Water management system

[cheerio]

Abstract:
[spoiler]
As we live i a time where climatechange is actually happening we need to think about our water supply. I own some land that needs watering and the last few month the waterlevel nearly dropped below my well. If the pump started the waterlevel dropped below in a few seconds. (I live next to a river that is predicted to fall down to a brook in the future... i love the river )
At this point i startet cleaning up an underground cistern (~20000 Liter) that is still full of dirt. I am confident that i clean it up until september and make it waterproof.
I also got a big fountain that can store an incredible amount of water. But as the water in my region carries Fe oxide i cannot fill the water in the fountain directly as it paints the bottom of the fountain orange-red.
[/spoiler]

Time for the plan:
[spoiler]
I use the cistern to clean the water of the Fe oxide (it falls down after a few weeks) and fill the fountain with it. In addition i use the rain from the roofs and direct them into the cistern and the fountain. Now i got 2 very good water storages.
The underground cistern and the well need level indicators.
The pumps need to work as a team to
1. keep the pressure above a limit
2. jump in for a pump with malfunction
3. save energy (we do not want a climate change, do we?)

The system also should be able to
1. predict how long the water will last (dependent on the weather forcast)
2. build statistics to make us use the water and energy more efficiently (watering on many locations the same time is not so good i guess )
3. detect malfunctions and try to fix or work around them
4. prevent an overflow of the water storages
5. Send SMS if smth bad happens and send emails for notifications
[/spoiler]

Status: in progress
[spoiler]
Bus: CAN (still works on 5km cable, so it is good for extensions)


I made a pressure sensor that sends the water pressure in the system over the CAN bus. In early August i will hook on a logger to see if the sensor works correctly 24/7 and how much Current it will drain max.

I bought some solenoid valves that will do the job. I did make some notes on how i organize everything but i cannot find them anymore so i need to do it all again.
[/spoiler]

Timeplan:
[spoiler]
This project IS heavy. I continue on August and try to make a light version of this system until october. The timeplan to finish it will be in about one year. (i need that time because i cannot work on this much)
[/spoiler]


Please give me some feedback and some suggestions. I will post my notes and schematics when i start to continure the project.


*Watermeter:
[spoiler]
Those Watermeters are pretty expensive if you need to read them digitally and need 1 1/4"
So i tried a different approach. I write down the waterlevel in the Pressure tank when the pump stops, and i write down the waterlevel when the pump starts. The Pressure tank is kind of a cylinder so i can calculate how much L water will be pumped into the tank in 1 pump cycle. Now i detect the pump cycles in my statistics that i did witch my Breadboard can logger and multiply them with the value i calculated.

The result ist not exact as the watering is going on during the pump cycles.
Another problem is that the Pressure tank loses air over time and the L/pumpcycle changes because of this.
The aprox. error per day is <100L which is good enough for me.

I plan to compensate the ongoing watering by software. I also plan to automaticly keep the same "pre pressure" in the Tank by interfacing a compressor and reading the waterlevel in the pressure tank via capacitance.

[/spoiler]

Items:
[spoiler]
*pressure sensor: Honeywell 13C0500PA1K
*MCU(pressure sensor): Atmel AtMega32
*MCU(i2c 7segment): Atmel AtMega88
[/spoiler]

(02.09.2011):
*Watermeter

(08.08.2011):
*schematics(eagle) uploaded for pressure sensor(drucksensor)
*schematics(eagle) uploaded for 7segment display for pressure sensor

[kam]

well, this is a HUGE project. I totally suggest you use Mitsubishi Alpha PLC. It is cheap (compared to what it offers) and very powerfull. It takes port expansions for digital and analog I/O, and also it has GSM module to send SMS.

If you want to make it more sophisticated, you could use a PC with I/O PCI cards. There are cheap cards, but frankly, i would not rely on a PC.

[cheerio]

well, this is a HUGE project. I totally suggest you use Mitsubishi Alpha PLC. It is cheap (compared to what it offers) and very powerfull. It takes port expansions for digital and analog I/O, and also it has GSM module to send SMS.

If you want to make it more sophisticated, you could use a PC with I/O PCI cards. There are cheap cards, but frankly, i would not rely on a PC.

