Water Level Monitor
John Board
Posts: 371
G'day,
I was recently approached by the Ambulance (EMT) service in a town that I used to live in to design a potential system to monitor the river level of a nearby river. The river normally flowed at 4 meters (13ft), however it quickly rises during the monsoon season, often easily up to 12m (39ft), and while I was living in the town peaked at 16.5m (54ft). They've thought of a few ways to monitor it manually, but were wondering if I could come up with a pricetag for a theoredical automatically monitoring system.
I've come up with a few ways to sense the river level, but I was wondering if you guys had any better ideas. Here are my best ideas so far:
There's a bridge over the river, which we could attach something onto to measure the level. The only catch is the bridge is only 8.5m (28ft) tall. The ideas for attaching to the bridge is simple, the standard system of having a long strip where at certain intervals you measure the resistance/conductivity between two points. Depending on how many points show conductivity, shows how many are underwater, and as such shows you how deep the water is.
Second idea is having yellow markers painted at meter (?) intervals on a tree or really long pole in the river. Then you can use CV to figure out how many yellow markers are visible, and thus how deep the water is.
If we used the first idea we could easily put some kind of MCU on it and could power it or 6 months at a time, easily. If we used the second idea, I would likely look at some cheap embedded computer (Quad Core RPi!) on board with OpenCV. I would probably also have an external microcontroller with RTC that told the RPi to turn on at certain times in the day, the RPi would stay on for 15-30 minutes (allowing time for remote maintenance if required), and then shutdown again.
The values then could be reported to a computer via WiFi or Bluetooth to a laptop or phone. Alternately I could put a GSM modem on and have it report the values via SMS/internet.
The power could be an SLA car battery (or something like that), potentially solar charged.
Any ideas on better ways to measure the level of the river, how to report back, or how to power it would be greatly appreciated!
-John Board
[EDIT] Here's footage of the river when it's at 10 meters: https://www.youtube.com/watch?v=JwfvIg21YSU
I was recently approached by the Ambulance (EMT) service in a town that I used to live in to design a potential system to monitor the river level of a nearby river. The river normally flowed at 4 meters (13ft), however it quickly rises during the monsoon season, often easily up to 12m (39ft), and while I was living in the town peaked at 16.5m (54ft). They've thought of a few ways to monitor it manually, but were wondering if I could come up with a pricetag for a theoredical automatically monitoring system.
I've come up with a few ways to sense the river level, but I was wondering if you guys had any better ideas. Here are my best ideas so far:
There's a bridge over the river, which we could attach something onto to measure the level. The only catch is the bridge is only 8.5m (28ft) tall. The ideas for attaching to the bridge is simple, the standard system of having a long strip where at certain intervals you measure the resistance/conductivity between two points. Depending on how many points show conductivity, shows how many are underwater, and as such shows you how deep the water is.
Second idea is having yellow markers painted at meter (?) intervals on a tree or really long pole in the river. Then you can use CV to figure out how many yellow markers are visible, and thus how deep the water is.
If we used the first idea we could easily put some kind of MCU on it and could power it or 6 months at a time, easily. If we used the second idea, I would likely look at some cheap embedded computer (Quad Core RPi!) on board with OpenCV. I would probably also have an external microcontroller with RTC that told the RPi to turn on at certain times in the day, the RPi would stay on for 15-30 minutes (allowing time for remote maintenance if required), and then shutdown again.
The values then could be reported to a computer via WiFi or Bluetooth to a laptop or phone. Alternately I could put a GSM modem on and have it report the values via SMS/internet.
The power could be an SLA car battery (or something like that), potentially solar charged.
Any ideas on better ways to measure the level of the river, how to report back, or how to power it would be greatly appreciated!
-John Board
[EDIT] Here's footage of the river when it's at 10 meters: https://www.youtube.com/watch?v=JwfvIg21YSU
Comments
After posting I thought of that. Originally I rejected the idea, as flowing water would affect the readings. However, what I realized is that there is a weir about 30 meters upstream from the bridge. If we sunk a pressure sensor upstream of the weir, right at the bottom of the wall, there would be hardly any water flow, and the readings would be a lot more reliable.
Thanks for your response!
http://en.wikipedia.org/wiki/Tide_gauge
http://en.wikipedia.org/wiki/Level_sensor
I've used a long range ultrasonic transducer for a similar application.
Basically, it hangs over the water ( say on a mast at the weir) and gives an analog signal for a suitable range (say 0 - 60 feet). Its called ToughSonic and made by Senix in the states. Not cheap ( about $500) but made to survive.
You could even directly connect it to an Xbee radio and drive a chart recorder or similar device closer to civilization.
