Comparing Field and Modeled Rural Bridge Responses to Moving Loads

Erica O West Author

07/28/2021 Added

18 Plays

Description

This poster compares strain data from SAP 2000 models with field data on a small rural bridge. This work will help improve research and will hopefully lead to better overall rural bridge health monitoring.

Searchable Transcript

Search:

[00:00:00.000]Hello, as you might be able to see,
[00:00:02.280]my name is Erica West and today I will be
[00:00:05.230]telling you about my project,
[00:00:06.990]Comparing Field and Modeled Rural Bridge
[00:00:09.130]Responses to Moving Loads.
[00:00:10.730]At the moment, it is incredibly hard
[00:00:13.380]to check rural bridge health for a number
[00:00:15.887]of reasons that we will
[00:00:17.250]get into momentarily.
[00:00:19.020]To improve research and better systems,
[00:00:21.360]this project is comparing field data
[00:00:23.202]collected with BDI system strain sensors
[00:00:25.929]to model created in the SAP 2000 software.
[00:00:29.795]There are over 11,000 rural bridges in
[00:00:32.175]Nebraska over 20 feet long
[00:00:34.082]on the county system and almost
[00:00:36.142]2 out of 5 are at least 50 years old.
[00:00:38.602]The counties that manage these structures
[00:00:40.685]need to balance their aging infrastructure
[00:00:42.892]with tight budgets and limited state help.
[00:00:45.308]As a result, counties need an inexpensive
[00:00:47.402]way to determine which bridges
[00:00:49.270]need repairs the most.
[00:00:50.802]To make maters even more complicated,
[00:00:52.849]any monitoring system would need
[00:00:54.529]to stand up to dust from the dirt road
[00:00:56.399]and absolutely no access to electricity.
[00:00:59.489]This is why the team is looking into using
[00:01:01.709]iPhones to measure bridges health
[00:01:03.459]by comparing them to other,
[00:01:05.227]more established methods.
[00:01:07.216]All modern iPhones have a built-in device
[00:01:09.514]called an accelerometer that measures
[00:01:11.814]when the phone has changed positions
[00:01:13.744]based on how it accelerates.
[00:01:15.864]This is used in apps like YouTube so that
[00:01:18.116]it knows when to show a video
[00:01:19.896]in landscape view and after you've been
[00:01:21.813]scrolling in portrait view.
[00:01:23.723]If we can use something so relatively
[00:01:25.762]inexpensive that can fit in a pocket,
[00:01:27.845]rural bridge monitoring could be
[00:01:30.105]greatly improved.
[00:01:31.435]BDI used strain sensors attached under
[00:01:33.975]the bridge that had to be set up by hand.
[00:01:36.337]This took a team of four a full day to set
[00:01:38.697]up and would therefore put a strain on
[00:01:40.817]the average county's budget.
[00:01:42.427]This is how critical and non-critical
[00:01:44.270]points were instrumented.
[00:01:45.790]The diagram is facing west and
[00:01:47.670]the picture is facing east.
[00:01:49.480]There are also two sensors on the bottom
[00:01:52.164]of the critical point that you can't
[00:01:53.893]see in the photo.
[00:01:55.023]We did damages and tested bridge R-210
[00:01:58.043]which is located on a small dirt road
[00:02:00.169]about 17 miles outside
[00:02:01.767]of Lincoln, Nebraska in rural
[00:02:03.757]Lancaster County.
[00:02:05.097]It is a simply supported bridge with
[00:02:07.287]steel girders and a concrete deck that is
[00:02:09.685]very typical of smaller rural bridges.
[00:02:12.208]We used a U-Haul truck that weighed just
[00:02:14.688]under 8,000lbs. to excite the bridge.
[00:02:16.858]The team did three cycles of static tests
[00:02:19.408]where the truck just sat on the bridge
[00:02:21.298]for a bit and then we drove the truck
[00:02:23.765]over the bridge in cycles of ten
[00:02:25.593]for 20 passes total at both 5 and 15mph.
[00:02:29.593]That's a grand total of 46 test for every
[00:02:32.938]level of damage starting at
[00:02:34.920]a healthy structure.
[00:02:36.885]All of those damages that I've
[00:02:38.715]been mentioning were done with
[00:02:40.455]a plasma cutter to either I-beams
[00:02:42.026]or channels at various points
[00:02:43.606]on the bridge.
[00:02:44.615]We are only going to be looking
[00:02:46.093]at the healthy structure for this
[00:02:48.235]presentation since SAP 2000 is unable
[00:02:50.437]to simulate the types of damage that we
[00:02:52.597]looked at.
[00:02:53.757]We are also only looking at the maximum
[00:02:56.248]overall strain for the structure as we
[00:02:59.021]are really concerned with what can do
[00:03:01.537]the most damage at a single spot and we
[00:03:04.406]are also very concerned about the team's
[00:03:06.673]overall safety.
[00:03:07.863]Let's get our bearings before we head into