This Alpha is cool. But my plan is to set this up decentralized with many modules at different locations. So this Alpha is nice as a Main controller Station but for my plans it is the overkill.
The PC Solution is not an option for me.

I want to learn as much as possible with this project so i go the hard way

[kam]

hmmm i'm not quite sure if the hard way is the best way in this project...

[cheerio]

does not detect my edit a a new date ^^
so now (update)

[kam]

yes this SMF engine really sucks, i wish i could change it with another engine, but... what the heck... Regarding the water level. You say that you will use a pressure sensor and calculate the height of the water, which is IMO precise and cool. I do not understand why you need to start and stop the pump:

I write down the waterlevel in the Pressure tank when the pump stops, and i write down the waterlevel when the pump starts. The Pressure tank is kind of a cylinder so i can calculate how much L water will be pumped into the tank in 1 pump cycle. Now i detect the pump cycles in my statistics that i did witch my Breadboard can logger and multiply them with the value i calculated.

The result ist not exact as the watering is going on during the pump cycles.

In large ship tanks we use to have long PVC pipes with a floating piece inside. The higher the level on the tank, the higher the floating piece will be. Normally, the piece goes through limit switches, but you can do something else: You can put an ultrasonic transmitter on this piece, and measure the time it takes for the sound to go all the way up the PVC tube. I have not test it, but... it could actually work.

[cheerio]

yes but this is too complicated
The tank is only for keeping the pressure it is NOT smth like the cistern.
The tank is ~300L. The pump has to switch on and of to keep the pressure.
The pump can transport much more water per second than all the watering needs. Over a whole day i can save much energy because the pump is turned off most of the time.
The tank is pretty old so i will try to detect the waterlevel in it w/o changing anything at the tank. (i am happy it does not leak so much ^^)

[George]

I noticed that you are using 1 1/4 inch pipe - is that correct ?

I helped my brother put in a system quite a few years ago - we used 1 1/4 inch pipe - big mistake.

Reason was we had uneven water flow rates across the grid of relay controlled watering devices ( sprinklers etc), at the outlets.

For making your design - consider the diameter of the pipe as a resistance to water flow. Smaller pipe --> bigger resistance.

If you can go min 2" pipe, then reduce to smaller size at the water release point. This will help water flow at all outlets, whilst decreasing

the pressure needed to do so.

I hope that helps a little with the prectical side of your project.

Cheers

George.

[cheerio]

I did some brain work after the shovel work and finished the software for the Pump and Valve switches. When i finished making this: http://www.kreatives-chaos.com/artikel/can-debugger
i can use their CAN-Bootloader to flash my devices.
I encountered some stability problems on my pressure sensor so i will relayout the pcb and change some stuff on it.

Soon to be updated.
@kam sorry i had not time to look into your new project manager yet

[kam]

Good news are that you are still working on this (mega-)project!

[cheerio]

I set up the project on source forge. I think a SVN is the best idea to keep it updated the convenient way.
Code(AVR_Studio 4.19): http://sourceforge.net/p/watermanagement/code-0/2/tree/
PCB(EAGLE 6.2): http://sourceforge.net/p/watermanagement/eagle/2/tree/

The Directory structure of the svn is a bit messed, but i think you will have no problems.

I still have no time to check out the new project tool of this site, but i will when i have finished the project.

[cheerio]

I am currently building a waterlevel-meter module. I use an iron pipe (1") with a wire in the center. This is a capacitor that will increase its capacity when the water level rises. That is because the permittivity of Water is significantly higher than the permittivity of air.
If you connect  this sensor-capacitor as the timing capacitor to an astable multivibrator you get a waterlevel->frequency  transformation.

[kam]

hmmmm.... nice idea, i've never thought of it... Some questions:
1. The pipe and the wire must be insulated, right? How do you plan to insulate the pipe?
2. Rain water has different composition every time. Won't this ruin your measurements? Have you make tests with acids or salts?

But it really is a nice ides...

[cheerio]

the wire must be insulated. you should aim for teflon. the pipe is connected to gnd. If you connect the GND of the circuit to earth you do not need to insulate the pipe. The different composition of the water is a problem indeed. But i did not run field tests yet. I am confident that the quantity of the water (~20m³) will keep the water composition pretty much constant.

[kam]

sounds logical...