I thought this would be helpful for growers trying to scale up.
It is a BOTE design I made tonight showing how to properly set up multi-tank storage with UV-C sterilization and a distribution pump. I previously used this type of design, and it works very well. It’s how I’m building our system.
The RO system is a 14,000 GPD unit from Hyper-Logic, providing ~10,000 GPD RO permeate (purified water) assuming city water temp of ~60’F. The facility will use ~10,000 GPD for plant cultivation, ice water hash, the extraction lab, and the kitchen. All of which are fed from these tanks utilizing a series of automated solenoid control valves and downstream water tanks with level sensors.
I could use the 10,000 GPD RO system from Hyper-Logic, saving about $10,000, if I install a tankless water heater to provide 75-77’F supply water to the RO system. But doing so will cost more than the difference between the 10K GPD and 14K GPD RO systems and entail complex temp control mixing valves downstream of the water heater. Not to mention the increased operating cost from the gas used to heat the water. So overall, it’s better to account for cool water’s effect on RO membrane efficiency than trying to provide ideal feed water temp.
I’m using a dual install of DAB E.SYSBOX pumps as booster pumps for the RO system to provide 25+ GPM at 50-60 PSI at the first carbon filter. I still need to review the dual pump efficiency curve to choose the exact PSI for the required flow.
I’m using a Viqua VP950 UV sterilizer and a custom Franklin SS pump. The pump will provide the required flow (26 or 34 GPM) for three or four daily passes of the total water volume through the UV sterilizer.
From Hyper-Logic, after working with them to design my RO system and their review of my BOTE design I’m sharing here:
Your flow diagram is quite sufficient! Looks like you have (3) post-RO [permeate] water tanks that will equalize when filling the first tank from the RO. The UV Recirc will pull from the 3rd tank > UV > back to 1st tank. Exactly as it should be done. Be sure the UV Sterilizer pump is sized to match the UV Steriliser flowput. Depending on those tank sizes, you might need to go up to the VP950 by Viqua (or similar) so that you get the maximum number of “turnovers” per day of all the water. You want a minimum of 2x per day. I see that the total potential volume in the tanks is 12,000 gallon. @ 2x … 24,000gallon div by1440/min = ~17gpm AT MINIMUM. The VP600 under it OPTIMAL UV contact time throughputs @ 18gpm. The VP950 OPTIMAL is 26gpm thus meaning you could get ~3 turnovers per 24-hours. That would be my vote! I attached the VP600/950 manual for reference of my data. You will need a larger pump than the ‘March’ I suggested in one of the last emails. It’s possible that March pumps make a similar unit with larger GPM throughput to match the throughput potential of the VP950; 26 to 40gpm capable (reference page 18 of the Viqua manual attached). We formerly used some much more robust, all stainless steel impeller Goulds pumps set up with a robust controls system (low level shutdown, power control box, etc) to recirculate through the VP950 or larger UV Sterilizers.
My response to Hyper-Logic:
Excellent, thanks for such a detailed response!
To avoid pump interference, the UV recirculation pump pipe bulkhead will be installed 90 degrees (and high off the floor) from the distribution pump pipe bulkhead installation point.
Good point about dual DAB E.SYBOX for the larger RO system. It’s similar to a set up I used previously.
Great info about the UV, thanks! If I can’t find a large enough SS March pump, I’ll probably get a custom Franklin pump. I’m using a Franklin 3-stage vertical pump (stainless steel wetted surfaces) for distribution (50 GPM @ 50 PSI) because they have a solid 12-week lead time, paired with a Yaskawa VFD. Currently, custom-built Gould pumps have around 16-18 weeks lead time, and Berkley is approximately 24-28 weeks!
It’s important to note some things not shown in my BOTE design drawing. Which effect the functionality and effectiveness of air vents and the measurement accuracy of pressure transducers, flow meters, in-line pH and EC sensors, ion selective electrodes (ISE), etc. The design of mainline, sub-mainline, lateral manifolds, and laterals, along with combo and automatic air vent placement, is of utmost importance to prevent water hammers (particularly on long pipe runs).
- The end of the RO permeate pipe in the first RO storage tank is below the level of the lower float sensor. This prevents surface splashing when the RO permeate flow fills the tank, which can cause erroneous readings from the float sensors (making them register incorrect water levels).
- The distance between the automatic air vent before the pump and the upstream ball valve is ten times the pipe’s ⌀ (inner diameter).
- The distance between the irrigation pump’s suction port and isolation ball valve (between the tank and the pump) is 10-15 times the pipe’s ⌀.
- The distance between the dynamic air vent and the upstream ball valve is five to ten times the pipe’s ⌀.
- The distance between the combo air vent and the check valve is five to ten times the pipe’s ⌀.
- The pressure transducer for the VFD I’m using to control the irrigation pump’s speed is installed after the check valve, at a distance of ten times the pipe’s ⌀ downstream from the combination air vent.
- There is a straight pipe after the transducer the length of at least five times the pipe’s ⌀ before any turbulence point (elbow, valve, tee, etc.).
- The flow bypass (aka pressure relief) is required to cool and lubricate the pump because it runs 24-7, often operating against closed downstream valves. Without the flow bypass, the pump would overheat and cavitate when operating against closed vales, reducing pump lifespan. Multi-stage commercial pumps like this require a minimum liquid flow while in operation.
- I install a dynamic flow restrictor into the flow bypass that’s ~20% to 30% greater than the pump’s minimum liquid flow GPM. That ensures I don’t lose too much flow or pressure to the bypass. It’s essential to oversize the restrictor’s flow rate by at least 18% greater than the pump’s minimum liquid flow because most dynamic flow restrictors have +/-15% accuracy.
- The air vent at the elbow of the flow bypass should be above the water line. Otherwise, use an automatic air vent. However, an automatic air vent is suboptimal because you don’t want to install it at a turbulence point (in this case, an elbow).
