Plan is to put 650mm unperforated pipe at the inflow and outflow of wetland then use submersible pressure level sensor (something like this: Impress sensor - IMSL silicon level transmitter) within the pipe to measure the water depth. From that I can roughly calculate the flow into and out of the wetland to build up a water budget.
I have no experience with Arduino so at this stage I'm just seeing if this would be feasible as to buy something that does this off the shelf would cost me over £3000.
I have two main concerns:
How to power the sensors and Arduino unit? Could use sealed marine leisure batteries or solar
Compatibility of the sensors with Arduino Uno (or similar). These sensors are often marketed as 'plug in and play' so I'm worried that they would have trouble connecting and storing the data.
As I said I'm just scoping out this project so any constructive feedback or criticism is very welcome.
I'm not sure a Uno would be the best choice for power consumption reasons, but there is a wide range of Arduino devices for more "IoT" type applications.
The critical question for power is how long the system will need to operate before needing attention? Having said that one could consider using a sealed lead-acid battery like you mention with a small solar panel to keep it topped up.
The pressure sensor could have a wide range of supply voltage and interface specifications. It isn't clear what current it needs though one of the options is current loop 4 - 20mA which gives some idea. Would you need to take continuous readings or just samplings, and if the latter how often? The sensor current could turn out to be quite a significant fraction of the whole consumption. It seems to need an analogue interface which an arduino can manage but depending on the resolution you want an external analogue-digital converter might be needed.
The turbidity sensor looks like it could take even more current - same questions apply. But this has digital interfaces which are easier to interface to an Arduino.
How will you recover the data, or will it be communicated as it is gathered? What wireless comms interface? If it might help I know of some people who have deployed flood sensors using LoRaWAN to gather the data, I could put you in touch if that might help?
The overall product design would need to be done carefully to resist the wet environment. Sounds like an interesting project!
Would you need to take continuous readings or just samplings, and if the latter how often?`
Hoping to do continuous sampling (both sensors) for a year at 15/30 minute intervals. I will be visiting the site every two weeks to collect water samples so can check the system and top up batteries then if needed.
How will you recover the data, or will it be communicated as it is gathered? What wireless comms interface?
Because I'm visiting the site every couple weeks I was planning on collecting the data manually (USB or some other format?). One less thing to worry about if I eliminate wireless comms
The overall product design would need to be done carefully to resist the wet environment
Yes that's another issue. The likelihood is that at some point the flow going through the pipes will be too great and start to spill over the top which will render the water level measurement useless. To capture extreme highs and lows I was thinking of using a camera trap and water level meter to 'validate' what the range of water level measurements I can capture. At this point I'm just accepting that it will be an estimate of flow as proper measurement is too difficult in these messy systems.
It's not just the water you are measuring you will have to deal with, it's the entire environment. Rainwater tends to get in everywhere, especially when you add in pressure changes and temperature changes. Even excess humidity can condense and cause corrosion issues.
I'm not sure where in the world you are, so your conditions are probably different from mine. I'm in the Toronto, Canada area, and to give you an idea, here are a couple of problems I have to find solutions for:
Overnight winter temperatures around -20c (-4f) or so mean I can only use batteries if I can manage to put them in a heated enclosure.
Rainwater, melting snow, and humidity are probably going to require that my circuits are encapsulated in some form of sealant. I'll probably end up simply pouring casting resin over the entire board to seal everything. This is even with the boards being in some sort of 'weatherproof' container.
I have to watch my power budget very carefully. I have to assume that after a week of overcast weather in December, my system will still be able to get enough power from the solar panels to run through the night.
I'm doing this in Scotland so just as wet but we don't have the extreme temperature changes.
Was planning on putting it in some waterproof Tupperware but casting resin over it sounds like a good idea, especially considering it's going to be outside for a year.
As for the power issues. I might just hook up a large lead acid battery to keep it going for a while, from what I've heard they won't have an issue with our cold (max -5 Celsius)
You can buy proper environmentally rated electronics enclosures that have gaskets etc to seal them. Look on RS or CPC Farnell for IP65 enclosures. Don't rely on Tupperware!!
You could use a USB stick to gather the data but you would have to have an IP rated connector with a screw-on cover. You might consider instead using an Arduino with on-board Bluetooth and collect the data to a mobile phone - I believe there are terminal apps you could use for this. Then the unit could be totally sealed except for the power connection and there are connector types that will do that. Obviously you'd need connectors for the sensor cables anyway.
Use resin with care!! It's OK if you know it's working but once set repair is impossible, and you would at least need to make sure you bring out connectors for programming. Better to use well specified connectors that provide a moisture barrier in a sealed enclosure.
Powered circuitry that is exposed to high humidity will corrode very rapidly, sometimes becoming useless in a matter of weeks, especially if temperature changes lead to condensation forming on PCBs.
If circuit boards are not completely potted (e.g. buried in silicon compound), they at least need to be "conformally coated": everything painted with clear nail polish.
The battery runs the unit for 42 days 41 nights without a charge: no MCU sleeping.
In the shield the electronics are exposed to but protected from the elements.
That project has been running for several years. I've not replaced any parts, yet.
Pound a pipe at the inlet, allow water to seep into the pipe, measure the height of the water. Pound a pipe into the outlet, allow water to enter the pipe, measure the height of the water.
You could have a ESP32 at the outlet that sends its data to the inlet. The inlet ESP32 could do the storage and calculations.
Great advice. A so-called sealed container is NOT sealed if any wires pass through the walls. Air pressure changes will pump moist air in an out and when moist air condenses you get liquid water on the electronics.
Best to ensure air circulation to keep things dry.
Are you able to put the electronics out of the "flooding-zone" a few meters away?
Or in a height that will never be flooded and still is safe against windstorms to be not blown away from this exposured position?
Is this wetland at the sea = salted water or far away from the sea = no salted water?
The website says nothing about the power-consumption of the pressure-sensors.
They look very professional and durable but in your application power-consumption must be known prior to buying them.
Depending on the precisision you need maybe different principles to measure waterlevel could be used.
Can you post a picture of how the wetland looks like or draw a sketch freehand how ground, pipe and mimium and absolute maximum (everything overflooded to the max) waterlevels look like.
Rather than just dive straight into electronics I would first see if the sensor technology You are planning on using actually works and gives the results you want .
You can also do a lot of work at your desk working out power requirements , how you will power it and the costs of the parts .
After that as said don’t under estimate the environment - sandwich boxes rarely work long term , you need proper enclosures , cable glands etc . ( you can cheaply test your sandwich box !)
Really to develop a system for less than £3000 is a difficult ask , so you need to do your research carefully .
I have used The Things UNO to monitor river water level (using a JSN-SR04T ultrasonic sensor) and temperature (using a DS18B20 Waterproof Temperature Sensor) in our dyke
communication was over LoraWAN to the MyDevices/Cayenne desktop
communications depends on what is available in the area which can be very limited in remote wetlands etc
in the past I used the MKRFOX1200 but I am not sure how viable the Sigfox network is at present
never used the MKR WAN 1300 (Lora) board but the Things UNO Lora board worked OK (see post #14) - however, I did have a LoraWAN gateway gateway installed in the house
also used Quectel EC21 modem using the MQTT protocol but again mobile coverage can be limited in remote areas (until recently we had to stand on a chair on the upstairs balcony to use a mobile phone and even now some networks have very poor communication in the area))
