Greywater recycling MBR system preliminary design
Edit: I just realized I should probably explain what MLSS and MBR stand for, since reading all my previous posts might be a bit much. It is Mixed Liquor Suspended Solids, and refers to the amount, in grams, of particulate solids in the water, per liter. MBR stands for membrane bioreactor.
Some experimentation might be a good idea, and testing the water once or twice to make sure everything works as planned, by sending some off to a lab, is also a good idea. This system can, of course be scaled up, but let’s suppose it is made to filter 70 liters per day.
The mlss sensor here could be an optical turbidity sensor, but I chose the continuous level sensor with a long spring attached type.
Okay, now, going with the water flow, the first stage is the surge tank. This should be 80 liters, to accommodate worst case scenarios, and still have some headspace for bubbles to break. Water flows into chamber 1 by gravity of course, and bubbles also escape through the same opening, causing a limited amount of mixing and aeration in the surge tank. This matters because it is preferable that the the storage tank be aerated at least a bit, and there is another aerator foot inside the surge tank for that reason, because the bubbles here might not be enough to aerate the whole tank. Of course that depends partly on the shape of the tank. I’m thinking the tank would be a big piece of water pipe sideways, at a slight slope (which is not shown in the diagram) to ensure water heads toward the entry to chamber 1, even if the system is tilted slightly.
The aerator feet all have manual valves attached so you can adjust the air flow if you find it necessary. Remember, these aerator feet are just plastic tubes with holes poked in them with a thumbtack or something. That’s probably a good size of bubble, though finer bubbles will increase energy efficiency, I figure better to do things the easy way for a first generation prototype like this. BTW, water cannot flow into the aerator feet if the holes are small enough because of capillary forces. Otherwise a check valve on the air line would be a good idea.
The surge tank has the cutoff valve, which cuts off the water flow from the supply tank (which is also not shown in the diagram) to try to make sure that no water can flow out of the faucets etc. when the tank is full. You might make that a float switch instead, and have it turn off the drain valve or something. If there is a dishwasher or something, you need to account for that. You could have sensors in the halfway point of both the surge tank and the clean water storage tank, if the level is ever above the halfway tank in both tanks, the PLC knows there is too much water in the system, more than the 70 liters, and can sound a warning or something.
Okay, next is chamber 1, with the float switch so the plc knows when the surge tank is empty, and it can turn off the pumps etc. to save energy and avoid draining the MBR chambers. The aerator foot might as well be directly against the bottom of the chamber floor here.
Then there is the tube connecting chamber 1 to chamber 2, it passes through the float valve to cut off water flow when the level is too high in chamber 2. That makes sure there is some headspace in chamber 2. More importantly, this ensures that water doesn’t rush into chamber 2 in large quantities, reducing any spikes in the effluent output.
Next there is the MLSS sensor. In this case, there is a baffle nearby to reduce the amount of water flow and turbulence around the sensor, making it a bit easier to get a reading, though you might not need that. The PLC should read this intelligently, taking an average over a period of time, because I bet it would be a pretty noisy signal. As I mentioned, it is a continuous level sensor of some kind, could be through the wall if you want, whatever. The spring could be a stainless steel spring, or rubber, whatever, the more stretchy the material used (how long it can be stretched compared to it’s original length) the higher the sensitivity. Spandex from a sock? Hey it could work. Very small diameter Silastic tubing? You might also be able to use a weight attached in the right way. There might be a spring built into the float sensor that could also provide a force-displacement relationship that could be used.
Basically, the more dense the surrounding fluid is, the more upward force there is on the float sensor. It just has to change position when this happens. It should be able to sense a percent or two change in density.
Next is the membrane module, and the air scouring aerator foot, something I wrote about in a previous post.
Below that there is a baffle. It’s purpose is to reduce water flow and turbulence in the area below, where sludge is supposed to settle. The baffle might have to be more extensive than that, blocking off more of the pipe’s cross section, or you could put this area as a separate chamber. This is one of those discretionary things. Another advantage of a separate sludge settling chamber is that it can double as a pasteurization chamber. I chose to use the flow through pasteurizer, though.
