Gravitational Flow - Pond return
 

Hi people, new here, and fluid motion is not my forte, so I need your assistance.

I am installing a koi pond, around 200 gallons. The water level will be about 10 to 12 inches above G.L.

The filter box has a spill overflow, and I will have to hard-pipe the return flow underground. There will need to be perhaps 5 or 6 90 degree bends in the return flow pipework, for which I plan to use 50mm diameter pipework (waste pipe).

I will be filtering at around 1,000 lph, and realise that the spill overflow of the filter box must be higher than pond level to create the head to achieve that return flow-rate. I plan to make that distance much larger, do not want it marginal.

My question is "How much higher?". I do not have the knowledge or skills to work this out. If anyone can help I would appreciate it.



The attached sketch gives an idea of the layout.

Have I really got to do this by trial and error ? Surely there is a theoretical solution to this problem.

I do not want to elevate the filter box too much, as it would mean a pump from the bottom of the pond with a larger delivery head than might be necessary, wasting energy.

Ideally I'd like to maintain say 6" between the filter spill outlet and the top level in the return flow pipework ...

You have two sources of pressure decrement: The pipe and the filter. Only if you assume that most of the decrement comes form the pipe you may calculated the pressure difference needed by the Hagen Poiseuille equation:

V` = pi · r^4 · delta_p / ( 8 · eta · l)

V' ist the flow rate
r is the pipe radius
delta_p ist the pressure difference
eta is the viscosity
l ist the length of the pipe.

Thank you for the reply. It will only be the pipe, as the outlet of the filter is after the filter media.

I entered the formula into an excel spreadsheet, and then used Goal Seek to calculate delta_p for a flow of 16.6667 litres per minute. I got an unreasonable delta_p figure which I could not have achieved.

So I went looking on the net and found this handy calculator for the equation. Particularly good because you can change units to match your application, and you can get the result in ft H2O directly.

So for a 50mm diameter pipe, length 8m, (actually only about 6, but compensating for the bends), and a flowrate of 16.667 l/min (1,000 l/hr) it calculates the head required as 0.00630063 ft H2O, or approximately 0.075 inches. I used 1.3 cP as the dynamic viscosity of water at 10 deg.C

This is particularly great news for me, as I had planned to put the filter spill outlet into vertical pipework about 12" above the pond level, and therefore it will not "back-up" the pipework.

Does this sound reasonable ?

I used the values
r=0.025 [m]
delta_p = 0.025 [N/m²]
eta=1.3e-4 [P·s]
l = 8 [m]

and get a stream of around 3e-4 [m³/s], which is 17l/s.

Are you sure you don't mean 17 l/min ?

17 l/s is 1020 l/min, or 61200 l/hr - that's an awfully large flow-rate..... We used to clean brewery pipework at 5 l/s, where turbulence is at its peak, and to do that we had to use pumps to achieve that flow...

Anyway, I've used the online calculator to go down to 40mm pipework, because I can incorporate a 40mm vane flowmeter I already have in the outlet stream. This ranges from 5 to 150 litres/minute, so will be ideal in monitoring the actual flow-rate with an Arduino project I am designing.

This will also monitor water pH, temperature, UVC lamp detection, and a level sensor in the effluent standpipe in case the return pipework incurs any obstruction. The Arduino app will cut the pump off if the return flowrate falls below a preset value, indicating a problem with the pump or filter media.

All the data will be sent via a cloud server to a mobile phone app called RemoteXY. Alarming will be done with a GSM/GPRS module sending text messages on 2G/3G.

The 40mm pipework will produce a 0.0153824 ft H2O delta_p, or just under 0.2 inches.