Rainwater reuse planning

[dnb]

I have rainwater reuse in my plans, but the build budget is tight (thanks to current events) so I need to keep things cost effective. I've noticed that there's a lot of expensive kits for using captured rain water and most of them look like I can make something almost identical for considerably less outlay. Let's say the budget is around £750 to £1000. (This is based on rounding up 33% of our expected water bill for 5 years, plus an allowance for things getting expensive) I plan to use the water for toilets, garden, car washing (with a decent filter!) and possibly for clothes washing (again if filtering can be cost effective).

 

ICBs look like a good and cheap option for water storage, but I'm a little concerned about cost and complexity from burying them in clay and expecting them to stay connected, so I am considering an onion type septic tank as a rainwater tank. It seems considerably cheaper than a rainwater tank for the same (4500ish litre) volume and with very little work, e.g. to implement filtering and a calmed inlet, will do the same job. I assume the price difference is due to production volumes, because the calmed inlet and filter are often additions.

 

My plan was to fit a header tank in the attic since this looks to be the simplest arrangement with the lowest running cost. Mains water can be directed into the header tank with a simple float valve maintaining sufficient air gap and I can implement a control system to turn off the mains supply with a motorised valve (probably a ball valve since it doesn't use power when in either state unlike a solenoid valve - and it can be manually moved when there is a power cut) when there is water in the main tank.

 

This leaves the tank water side. A submersible pump, a pressure based controller and a float valve (probably one with a sharp cutoff) is again a simple and relatively inexpensive option. It does everything I need except sensible dry run prevention. (My theory is that dry run protection built in to the pump probably shouldn't be routinely relied upon.) So in addition I'll need to implement a tank level indicator of some kind and controller to switch the pump and the mains valve. (Sounds like another little microcontroller project, or something for a cheap "smart relay" device...)

 

The tank and the pump use up most of the budget but I have a shed full of random spare parts that will more or less do the rest. What am I missing (apart from a lot of implementation detail with many a devil lurking inside)? Any regs preventing me from using a new septic tank for this purpose for instance? Can I do anything cheaper or more simply? I note there are lots of German and Polish parts on ebay that are very well priced compared to UK parts. Are we that far behind the curve when it comes to sensible water usage?

[nod]

It’s funny really 

Myself and my wife where discussing this last night while we had no planning requirement to water harvest we wish we had and will on our next build 

 

Your plan seems a sensible and cost effective approach 

Ill watch this thread with interest 

[Conor]

I'm planning a row water system as well. But, as you say, commercial systems seem to be around £3k. Not justifiable at that price.

 

I assume you have a water main supply?

 

Could consider what I'm doing. I'm I stalling a seperate pipe network for the toilets etc. Initially, I'll just have this connected to the main cold water manifold. But it will allow me to disconnect and hook up to a different supply in the future. That would likely be a rainwater tank with pressure controlled sump pump.

[Gone West]

In my previous Victorian farmworkers cottage we had the original rainwater harvesting system which the early occupants used as their water supply. There a bit more about it in an earlier thread.

 

[dnb]

1 hour ago, PeterStarck said:

bit more about it in an earlier thread

Thanks. That will be very useful.

 

2 hours ago, Conor said:

I assume you have a water main supply?

I do. It is currently connected to a single outside tap near the old garage while we wait for the house.

 

2 hours ago, Conor said:

I'm I stalling a seperate pipe network for the toilets etc. Initially, I'll just have this connected to the main cold water manifold.

I will be doing similar this whatever the outcome of this investigation. I am currently of a mind to use a header tank to gravity feed water to all the toilets etc so that it's easy when the time comes to manage air gaps etc. so I can avoid potential expense with needing low volume production parts that require some kind of approval, and in theory can get away with a smaller (and hopefully more efficient) pump. We'll see if the maths supports this hypothesis. Space for the header tank isn't a concern.

 

3 hours ago, nod said:

Your plan seems a sensible and cost effective approach 

Thanks. Nice to have a good review! I am not obliged to do this by planning, but it helps with the water use part of building control. SWMBO wants a large bath and some kind of high volume pressure washer as a shower so the deductions from water reuse are extremely helpful. (Our BCO has indicated he's keen to see compliance with the water usage)

[Gone West]

13 hours ago, dnb said:

I am considering an onion type septic tank as a rainwater tank

Using  an old type onion cesspit would be better than a septic tank because septic tanks usually have baffles or separators built into them which would make using them as a water storage tank more difficult. The old type ceespits were just a simple tank which had to be emptied when full as there is no outflow, as there is with a septic tank.

[arg]

14 hours ago, dnb said:

This leaves the tank water side. A submersible pump, a pressure based controller and a float valve (probably one with a sharp cutoff) is again a simple and relatively inexpensive option. It does everything I need except sensible dry run prevention. (My theory is that dry run protection built in to the pump probably shouldn't be routinely relied upon.) So in addition I'll need to implement a tank level indicator of some kind and controller to switch the pump and the mains valve. (Sounds like another little microcontroller project, or something for a cheap "smart relay" device...)

 

A couple of thoughts (not based on any experience, yet):

 

If you are doing a microcontroller project, you could potentially detect the pump running dry by sensing the mains current drawn by the pump (with a simple sense coil).  As you say, you wouldn't want to leave the pump running dry, but just for a few seconds as the tank empties and then locking it out until it rains again probably wouldn't hurt.  Probably more reliable over the long term than trying to sense the water level in the tank.

