Water purification system for household

[mygoggie]

I have been very quiet on my Quad 33 and Quad 303 restoration threads mainly because I have been spending most of my spare time designing and installing a household sized water purification system.

The design incorporates the following components:

• Rainwater harvesting to a tank farm.
• Condensation harvesting using an underground piping system.
• A hand drilled well point.
• Wetland system to purify the house's greywater.
• Irrigation system to distribute the purified greywater into the garden.
• A biogas digester to generate gas from the blackwater and liquid fertilizer from the digestate.
• A slow sand filter to purify the rainwater, condensation and well point water to potable standards.
• A water purification skid to control all of the above and distribute the potable water into the house.

If anyone is interested I can keep posting as I proceed.

Here is the skid frame and cabinets assembled. Now busy with the internal pumps and piping. Then to tackle the electrical and control design.








Revision 10 of the greywater wetland is almost complete.





The purified greywater is pumped from here into the garden using various pipe circuits.





The slow sand filter being built. It was a experimental tank that was gifted to me to convert.

[HB]

Great thread Herman. Please continue.

Questions will follow. 

[peterc]

Very interesting!

Thanks

[mygoggie]

Today was spent rerouting my UPS system to watch the RWC while EKSDOM did its bit to alleviate our energy levels. In between this I spent some time to make the pump mounting plate.

This is basically a 200 mm x 60 mm extruded fibreglass channel onto which the pumps will be mounted. The idea is to provide a base that can be mounted onto the skid as a separate unit. I will be using rubber mountings to fix the mounting plate to the skid's base. Hopefully this will get rid of some vibrations and noise.

Here I have sawn the piece of channel to length and cut out the little corners to allow fitment in between the base frame members.





The hole locations laid out. I worked back from the required pump centrelines.  Yes, I had to work outside in the sunlight due to some strange absence of light of my workshop.





The first fitment with pumps mounted. I spent some more time later in the afternoon to determine the plate's final position and mark out the mounting holes in the skid and in the plate. Will take some pics when this is drilled and the plate finally located.

[Family_Dog]

Following with great interest, looks very neat so far Herman. 


-F_D

[vleisman]

What a great thread, subscribed.

[mygoggie]

I have been thinking about the switchgear equipment to use in the skid and for this reason asked some questions on another forum where I always get solid answers.

So basically after a lot of back and forth the electrical and electronic design will entail the following:

Electrical:

• A 4mm² cable from the existing household DB to the skid's control panel. The cable to be protected by a two pole 32A D-curve miniature circuit breaker. <= DONE
• At the control panel a lockable isolator and power on light. <= DONE
• For the control of the pumps there will be a manual or automatic switch.
• A stop/start push button for each pump.
• A running indicator light for each pump.
• A manual motor starter circuit breaker with 9A AC3 rated single phase contactor for each pump.
• A MCB for each other electrical item such as controller, transformer etc.

Electronics:

I will be using two Raspberry PIs to control things.

• The one PI is used as the computing system for an Opensprinkler irrigation controller expansion board. This is 24VAC as the solenoids and the controller itself uses 24VAC. This to form part of the system as all household greywater and waste water from the purification system will go into the garden via an irrigation network.
• The second PI is used in the same way, but with another RIO (Raspberry input output) expansion board for which I have written most of the C++ code to control circuits of a research biogas digester I have designed and built. This RIO can accommodate 8 to 40VDC operating voltage. Fact is I have stock of this so why not put it to use? I can just as well change this into an Ardinio of a PLC unit if this does not work in the long run.

Voltages I am aiming for will be 24V for controlling and 230VAC for the pumps and the ozonator. I would like to stick to these two to keep things simple and to a standard.

For now I would like to see version 1 of the skid up and running which will allow me to change and improve on the design as data is gathered. No system is ever perfect and the best route is just to get it working and then refine the design until you get to a durable and lasting system. Just my design philosophy.

[mygoggie]

I did a layout of the pipework required in the skid and then revised it and revised it and again revised it. I can do the design in CAD, but I love free hand sketching as using a pencil on paper is more "warm" for me than a mouse and screen.

So the final layout look something like this.









The final step was to mark out the area on the floor and did a draft layout to 5 mm accuracy of the pumps, physical pipes and fittings.







