The loop as a system
A hot WFI system is a tank, a pump, a loop and a heater. The still or membrane unit fills the tank; the pump circulates the loop past every use point and back at 80 C or above; the heater replaces what the tank walls, the pipe, the use-point coolers and the sub-loops take out. What is hard is the accounting: the heater is sized on losses nobody has measured, and the sub-loop coolers on draws that happen ten minutes a shift.
This page covers that accounting, the cold sub-loop, the condensate leg and the return temperature control.
Once the system is hot the heater sees only what the loop loses. A 10,000 gallon insulated tank at 82 C in a 22 C room loses 20,000 to 40,000 Btu/h; 800 ft of 2 in insulated pipe loses about the same; each use-point draw takes out 80 C water and returns nothing. On a mid-size system the steady loss is 60,000 to 150,000 Btu/h, and a 200,000 Btu/h heater holds it with margin. The cold start from ambient is done by the still filling the tank hot, not by the loop heater. Sizing the heater for a two-hour cold start it will never do gives a unit that cycles its steam valve all day and hunts the return.
The heater is a vertical double tubesheet shell-and-tube with the loop on the tube side and steam at 15 to 30 psig on the shell, a shell temperature of 121 to 134 C: enough driving force for a compact unit and low enough that a stuck valve cannot flash the loop.
A formulation suite that draws WFI at 20 to 25 C for an hour at a time is badly served by a use-point cooler that starts and stops on each draw. A cold sub-loop is the alternative: a branch off the hot loop through a sub-loop cooler, circulated cold past several use points, and returned to the tank. The cooler sees the full sub-loop flow while the suite runs, 82 C to 22 C on chilled water, and is sized on that flow. When the suite finishes, the sub-loop is drained and steamed, or returned to the hot loop and reheated by the main heater, which is the other reason the heater carries margin.
Steam condensate from the heater shell is plant condensate and goes back to the boiler through a trap at the shell low point. It never runs to a floor drain near the WFI tank vent or a use point, and it never shares a header with a sanitary drain. The trap sits below the shell so the shell drains dry when steam is off, because a shell holding 100 C condensate against 80 C tubes is a thermal shock at the next start. An air gap between the condensate line and any drain is the rule; a direct connection is a cross-connection finding.
The control loop is simple: a transmitter on the return to tank, a modulating steam valve on the heater, a setpoint of 82 to 85 C and a low alarm at 78 C. The heater outlet is not the control point, because the return is the coldest place in the loop and the SOP names the return. A well-insulated tank with a loop heater needs no jacket; what it needs is a heated vent filter and a spray ball fed from the loop so the headspace and walls above the water line are wetted at 80 C, because that is where a hot system grows organisms if it grows them anywhere.
Send HeatX the tank size and insulation, loop length and pipe size, the use points and the steam pressure, and the heater and coolers are sized as a set.
Call and talk it through with an engineer: 1-805-484-2992
This table describes a heater and sub-loop coolers sized to keep a storage loop hot and alarmed. Check it against your tank, return temperature and steam supply, then send those figures with your quote request.
| Condition | Detail |
| Product | WFI in a recirculating loop; 1,000 to 20,000 gallons stored |
| Flow range | 30 to 400 gpm loop; 5 to 60 gpm on a cold sub-loop |
| Temperatures in / out | Heater 78 to 82 C in, 82 to 85 C out; sub-loop cooler 82 C to 20 to 25 C |
| Hold | 80 C minimum at the return, continuously, alarmed at 78 C |
| Utility | Plant steam 15 to 30 psig; chilled water 7 C on sub-loop coolers |
| Approach | Heater 15 to 25 C; cooler 3 to 5 C |
| Construction | Double tubesheet straight tube, vertical, steam on the shell |
| Finish / class | 316L, 15 to 20 Ra electropolished; ASME BPE; drains to one low point |
| Condensate | Steam trap at the shell low point, to plant return through an air gap, never to a drain near WFI |
| CIP / SIP | Derouging with the loop; a cold sub-loop is steamed at 121 C after each use |
Measure the loss with a cool-down test and size on that. With the loop circulating and the use points in normal service, close the steam, log the return temperature as it falls, and stop well above the 78 C alarm. A 12,000 gallon system holds about 100,000 lb of water, so a fall of 1 F per hour is about 100,000 Btu/h of loss, plus what the tank and pipe steel give up, and the 3.6 F from 82 to 80 C takes over three hours at that rate. That one afternoon's data is more reliable than insulation tables, and it often shows the existing heater at two or three times what the loop needs, which is why it's steam valve hunts.
Two in parallel is common on loops that cannot be allowed to cool for a gap inspection, a derouging or a repair. Two full-size units give true redundancy; two half-size units give better turndown, because one can be valved out when the loop is quiet and the steam valve on the other works in the middle of its range. The idle unit must not become a dead leg, so the isolation uses zero-static valves at the loop and the unit is drained and left dry or kept hot in a short recirculation, and the SOP says which.
It upsets the return RTD before it upsets the tank, and that is still a logged excursion. Three hundred gallons at 22 C mixed into 10,000 gallons at 82 C only drops the tank about 2 C, but the cold slug passes the return transmitter first and pulls it well under the 78 C alarm. Reheating that 300 gallons from 22 to 82 C is about 270,000 Btu, and with 50,000 Btu/h of spare heater capacity that takes more than five hours. The fix is to drain the sub-loop, or to reheat it through its own branch on the main heater before the return valve opens, and schedule it after the suite finishes, since even the 140,000 Btu/h a quiet loop leaves spare on a 200,000 Btu/h heater takes nearly two hours.
Dry plant steam with the air removed; it never touches the water, so compendial purity is not required. A separator and drip trap ahead of the control valve take out the condensate, and an air vent or a thermostatic trap at the shell high point removes the air that blankets the tubes on every start. Wet steam or air shows as the valve driving further open for the same return temperature. A dedicated pressure-reducing station holds the shell at 15 to 30 psig whatever the boiler header does, which keeps the shell temperature at 121 to 134 C and the tube wall where the design assumed it.
Route the whole loop return through the spray ball and size the pump for its pressure drop. That way the headspace and the walls above the water line are wetted at 80 C whenever the loop runs, instead of on a timer that leaves them dry most of the day. The loop still has to hold its velocity: a 2 inch return needs about 26 gpm to keep the 3 ft/s minimum, and every open use point and sub-loop take-off subtracts from what reaches the spray ball.
The exchanger a WFI system is validated around: 80 C, double tubesheet, 15 to 20 Ra, and nothing added.
+ Learn More
Ambient loops run cheap and sanitize on a schedule; the exchanger has to do both jobs on the same tubes.
+ Learn More
UV lamps and pumps warm an ozonated loop all day; the cooler removes that heat and gives ozone nothing to eat.
+ Learn More