The air output of this chamber goes into the top of the surge tank, so that water droplets have time to settle out a bit before going into the air exhaust line. That’s optional, but it might be a good idea to use a polypropylene air filter, which can withstand getting a bit wet. The air filter should probably be there, in case the MBR malfunctions and anaerobic bacteria grows and gets released out the air line. Aerobic sewage treatment systems don’t seem to filter the air, so this is probably optional.
The “airlift tube” is used to move some of the settled sludge back into the first chamber, which is done when the MLSS in the second chamber gets too low (since the sludge will flow into the second chamber anyway). When air escapes from the aerator, one way to thing of what it does is to reduce the effective density of the fluid in the column, so there is a net buoyancy force acting on the water in the column, and water flows into the column until the pressure at the bottom of the chamber in the airlift tube is the same as the pressure at the bottom of the secondary chamber. That equilibrium should never be achieved, though, because the water + bubbles escape from the top of the column before the height of the fluid in the airlift tube is achieved. How much air flow is required to do the airlifting depends on the bubble size, because small ones rise more slowly relative to the water. I noticed you can get 20 micron stainless steel aerators for aerating wort that would certainly produce plenty small bubbles, but I would be concerned that the holes are so small they might get clogged (mineral scaling might do that). I think just poking some holes with a sewing needle in some plastic tubing should be fine, anyway. A manual valve could be used for fine adjustment, maybe I should have included one, but it is optional, as long as the airflow is sufficient, there is no need to worry about it being excessive (except that the air pump has to be able to provide it.)
Next is the flow sensor, which is used by the plc to monitor the MBR’s performance, slow down the flow if it is too fast, by turning the pump off or down.
Next is the pump. What pump this is, I don’t know, but I have heard that aquarium pumps are usually made from food grade materials. Still, might be better to get a pump made for pumping potable water. The pump should not be capable of exerting too much suction force, or that could damage the membrane. How much exactly depends on the membrane used. You could run a DC pump at lower than the intended voltage or something, just run it off a wall wart.
Next is the sterilizing filter. Sound expensive, but you can get these for $50. There should be redundant disinfection at least, and this counts as one disinfection stage. The UF or MF membrane used in chamber 2 sort of counts, but it’s not a very good disinfection stage, because it will probably get a hole in it at some point, and when it does, the plan is to not fix it (since it will still work fine to filter almost all the sludge out). This can be checked periodically with the bubble point test, probably would have to buy a little hand air pump for that purpose, and keep it nearby the MBR.
Next is multifiltration. I mentioned that in a previous post. Deciding exactly how much of what filter media needs to go in here depends on what the output of the MBR proper is. See my previous post on multifiltration.
Next is ozonation. Again, you need to check this is working by smelling it or something. Don’t inhale too much, though, and the top of the column should be vented outside, for sure. I don’t know if the column and the contact tank after it should be made of metal or glass or something, but it’s probably a good idea. First the water goes into a column, it’s a column so the bubbles of ozone + air have a longish way to travel, ensuring a lot of the ozone is dissolved (though it doesn’t take much, actually). Then it travels through a tube for a while, that’s the contact tank. The ozone needs to be removed, too, which could be by passing it through a GAC filter.
Come to think of it, now, it might be a better idea to have way to ensure no bubbles enter the ozone contact tank, or the air will get around to the gac filter and cause minor malfunction after it accumulates (since it would not readily pass through the gac filter bed).
Last is chlorination, and you might optionally have a remineralization stage, or maybe you could add some salt to the chlorination solution, because demineralized water is fairly corrosive. But maybe there would be no need. I’ll try to put up another post about that later, an easy chlorinator that should be a bit more dependable than the other one I mentioned.
The chlorination could also be using in place of the ozone for the second redundant disinfection, with a few ppm of chlorine and a few liters of contact tank you could kill everything except crypto, but in a homemade system, maybe it wouldn’t do any harm to have the extra assurance of the ozone.
Then it goes to the clean storage tank.