doesn't have two-way signal progression anyway – are not meaningful deficiencies of an interchange at this location.

Do you want to try to find the study in which ODOT evaluated different interchange designs at I-270 & Roberts Rd?  I'm sure I saw at least a comparison with different evaluation measures, including something like "operational efficiency", for each design.  I'm not saying ODOT concluded the DDI would have the best operational efficiency – though they may have – but they certainly modeled and simulated and analyzed, and didn't just build the thing and hope it would work.  I remember seeing concept drawings for a diamond with roundabouts, and maybe a regular diamond with a bunch more lanes, among the alternatives considered.  I don't remember whether any parclo designs were evaluated, and I'm pretty sure there was no SPUI alternative at that stage; such designs were likely eliminated early in the process due to cost, and if kept and chosen, likely wouldn't be funded yet.

My whole point is not that DDI interchanges are awesome and never a poor choice.  My point is that they can be a sensible choice in some locations, including the one in Columbus.  My uninformed assumption that any other DDI was not a poor choice for that location is based not on bias towards this interchange type, but on a general (perhaps uninformed) assumption of competence within the agencies that build them, paired with the fact that they chose to build a DDI. (I try not to express this as a certainty when talking about interchanges I'm not personally familiar with, hence the "I suspect" phrasing used in a previous post.)

Just as you perceive that I have a bias in favor of DDIs, you are perceived to have a bias in favor of considering signal progression as the single most important factor in roadway & interchange design, with reduced attention to other factors typically considered. Other people, myself included, recognize the benefit of good signal progression, but do not give it enough weight to sink the DDI concept completely. I believe that discussion may be worth having, but not in a thread that's ostensibly about diverging diamond interchanges. Or it may devolve into a religious war, which definitely should be kept separate from a thread that's about anything else.

[6a]

After thinking a bit more about it, I decided those figures still don't quite illustrate my point as clearly as I'd like. 

I believe your point was the DDI at I-270 and Roberts Rd works fine because the amount of through traffic on Roberts Rd is "˜insignificant'.  After reading the study you have backtracked on that a little and now describe the Roberts Rd through traffic as "less than 20%" .   But if you consider the Roberts Rd traffic that travels through at least one of the DDI signals the percentages exceed 50% (whether you look at 2003 or 2030 volumes).
 
So far we have only been considering the potential delays on Roberts Road but consider the delay drivers exiting I-270 experience.  Every driver exiting the freeway encounters a "NO TURN ON RED"  sign regardless of what off-ramp they take.  Drivers have to wait for a green light to proceed and are not allowed to make a simple right turn on red in gaps.  There are plenty of interchange designs that allow drivers to turn right on red exiting the freeway but this DDI isn't one of them.

Please help me out here, I'm genuinely trying to understand you completely. It has always been my understanding that the selling point of a DDI was to assist with left-turning traffic. That said, at this particular location, the existing bridge deck caused real problems in this regard, as a left turn lane either wouldn't fit, or reduced through traffic to one lane. In addition, although I don't see it mentioned anywhere, a significant part of said traffic is trucks. Now that the DDI is operational, it makes sense to me that a majority of traffic would only encounter one signal. It would seem to me that was the intention of having this particular design at this location...help with the large left turning traffic by having it bypass the second signal.

Before it was built, signal progression was hopeless anyway, since traffic routinely queued past the next set of lights anyway. What am I missing? I feel like there's one piece I'm not catching somewhere.

[tradephoric]

Here is one of my favorite theoretical interchanges known as the folded interchange:

This interchange would have been completely impractical to build at I-270 and Roberts Road but it would have had better measures of effectiveness than the DDI.  The problem I have is when someone suggests that other interchange designs - regardless of how much money they cost -  would perform "comparably well"  to the DDI.  If money was no object they would stop building DDIs entirely (since they are not very efficient).  It seems like VTK is trying to backtrack from their "comparably well" statement but just can't do it.

[jeffandnicole]

Here is one of my favorite theoretical interchanges known as the folded interchange:

This interchange would have been completely impractical to build at I-270 and Roberts Road but it would have had better measures of effectiveness than the DDI.  The problem I have is when someone suggests that other interchange designs - regardless of how much money they cost -  would perform "comparably well"  to the DDI.  If money was no object they would stop building DDIs entirely (since they are not very efficient).  It seems like VTK is trying to backtrack from their "comparably well" statement but just can't do it.

I see some uncomfortably close calls with RTOR in that video.

[tradephoric]

I see some uncomfortably close calls with RTOR in that video.

Also, drivers exiting the loop ramp take it at 70 mph lol.  Obviously this is a conceptual model and it took me about 20 minutes to put together.  Sometimes drivers do take loop ramps at 70 mph though:

https://www.youtube.com/watch?v=2dvm7-5yMMo

^This scenario is probably one of the strongest arguments against loop ramps, specifically exiting loop ramps.  Of course, based on some of the callous comments made on the crash prone roundabout thread regarding fixed objects in the middle of roundabouts, I doubt many people on this forum have a problem with drivers taking loop ramps at 70 mph (since it's their own fault.. personal responsibility... all that good stuff).

[vdeane]

If money was no object they would stop building DDIs entirely (since they are not very efficient).

I doubt people who own land next to interchanges would be very happy with that.

[jeffandnicole]

I see some uncomfortably close calls with RTOR in that video.

Also, drivers exiting the loop ramp take it at 70 mph lol.  Obviously this is a conceptual model and it took me about 20 minutes to put together.  Sometimes drivers do take loop ramps at 70 mph though:

https://www.youtube.com/watch?v=2dvm7-5yMMo

^This scenario is probably one of the strongest arguments against loop ramps, specifically exiting loop ramps.  Of course, based on some of the callous comments made on the crash prone roundabout thread regarding fixed objects in the middle of roundabouts, I doubt many people on this forum have a problem with drivers taking loop ramps at 70 mph (since it’s their own fault.. personal responsibility… all that good stuff).

Many years ago, a woman was killed near my house when she took a diamond interchange ramp at 90 mph and hit a house across the street.

This scenario is probably the strongest argument against straight ramps, especially exiting straight ramps.

[tradephoric]

Many years ago, a woman was killed near my house when she took a diamond interchange ramp at 90 mph and hit a house across the street.

This scenario is probably the strongest argument against straight ramps, especially exiting straight ramps.

Well, good point.  Fatal crashes can occur at straight diamond ramps too.  I'll always remember that Atlanta bus crash on I-75 where the bus driver mistook an exit ramp for a freeway-lane.  The bus blew through the bridge deck and landed on it's side on the freeway killing 6. 

http://usatoday30.usatoday.com/news/nation/2007-03-02-atlanta-bus-wreck_x.htm

Ok J&N, I've always wondered why agencies seem to favor entering loop ramps onto the freeway as opposed to exiting loop ramps.  Agencies must have a reason they favor entering loop ramps.  What is it?

[tradephoric]

If money was no object they would stop building DDIs entirely (since they are not very efficient).

I doubt people who own land next to interchanges would be very happy with that.

VTK argued that other interchange types would perform "comparably well" to the DDI in relation to operational performance.  We aren't looking at cost or the environmental impacts of the interchange.  The focus was on total network delay.  You are muddying the waters by bringing up people crying about owning land next to a freeway??   Thanks for that irrelevant comment.