 

The roof tank with effectively two float valves set at different heights does seem like a neat option, though I'm not sure about the regulatory issues (need to make sure the overflow is such that runaway pumping of the rainwater can't reach the level of the fresh water inlet?).   But as a completely different approach,, how about just running the mains water into your underground tank (perhaps manually) when it gets empty?   If you think your use vs roof area vs storage volume means you only need the mains water occasionally,  the effort of manually topping up when the pump system has alarmed and shut down is offset by the fact that the pump controls etc are now very simple and need little maintenance.

[PeterW]

23 minutes ago, arg said:

how about just running the mains water into your underground tank (perhaps manually) when it gets empty?

this is what I considered - just a solenoid valve that triggered for 5 minutes then locked out but also an indicator light that was reset manually to ensure it didn’t run constantly without being noticed. 
 

I’m thinking one of the oil tank sensor clones for the level sensor as it should work pretty well. 

[dnb]

2 hours ago, arg said:

you could potentially detect the pump running dry by sensing the mains current drawn by the pump

Nice idea. I can reuse code (and some hardware) from my solar diverter project. Would be worth doing this in addition to my 1st plan because it will help to detect more fault conditions.

I have a prototype microcontroller based tank level sensor I made for my father-in-law's yacht water tanks - it works on air pressure so has no moving parts in the tank itself (he thought it important for improving reliabiliy and repair time) , so was going to use this to switch off the pump and switch on the mains water when the rain water tank level got low.

 

3 hours ago, arg said:

only need the mains water occasionally,

I have to admit I haven't got a good computer model of this yet, (It's what I do at work) but my best guess is that rain capture will be almost seasonal. There will be much too much in the winter, and then no rain for ages in the spring and early summer, followed by vague sanity in August, September and October where the rain is distributed more evenly.

 

3 hours ago, arg said:

not sure about the regulatory issues

Neither am I. But I suspect that an overflow that can keep up with the pump and the fresh water together will be OK if it keeps a specified minimum separation (20mm comes to mind for no reason I can remember) of the mains outlet from the level in the tank.

 

3 hours ago, PeterStarck said:

old type onion cesspit

Good call. Baffles and chambers won't be useful here.

 

3 hours ago, arg said:

how about just running the mains water into your underground tank

I have thought about this and I'm still undecided if it is more efficient to do this or my first plan. It might prove easier though and that's worth a lot. The weather modelling will tell all here.

 

[arg]

1 hour ago, dnb said:

my best guess is that rain capture will be almost seasonal.

 

I thought that, but having spoken to someone local who has had RWH for some years, he claims that in this area (East Anglia) it's not very seasonal.  Met Office data seems to bear this out.

 

Usage on the other hand might be - in my case I'm intending RWH primarily for garden use (so very seasonal), but looking at cost/benefit of adding toilet flushing capability.

[dnb]

Nice data source. Working on some Matlab tonight so I can have a MonteCarlo rainfall model.

[Gone West]

I've been using rainfall data from my local Met. station to compare groundwater levels and rainfall for the last few years.

 

https://www.metoffice.gov.uk/pub/data/weather/uk/climate/stationdata/manstondata.txt

 

 

 

[JOE187]

From my experience keep it simple use  a float switch in both tanks  to controls all levels.

3000litres minium in ground tank and  200l in roof.

Use a floating intake to draw cleanest water, very easy to make using flexible suction pipe, a  plastic  ball float and stainless filter.

Filter on outlets of roof tank that can easily be changed.

Mains water back up to  roof tank with ball valve and manual lever valve.

 Only find back up occasionly needed in dry summer.

 

[dnb]

I promised some mathematical modelling. Not the clothes kind. I'm much better at the former. Here are some charts and things.

 

It turns out that modelling rainfall is almost as difficult as modelling sea clutter as seen by radar. I ended up with a uniform distribution for deciding if it would rain on a particular day, and then a gamma distribution to determine how much rain there would be. It sort of worked but plan B - getting real world daily rain data for my area for the past 89 years - seemed a considerably better option when it appeared!

 

I wrote a quick and dirty Matlab script to read in the rainfall data, then in a loop (couldn't be bothered to vectorise) of 1000 runs select a year at random and run through each day to work out how much rain my roof and a 4000 litre tank could capture less a fraction for efficiency of capture, followed by "using" 150 litres of water that day (no idea if that's the right amount, but it seems sane for the moment). Anything that drained the tank counted as mains usage.

 

This gave me the following charts, averaged over the 1000 runs:

The rainwater tank ranges from zero to full for most days at some point over the 1000 runs, but there are more zeros in the spring than in winter - hence the higher average mains usage. Means are pretty useless statistics though when outliers are involved (i.e. full and empty tanks) so let's look at the median.

This clearly says the expectation is that mains water will not be used (on average) and that 4000 litres is sufficient storage for using 150 litres per day (again on average). So let's look at the 95% "worst case" level, since average is wrong half the time...

In this case, the system only provides 30% of the water we need. But 95% of the time, it will do a lot better. Let's now look at how much better by plotting the mains use percentile for each year. It shows that the system usually provides sufficient water to sustain near zero mains use provided the assumptions on water use of 150 litres per day are sound.

 

All ready for you lot to throw rocks at the charts now ?