Doing such a draft layout is invaluable as it allows me to look at things from other angles and allow improvements before construction. For example I could change the pump intake layout to streamline it a lot more to ensure smooth flow in the approach to the pump. You will be surprised at how quickly any water turbulence in the pump intake line will increase pump impeller cavitation, shortening the life thereof very quickly.

Just for interest, the golden two rules for pump intake piping design is that the size of the pipe must be 1,25 to 1,5 times the pump intake port diameter and secondly then the approach pipe must be straight for a minimum length of 10 times the chosen diameter.

In this case then the pump intake port is 32mm diameter and I used a 1,25 ratio because the flow volume is low and the suction height is very low. If the suction height approaches 7m then you need to increase the pipe diameter to at least 2 to 2,5 x pump port diameter to prevent boiling of the water under the negative pressure. In this design I could get away with a 40mm diameter pipe. This then requires a final approach length of 400mm.

The following is NOT correct and will destroy the pump impeller within a very short while.

[mygoggie]

The past few days were spent stealing time here and there to refine design of the pipework in the skid and comparing it to what is practically possible.

I first completed the through hole layouts and drilled the holes for the piping. Yes, the wind blew for a few days at 40+ knots so leaves are part of our life and I need to clean in any case after cutting the holes. Here are the holes for the service pipes.





The holes at a closer range.





The holes leading to the front valve cabinet for easy operation by the users.





Yesterday I completed the installation of the external pipe clips and the entry pipe work.





Then onto the internal pipe clips. Here the holes are drilled, the rivnuts installed and the first pipe clip screwed down.




The pipework leading to the valve manifold are installed and clipped in place.





The pipe manifold which allows the user to flush the wetland and cross feed the tanks. The manifold also serves as an air trap for any air released during suction when pumping water from the feeding tanks to the slow sand filter. I first had a separate air release circuit, but this morning when I was designing the finer details of the manifold I realised I could do away with it as I have changed the inlet layout to the pumps. So now there are just the two valves with air releasing via the wetland scour circuit. I must say the manifold took some time to design as I had to work to 1mm accuracy to ensure the valve spindles line up perfectly with the holes of the to be installed valve cabinet.

All done and I tested the scouring this afternoon under gravity. Works great and all air from the feed lines are captured and released to atmosphere as planned.





The new look of the skid. Next step will be to mount the pumps and complete the pipework leading to and from the pumps.

[mygoggie]

It has been a bit back and forth to find the correct anti-vibration mounts for the pump mounting plate. The main issue was to find something that will fit below the channel section, be able to resist corrosion class 3 and still be not too stiff to absorb the cyclic rotation based vibrations of the pumps.

I finally settled on this mount which is specifically designed for the above type of application.




Will order tomorrow and hopefully it will arrive early next week.

[chrisc]

Those PD mounts could double as speaker supports too...!

[mygoggie]

Victor of www.microsolve.co.za was so kind as to provide me with a Franklin Electric motor control unit to test in my design. It has arrived at my local PostNet and I need to go and fetch it. Thanks Victor!

It should allow me to provide a robust user friendly system that even SWAMBO will be able to understand. Here is what the manual state about the various protections offered by the unit.

[mygoggie]

The weekend was a resting one mainly due to the RWC and the rest of the weekend spent recovering from the celebrations ...

This afternoon I spent installing the manifold pipe clips, drilling all the holes in the valve cabinet and doing the first assembly. Golden rule .. measure twice and then again and THEN trust your measurements, not what you perceive to be correct. Anyway, the valve spindle holes came out not in position as measured, but I could fix the situation by enlarging the holes one size up.

The valve clips in position. The rivnuts did not work. I replaced them with long hex nuts I had in stock.





Valve cabinet fitted with valve handles installed.





The valve cabinet secured by four cap screws on the sides. This will allow future removal of the cabinet if the valves needs to be serviced.





The final look of the skid with valves. Next step will be to install a third pipe clip on the top of the manifold as there is a little bit of movement of the manifold I am not happy with.

[d0dja]

I was following this thread to perhaps do something similar. But now that I've seen how one should do it properly, I may just chicken out. Fantastic work.

[mygoggie]

I was following this thread to perhaps do something similar. But now that I've seen how one should do it properly, I may just chicken out. Fantastic work.

Thanks @d0dja, feel welcome to ask if you need help to build your own. I have had a number of people express interest in getting such a skid in place. I think I must consider the option that this could become new venture. Let's first get this one working as envisaged and then I can explore such visions.