Determining when ice will form

Hey all,
I would like some help with determining when ice will form on a heat exchanger surface other than trial and error. We run numerous PHE with set of -1.5degC supplying 2degC chilled water with no icing issues, blast tunnels down to 6degC with set of -6degC with no defrost other than off cycle needed. I have built ice banks with set of -12degC and ice started to form at 8degC water temperature.

The reason I ask is we have a loading areas at a Coolstore complex around +10degC to +5degC that require some optimization and lowering of set but I can't take the risk of turning the fan coils into ice blocks. Assume the coils have enough capacity and reach set point for an off cycle defrosting, there is no supplementary defrost installed.


Thanks

Superfridge,
What about other angle of determining calculated heat load,
Calculate actual capacity of units installed measuring td, airflow etc.
With varying SST to match estimated load optimise for maximum SST.
Infiltration, humidity will play a part as well.
Load out docks air coolers can get quite dirty, so also sure they are at optimum.

How ow are you controlling suction pressure, plantroom or back pressure regulator, other?

I would start at 0C. May be at 0C coil will provide required cooling. Than gradually step by step lowering this temperature. Assume that at -4C frost start forming. Initially it will be thin layer and will defrost during off cycle. At -5C this layer will become thicker. Ice block will not be created right away. Finally you will find the temperature what is good for you.

Thanks for the replys
RANGER1, evaporator temperatures are controlled by mechanical back pressure valves. The LP vessel that provides liquid NH3 to these areas has a set of -12degC and pumped liquid pressure of 600kPa and provides liquid to storage rooms as well at various temperatures down to -2degC. The coil maintenance side of things is in hand but good point.

These two load out areas (+10 and +5degC) BPV's are currently set to +2degC.
As you can see the liquid provided to these coils is subcooled by 26K (-12 to +14) Therefore (correct new if I'm wrong) As the liquid enters the coils the liquid must loose 26K of subcooling and then raise it temperature 14K all by sensible heat process until it finally reaches saturation and can start boiling hence the need to lower the coil set. I can't raise the LP pot set due to the -2degC room.
Am I on the right track here?

Your subcooling is 14C (-12 to +2). Regardless of BPR setting first you should raise temperature of the liquid from -12C to +2C. This can be the problem. -12C liquid can create ice block.

Cost money, but what about something like Danfoss ICM in suction
line controlling air off, or pulsing liquid line solenoid to give it time to warm up liquid.
Fine tuning ref valve as well
Is it a new or old system giving you grief?

Some interesting information on subcooled liquid being fed to coils here.
Top feed coils are the worst offenders.
This is not the article I wanted to link to, but I did not find it.

http://www.nhtres.com/Mechanically-P...vaporators.pdf

RANGER1 the client will not spend money. The site is relatively new, 8 years. They have moved the goal posts and I'm trying to make it work but I think it may not have been right from day one.
NH3LVR good link thanks. More to industrial fridge than meets the eye eh.

I'm thinking I will attach a temperature probe to the coil and set it up that way.