Cheers,
I was asking around last night, and someone pointed out that as well. It's funny, it's stupidly simple but I didn't think of it! Thanks giving me part names!
-John
we need like and dislike buttons here...
Enjoy!
Mike
The best systems use a stilling well off to the side of the river or body of water, with a pipe going horizontally out deep under the water. The pipe is positioned to minimize flow pressure and to average out the hydrostatic pressure.
The combination of an encoder with a beaded cable and float on one side and counterweight on the other can provide a huge range, easily your 20 meters, with millimeter resolution. They are reliable but do take some maintenance to keep the inlet free of silt and debris. Strange things can happen, for example, in one glitch a rat fall into the well and crawled up on top of the float to expire.
Tall stilling wells from the USGS have a ladder inside so that someone can climb down, made out of steel culvert material. Smaller ones are made out of something like 8 inch PVC. These are usually located in a segment of the river where the banks are rocky or constrained in a manner that allows for a calibration curve, so that the volume flow of water can be derived from the level. Sometimes there is a V-shaped flume with a stilling well at each end, which allows for the best volume flow estimates.
Hydrostatic level probes are convenient, but be sure to get one that will withstand the pressure at maximum flow. Pressure transducers tend to break at something like twice their rated pressure. The tradeoff is that the repeatability is something like 0.1% of full scale, so a wide pressure range means lower resolution at low water levels. The ones sold by Unidata use a differential pressure sensor and a cable with a vent tube to atmospheric pressure. The vent tube has to stay above the maximum water level and out of the weather. Watch out for kinks. Some hydrostatic probes use an absolute sensor underwater and for best results a barometric sensor above to compensate for atmospheric pressure. That works okay for deep water.
Bubbler systems commonly use a low flow provided by a peristaltic pump. Rapid or uneven flow creates errors. It just needs to keep the tube full of air to the bottom bubble release point.
Never underestimate the force of moving water or the devastation that can be caused by debris flow. Don't leave anything exposed that can be caught. People who work in that field usually have stories to tell. One case I recall, a 1200 pound steel flume was pulled out of the banks and carried a half mile downstream until it crumpled against a bridge abutment.
It sounds like an interesting project. Maybe it doesn't require exact measurements, just a warning about when it reaches certain benchmarks?
Thanks for all the interest and responses.
The Stilling Well idea seems interesting, the only problem is, I can see it being quite expensive. This town has 500 inhabitants, and I don't imagine the council would want to spend that kind of money on them. That said, there is a large coal mine near by, which is affected severely when the bridge gets covered with water (can't transport the coal out, loose tons of money a day). If a case could be made to them for the usefulness of the project, they could easily provide the funding (they fully paid for an entire new bridge over the river because they were loosing so much money when the old one got flooded over).
As I said before, we are concerned that the flowing water would affect the readings of a pressure sensor - however with those suggestions we could get around that. The only concern with that is the entire rig is under water - maybe I'm being padantic, but I'd rather keep as much of it out of the water as possible.
I've done some more talking with the Paramedics up there, and they've also had a chat with the Police, and they're interested in the following idea:
If you look at the video in the original post, you can see on the weir there is a metal gantry - this metal gantry has always stayed above water in all past floods. We are looking at putting some kind of ultrasonic (perhaps the one suggested earlier in the thread) on the gantry along with a processor, and a big battery. It would then take readings at intervals. Those values could then be sent back to base using either a GSM modem or some kind of RF setup (XBee, although I've been looking at the AC4490-1000) which would send the data to either the Ambulance or Police stations which are at most 2-3km away. While we are doing it we may as well put a camera, and some other sensors (flow sensor, rainfall sensor, temp, humidity, barometric, etc) on it too.
The advantages of this is that it's reasonably low maintenance (no cleaning for dead rats or silt), and it's all above water.
The disadvantages is that you might not get mm or cm accuracy.
Ultimately the only real maintenance that you would have to do is recharge the battery every 3-6 months.
[EDIT] Here's the location of the weir. To the right of the weir on the map is the new bridge (not shown in the original video), and a secondary smaller bridge further down stream. The gantry is on top of the weir, towards the south. It would be transmitting coordinates to 34 Stopford street, or 79 Stopford Street (Main road of the town, south. 0.5 - 1.09 km from gantry)
https://www.google.com.au/maps/place/Baralaba+QLD+4702/@-24.1710079,149.8140124,194m/data=!3m1!1e3!4m2!3m1!1s0x6bc149047f5e68f1:0x400eef17f20c980
Their website doesn't show any design details (still "under construction" ) but the original used an old Siemens cellphone for sending data from diff pressure submersed xmtr.
I believe the kayakers use an updated version of the same contraption...
... quick update - for more info see http://wiki.iww.ie/index.php/River_Gauges