[00:03:10.134]the results.
[00:03:11.084]The horizontal line is
[00:03:12.374]our critical girder.
[00:03:14.199]Specifically, this is the spot we are
[00:03:16.094]looking at for our first graph with our
[00:03:18.104]main results.
[00:03:19.464]Then, you can see where our
[00:03:21.051]two quarter points are, Q1 and Q3.
[00:03:23.394]The midspan is the middle with all
[00:03:26.674]the sensor locations that are not on
[00:03:28.844]either Q1 or Q3.
[00:03:32.564]As you can see here in the animation,
[00:03:34.834]the truck was driven from Q3 towards Q1.
[00:03:38.319]This might seem a bit counterintuitive,
[00:03:41.319]but remember we were driving the truck in
[00:03:43.609]both directions, and this is just how the
[00:03:45.799]westbound direction happened to play out.
[00:03:50.456]The fact that the field data seems
[00:03:52.496]to take place later than the model data
[00:03:55.599]is completely inconsequential.
[00:03:57.587]This is just because the field test
[00:03:59.577]started to run before the truck actually
[00:04:02.759]got to the bridge while the model just
[00:04:05.271]has the truck immediately go
[00:04:07.071]over the bridge.
[00:04:08.401]The main thing that we see here is that
[00:04:10.781]the model is predicting signigicantly
[00:04:13.331]more maximum strain for both axels
[00:04:16.738]as it goes over the bridge than what was
[00:04:19.058]actually recorded in the field.
[00:04:21.046]Here, looking at our secondary results,
[00:04:22.975]we can actually see the truck moving
[00:04:24.765]over time.
[00:04:25.705]You can see each of those axles
[00:04:27.847]as they hit those sensors.
[00:04:30.155]So, our back axle weighed a bit more than
[00:04:32.385]our front axle which is why the second
[00:04:34.635]peak always looks a bit higher on there.
[00:04:37.035]But the main thing that we need to take
[00:04:38.913]away from this graph is that the midspan
[00:04:40.995]is in fact our highest maximum strain
[00:04:43.697]and as I mentioned earlier, that's all
[00:04:45.567]that we really care about for this
[00:04:47.257]particular poster and this
[00:04:48.977]particular project.
[00:04:50.257]And so, we can go forward feeling
[00:04:52.364]confident that we are looking at the
[00:04:54.437]correct area of our bridge.
[00:04:56.177]So, to quickly review and analyze,
[00:04:57.883]the model predicted about 150%
[00:05:01.594]of the maximum strain from the physical
[00:05:03.784]bridge when we measured it
[00:05:05.534]at the midspan.
[00:05:06.714]And then our highest quarter point model
[00:05:08.769]strain was right around 32.03 Microstrain
[00:05:11.229]while the highest we actually measured
[00:05:14.413]in the field was only 23.39 Microstrain.
[00:05:18.413]And so that means that the model is
[00:05:21.903]overestimating at those quarter points
[00:05:24.063]by about the same amount proportionally
[00:05:26.606]as it was doing at our midspan.
[00:05:30.350]In conclusion, while the model seriously
[00:05:32.730]overestimates the maximum strain, it still
[00:05:35.264]captures the overall shape of the data.
[00:05:37.444]This is likely due to the physical bridge
[00:05:39.714]being stiffer than the model and therefore
[00:05:42.120]putting less overall strain on the I-beams
[00:05:44.340]and a bit more on the edges.
[00:05:46.081]While the model could be made
[00:05:48.041]more accurate before testing, that would
[00:05:50.110]require estimations of stiffness
[00:05:52.117]that could be wildly wrong.
[00:05:54.190]In extreme cases, this could lead to
[00:05:57.184]an underestimation of stress that might
[00:05:59.449]lead to dangerous damage plans.
[00:06:01.559]The model currently provides very accurate
[00:06:04.079]data on where the highest stress points
[00:06:06.329]will be while ensuring that stresses
[00:06:08.389]are not underestimated.
[00:06:10.238]As a result, the current model set up
[00:06:12.825]is incredibly useful for its purposes
[00:06:15.080]in this research, even if the strain
[00:06:17.114]numbers are unlikely
[00:06:18.676]to match the real results.
[00:06:20.676]Here are the sources I directly consulted
[00:06:23.540]for this particular presentation.
[00:06:26.605]I would like to thank the rest of
[00:06:28.179]my research group, Qusai and Emmanuel,
[00:06:30.585]the Lancaster County DOT,
[00:06:32.705]University of Nebraska School of
[00:06:35.045]Engineering, and funding for this project
[00:06:37.333]was provided in part by the
[00:06:38.793]National Science Foundation.
[00:06:40.485]Thank you for listening.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Comparing Field and Modeled Rural Bridge Responses to Moving Loads

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments