Duty and temperatures tell you how much surface. Viscosity and particulates tell you what shape that surface can take. A gasketed plate pack has channels a few millimeters wide, and every gain in coefficient it offers comes from pushing product through those channels fast enough to stay turbulent. Thicken the product or put pieces in it and the channels fight back, first with pressure drop and then with blockage.
The thresholds below are working rules, not laws, and a product near one of them is worth a conversation. But they sort the large majority of sanitary duties into the right construction on the first pass.
Yogurt base at the cold endA product entering at 300 cP and leaving at 2,000 cP is sized on the thick end, because the last channels or last module set the pressure drop for the whole unit.
The viscosity that matters is the one the product has where it is thickest inside the unit, and that is the cold end. Yogurt base cooled from 110 F to 40 F might be 300 cP going in and 2,000 cP coming out; the last few channels or the last module see the 2,000 and set the pressure drop for the whole unit. Sugar syrup, fruit concentrate and chocolate behave the same way in reverse on a heater: thick at the inlet, thin at the outlet.
Give the viscosity at both temperatures if you have it, and the temperature at which it was measured if you only have one number. A single figure at room temperature is almost always the wrong one for sizing.
Viscosity attacks the coefficient twice. It thickens the laminar film at the wall, and it lowers the Reynolds number so the flow stops being turbulent at all. A gasketed plate that gives 1,000 Btu/h-ft2-F on milk will give 300 on a 200 cP syrup at the same flow, and 150 on a 500 cP one, and to hold even that the pressure drop rises several times over. Somewhere around 500 cP the pump head needed to keep a plate channel working exceeds what a sanitary centrifugal pump delivers and a positive displacement pump is needed, and at that point the plate has lost its main advantage. Where the flow curve of a product is not known, a sample can be tested; see rheology testing on our tube-in-tube site.
A wide-gap plate buys some room. Its channels are roughly twice the width of a standard plate, its coefficient is lower, and it passes pulp, seeds and small fibers that would bridge a standard channel. It is the right answer for citrus juice with pulp, fruit purees without pieces, and moderately viscous products in the low hundreds of cP where the plate's counterflow and close approach are still wanted. It is not the answer for diced product or for anything above about 500 cP.
Tube-in-tube handles what plates cannot because the inner tube is a single open passage, typically 1 to 3 inches, with nothing to bridge and nothing to shear the product against. The coefficient on viscous product is 100 to 250 Btu/h-ft2-F, sometimes less on a paste, so the surface is large for the duty and the unit is long. That length is the price of passing the product intact and cleaning it with a CIP flush that goes end to end with no dead spots. A corrugated inner tube adds turbulence at the wall and can roughly double the coefficient on laminar products that tolerate the mixing.
Particulates are sized on their largest dimension, not their average. A sauce with quarter-inch diced peppers and the occasional half-inch piece needs an inner tube that passes the half-inch piece with room to spare, which in practice means an inner tube of 1.5 inches or more. Multi-tube modules, with several smaller tubes in one shell, increase surface per foot but each small tube must still pass the largest piece, so the piece size sets the minimum tube diameter and the surface follows from that.
Pressure drop is a budget, and on viscous product it is the tightest budget on the sheet. Across a plate pack 10 to 15 psi is normal on thin product; on a cooler that finishes at 2,000 cP it may not be achievable at any plate count. On a tube side 5 to 10 psi is the target, and on a viscous product it is reached by shortening the series path and adding parallel modules rather than by opening the tube up, because a larger tube at the same flow drops velocity and coefficient together. Give us the discharge pressure of the pump feeding the exchanger and the pressure the downstream equipment needs, and the unit is sized inside that window.
Fouling and viscosity go together. Thick product moves slowly at the wall, the wall runs closer to the utility temperature, and burn-on starts. On a steam-heated tube-in-tube the fix is a lower utility temperature or a hot water loop, and a longer unit; on a plate the fix is usually to leave the plate behind. Products that gel, set or crystallize as they cool, starches, pectin, some fats, need the outlet temperature held above the set point or a scraped surface unit, and are worth a call before anything is sized.
The common mistake on viscous product is to size the exchanger for the liquid and forget the pieces, or to size it for the room-temperature viscosity and be surprised at the cold end. The second common mistake is to reject tube-in-tube because it looks large, then spend the difference on plate gaskets and a positive displacement pump that was not in the budget. State the viscosity at both ends, the largest piece, and the pump you have, and the construction follows.
Call and talk it through with an engineer: 1-805-484-2992
The apparent viscosity at the shear rate the product will see in the exchanger, at both the inlet and the outlet temperature; failing that, the whole flow curve. Yogurt base, purees and sauces can read several-fold differently on a viscometer depending on spindle speed, so a single number taken at an unknown speed and at room temperature is nearly useless. A plate channel shears product hard and a 2 inch tube hardly at all, which is why the same yogurt can pump through plates on a trial and then stall a tube-in-tube cooler sized on the plate reading. Record the instrument, spindle, speed and temperature with every reading, or send a sample for a flow curve.
Far less than a pump or a plate does, provided the velocity and the return bends are chosen for the piece. The inner tube is an open passage of 1 to 3 inches with nothing to shear against, so fruit cubes, diced vegetables and curd arrive whole at moderate velocity. Damage usually traces to the feed pump or to short-radius bends where pieces strike the wall. Tell us how intact the pieces must be and how that is judged, and the velocity and bend radius are set to it.
No; the liquid carrier reaches temperature well before the center of a large piece does, so a unit sized on liquid outlet alone can leave pieces underprocessed. Heat has to conduct from the surface of a half-inch cube to its center, and that takes time the exchanger cannot supply on its own, so the hold tube downstream is sized on the largest piece and the required center temperature, not on the liquid. Validation then means a thermocouple in a simulated particle, not just the outlet probe. Send the largest piece dimension and the center temperature the process requires, and the exchanger and hold tube are sized together.
Usually on the heating duty, rarely on the cooling duty, because the cold end is where the product is thickest and the last channels set the pressure drop for the whole pack. A product that enters a cooler at 300 cP and leaves at 2,000 cP, like the yogurt base on this page, is well past the few hundred cP a wide-gap plate handles, even though the same product at 110 F would heat in one without trouble. The plant symptom is a positive displacement pump running at its relief setting while the outlet temperature creeps up, because flow has fallen. The usual split is a plate for heating and a tube-in-tube for the final cooling, each sized on its own outlet viscosity, with the changeover temperature chosen where the plate pressure drop is still inside its 10 to 15 psi budget.
Because every one of the smaller inner tubes must still pass the largest piece, so the piece sets the minimum tube diameter and the surface follows from that. A sauce with occasional half-inch pieces needs an inner tube of 1.5 inches or more; splitting the flow across several 1 inch tubes to gain surface per foot would plug the first module on the first shift. Where the surface still has to grow, add modules in parallel rather than shrink the tubes.
No; it roughly doubles the coefficient on a laminar product that tolerates mixing, and it is the wrong tube for products damaged by shear or carrying pieces that can catch. Smooth syrups, concentrates, honey and liquid egg take it well, and the shorter unit pays for itself. Fruit preparations with soft pieces, cultured products whose texture must survive, and anything with fibers that can bridge a corrugation are better in a smooth tube, accepting the extra length. The pressure drop also rises with the corrugation, so a product already near the 5 to 10 psi tube-side budget may lose more in pump head than it gains in surface. Where a product sits near the 500 cP line and its texture is not critical, pricing smooth and corrugated side by side is cheap and usually settles it. Say whether texture, piece integrity or pressure drop is the constraint that matters most.
Duty, log mean temperature difference and a realistic U-value, worked through on a milk heater.
+ Learn More
State the flow so it cannot be misread, choose the approach on purpose, and let the product heat itself.
+ Learn More
Duty, temperature difference and the U-value your product will actually give you.
+ Learn More