Comparative Analysis of Intersection Designs for Seven Corners Redesign
Advisor: Thomas Casalaspi
Internship Location: Wells + Associates
Onsite Mentor: Larry Sefcik
Seven Corners in Falls Church, Virginia, is notorious throughout Northern Virginia as one of the region’s worst intersections. Drivers complain about missing turns, dangerous merges, and confusing lane configurations. Many actively avoid the area entirely, taking longer routes just to escape the chaos. With 110,000 to 130,000 vehicles daily, a Level of Service grade of F (fail), and average delays exceeding 3 minutes during rush hour, Seven Corners exemplifies how poor intersection design affects entire communities. Approximately 25% of traffic fatalities and 50% of traffic injuries nationwide occur at intersections, making poor design not just frustrating but dangerous. Fairfax County has proposed a $130-140 million Ring Road reconstruction based on studies from 2012-2015. However, no recent independent comparison to alternative intersection types has been conducted.This project fills that gap by designing and testing alternative intersection configurations using professional traffic simulation software. Working with Wells + Associates and HNTB, local traffic engineering firms, I will design 2-3 alternative intersections and compare their performance to both the Ring Road and the current layout. The final product includes a detailed road model, traffic simulation results, and a technical report with data-driven recommendations.
Project Blog
Post 1: Introduction
If you've ever driven through Seven Corners in Falls Church, you know the frustration. I've missed my turn there countless times, watched drivers cut across multiple lanes in panic, and even rerouted my GPS to avoid the intersection entirely, taking a longer route just to bypass the chaos.
But what started as a personal annoyance turned into genuine curiosity. Why is Seven Corners so bad? And more importantly, could it be better? That question led me to discover that Fairfax County has proposed a $130-140 million Ring Road reconstruction. The price tag alone caught my attention, but what really stood out was that no recent independent comparison to alternative intersection types has been done since the design was selected over a decade ago.
Over the next three months, I'll be working with two local traffic engineering firms (Wells + Associates and HNTB) to learn Synchro, a professional traffic simulation software, and design alternative intersection configurations for Seven Corners. I'll compare these alternatives to the county's Ring Road proposal using traffic data and engineering metrics like Level of Service, delay times, and conflict points. Unlike professional engineers constrained by budgets and political pressures, I can objectively evaluate all options based purely on what works best.
Approximately 25% of traffic fatalities and 50% of traffic injuries in the U.S. occur at intersections. Seven Corners handles over 110,000 vehicles daily with a failing grade (Level of Service F) and delays exceeding 3 minutes during rush hour. Poor intersection design isn't just annoying. It's dangerous.
This project is my chance to tackle a real problem affecting my community every single day. I'm excited to learn whether my independent analysis will confirm the county's choice or suggest better alternatives. Either outcome would be valuable. Validation gives the community confidence in the investment, while discovering better options could save money or improve safety.
Week 1: First Steps
In the first week, though I didn't do much on Synchro, I've been figuring out where to start. Before beginning my work on Seven Corners, I decided to work on an upcoming traffic signal right outside of BASIS McLean first.
It's a good learning project because it's less complex than Seven Corners, but it still lets me test out different signal configurations and see how small design choices affect traffic flow. For example, should left turns get their own dedicated arrow, or should drivers yield and find gaps in traffic? Though these decisions seem minor, they have big impacts on the grade and flow of the intersection.
I'm also reading through a signal warrant analysis study from an engineering firm that previously worked on the BASIS traffic signal. Signal warrants are the criteria engineers use to determine whether an intersection needs a traffic signal, considering factors such as traffic volume, crash history, and pedestrian activity. Understanding this methodology is helping me see how professionals approach these decisions.
Testing concepts on the BASIS signal is helping me learn the fundamentals before tackling Seven Corners and all its complexity. Next week, I'll start building my first Synchro model. For now, I'm focused on learning the principles that will guide my design work later on.
Week 3: Review of the Signal Justification Report
This week, I got the full 77-page signal justification report that Wells + Associates wrote for the BASIS intersection. It's the same study I mentioned last week, but now I've actually read through it instead of just skimming. The report breaks down all nine MUTCD signal warrants, which are basically requirements that determine whether an intersection legally needs a signal. Only two warrants were met: four-hour vehicular volume and peak hour delay. Although the other seven warrants weren't met, the traffic volumes and delays were bad enough to justify the signal anyway.
What stood out to me was how detailed the data collection was. They counted every vehicle for 13 hours straight and even did a right-turn-on-red adjustment where they reduced certain volumes by 25% based on driver behavior at a nearby signal. The intersection currently operates at Level of Service F during the evening rush, which is the worst rating possible. With a signal, they're projecting it improves to LOS B. They also analyzed a roundabout alternative that would perform even better at LOS A, though cost or space constraints likely ruled it out.
I also have the actual Synchro files that Wells + Associates used for this study, which means I can open their model and see exactly how they set everything up. My plan for this week is to explore the existing model, figure out how the software works, and start testing some basic changes to see what happens to the results.
Week 4: Experimenting on Signal Timing
This week, I got into the actual Synchro files that Wells + Associates used for the BASIS intersection study. The model was already built, so I spent time figuring out how it was set up and trying some basic experiments with the signal timing. I changed the cycle length from the original to 120 seconds to see what would happen to the level of service. The results were interesting but also confusing because some of the delay calculations started throwing errors, which means I probably messed up the phase timing somewhere.
I also experimented with changing some of the left turn movements from permissive to protected. The eastbound left turn into the school is in a shared lane with through traffic, which means if someone's waiting to turn left, they're blocking everyone behind them. I thought making it protected would help, but when I ran the analysis, the LOS actually got worse in some places. I'm still trying to figure out exactly why that happened and whether it's a timing issue or if I assigned the phases wrong.
Right now, I'm out of town on break, so I'm not actively working on the model. When I get back, I need to go through the original configuration more carefully and understand what Wells + Associates actually recommended before I start changing things. I think I was making changes without fully understanding how the ring and barrier system works, which is probably why I kept getting errors. Next week, I'll reset and approach it more systematically.
Week 5: Shifting Focus to Seven Corners
Over spring break and this past week, I've been transitioning from the BASIS intersection mini-project to my actual senior project on Seven Corners. The BASIS work was useful for learning the basics of Synchro and understanding how professional traffic studies are structured, but Seven Corners is a much bigger challenge.
I'm currently in the process of getting the VDOT Synchro files for Seven Corners. The problem is, I'm not sure exactly who to contact or what the process is for getting these files. I've been reaching out to different people at VDOT and Fairfax County to see if they can share the models, but it's taking longer than I expected. Once I get the files, I can start testing different configurations and comparing them to the county's Ring Road proposal, but for now, it's mostly waiting for a response.
Week 6: Reading User's Perspective
I didn't make much progress on getting the Synchro files this week, but I'm planning to call Fairfax County DOT tomorrow morning when they open to request them directly. The person I originally emailed hasn't responded, so I am going in with a phone call.
While waiting, I've been reading more about the specific problems at Seven Corners to understand why it's so bad. One of the worst parts is the intersection of Arlington Boulevard Service Road and East Broad Street, where the turning-lane configurations are confusing, and people often go down the wrong road. There's also a random merge further down Arlington Boulevard right after the intersection that forces traffic to compress from multiple lanes down, which not only is puzzling, but it also creates a bottleneck and backs up into the intersection itself.
Trying to fix an intersection when all you have is "this sucks" is way harder than having people tell you "X road sucks because of Y." General complaints don't tell me what to actually fix, but specific problems like "Arlington Blvd has two back-to-back right turns onto East Broad St, and it's unclear which lane you need to be in for the second one" give me a good foundation.
Week 7: Running into a Problem
This week, I finally heard back from Fairfax County DOT after submitting a more specific FOIA request. The good news is they have the Seven Corners files and are willing to share them. The bad news is they're all in VISSIM format, not Synchro. I've been learning Synchro through my work on the BASIS intersection, so switching to completely different software at this point is a major setback. Not to mention, during my meetings with my advisors, they both agreed that VISSIM would be much more complicated and technical to learn in such a short time.
I'm planning to reach out to Dr. Du at HNTB to see if he has any suggestions or can help me figure out the best path forward. However, I don't think learning an entirely new software is realistic, especially because I'd have to reach out and try to obtain a demo license again. I'm wondering if I should go back to focusing on the BASIS intersection analysis I've already started, or find some middle ground. I have a lot of other work coming up and limited time to figure this out, so I need to make a decision soon about what's actually feasible to complete by May.