1-805-484-2992
Quotes - Engineering - SalesSize Is a Calculation, Not a Catalog Lookup
The question arrives as a model number question: which unit do I need? It is really a surface-area question, and surface area comes from three inputs that you already have or can get in an afternoon. The duty is how much heat has to move per hour. The log mean temperature difference is how hard the utility pushes it. The overall coefficient is how easily the wall and the two fluids let it through. Area equals duty divided by the product of the other two.
Below is the milk heater worked all the way through. Substitute your own flow, temperatures and product and the same steps give your own answer, and the quote form is laid out to take exactly these inputs.
The inputs you need on hand
- Product mass flow, in lb/h or gpm with a density
- Product inlet and required outlet temperature
- Product specific heat, or the product name so we can look it up
- Utility type and its supply temperature, and return temperature if fixed
- Product viscosity at the inlet and outlet temperatures
- Any particulates, fibers or pulp and their largest dimension
Log mean difference, milk heaterA 25°F hot end and a 135°F cold end work out to 65°F, and that number, not a catalog, sets how many square feet the duty takes.
The Milk Heater, Worked
Take 10,000 lb/h of whole milk from 40 F to 165 F using a 190 F hot water loop that returns at 175 F. Specific heat of milk is about 0.93 Btu/lb-F, so the duty is 10,000 times 0.93 times 125, which is 1,162,500 Btu/h, call it 1.16 MMBtu/h. In counterflow the hot end difference is 190 minus 165, or 25 F, and the cold end is 175 minus 40, or 135 F. The log mean of 25 and 135 is 65 F.
Now the coefficient. A sanitary gasketed plate on milk with a dairy fouling allowance is about 900 Btu/h-ft2-F in service. Area is 1,162,500 divided by 900 times 65, or just under 20 square feet, which is a modest plate pack. The same duty in a sanitary shell-and-tube at 250 needs about 72 square feet, a 6 inch shell with a bundle several feet long. The numbers below are what change that answer.
- Drop the hot water to 175 F and the LMTD falls to about 48 F: plate area rises to 27 sq ft
- Add 90% regeneration and the heater only lifts milk from 152 F to 165 F: its duty drops to about 0.12 MMBtu/h
- Switch to cream at 0.85 specific heat: duty falls 9%, but viscosity roughly triples and the plate coefficient drops with it
- Apply a raw milk fouling factor of 0.001 h-ft2-F/Btu: a 1200 clean coefficient becomes about 550 fouled
- Halve the milk flow with the same utility: LMTD barely moves, but the plate coefficient falls as channel velocity drops
- Use 150 psig plant steam at 366 F instead of hot water: LMTD jumps, area falls, wall temperature burns milk on
Area equals duty over LMTD times a fouled coefficient, never clean.
- Milk 0.93 Btu/lb-F, cream 0.85, juice about 0.90
- Plates 800 to 1200 Btu/h-ft2-F, shell-and-tube 150 to 350
- Tube side 4 to 8 ft/s keeps the wall clean
Duty, Temperature Difference and Coefficient, One at a Time
Duty is the term nobody gets wrong and everybody forgets to state. It is mass flow times specific heat times the temperature change on the product side, and it is the same number whether the exchanger is plates or tubes. Give flow as lb/h if you can; if you have gpm, give the product density or the product name and we convert it. Specific heat for milk is 0.93, skim slightly higher, cream around 0.85, single-strength juice about 0.90, and 40% sugar syrup near 0.75. Water is 1.0, which is why a WFI cooler at the same gpm carries a slightly larger duty than a milk cooler.
The log mean temperature difference is where the design choices live. It is not the average of inlet and outlet differences; it is the log mean, and it drops fast as the two ends get close. On the milk heater the hot end at 25 F and the cold end at 135 F give 65 F. If the plant only has 175 F water the hot end becomes 10 F, the cold end 135 F, and the LMTD is 48 F: the same duty now needs a third more surface. This is the reason a hot water set is specified at 185 to 195 F for dairy heating: warm enough for a workable LMTD, cool enough not to burn on.
Counterflow matters. A gasketed plate unit is true counterflow, which is what allows a 2 to 3 F approach in a regenerator, and what makes a temperature cross possible, where the cold product leaves hotter than the hot product does. A one-shell-pass, two-tube-pass shell-and-tube is part parallel flow and carries a correction factor below 1.0, which is one more reason its economic approach is 10 F or more. A tube-in-tube in series is true counterflow too, which is why it can run a tight approach on a viscous product that could never see the inside of a plate pack.
The overall coefficient is where sanitary duty differs from utility duty. On water-like product a gasketed plate runs 800 to 1200 Btu/h-ft2-F, a sanitary shell-and-tube 150 to 350, and a tube-in-tube on viscous or particulate product 100 to 250. Those are service values with fouling included. The clean values are higher, and quoting the clean value is the classic way to buy an exchanger that is fine on Monday and short on Friday. Raw milk deposits protein and calcium phosphate on any wall above about 150 F, and a fouling factor of 0.0005 to 0.001 h-ft2-F/Btu is the normal allowance; on a plate whose clean coefficient is 1200 the higher figure alone takes the service value to about 550.
Fouling allowance also sets the run length. A pasteurizer sized with a generous allowance holds temperature for a full shift and is cleaned once; one sized on the clean coefficient starts short of set point after a few hours and the utility valve opens to compensate, raising wall temperature and fouling faster. On viscous products the deposits are worse and the coefficient lower to begin with, so the allowance is a larger fraction of the total, which is part of why tube-in-tube units on sauce look large for their duty.
Velocity is the lever that keeps the coefficient up. On the tube side of a shell-and-tube, 4 to 8 ft/s on thin product keeps the film coefficient high and the wall clean; below 3 ft/s the coefficient drops and deposits build. On plates the corrugation does the work, but channel velocity still falls if the flow is well below design, so a unit that will run at half flow on the night shift should be told that. The cost of velocity is pressure drop: 10 to 15 psi across a plate pack and 5 to 10 psi on a tube side are normal, and both are inputs to the size rather than results of it.
Put the three terms together and the area is fixed, but the shape is not. Twenty square feet of plate can be a few large plates or many small ones; 72 square feet of tube can be a short 8 inch shell or a long 6 inch one. The choice is made on footprint, pressure drop, connection size and how the unit will be cleaned, which is where the construction page of this section picks up. Send the duty and constraints as they are and we will size it both ways where it is close. Call and talk it through with an engineer: 1-805-484-2992
Quote the clean coefficient and you buy a unit that is fine on Monday and short on Friday.
Request a QuoteCommon FAQs
For clean, water-like product with hot water or chilled water, roughly one square foot of gasketed plate per 60,000 Btu/h at a 65 F LMTD, and three to four times that in a sanitary shell-and-tube. It is only good for a first look; fouling, viscosity and approach change it quickly.
Because the driving force falls off along the exchanger as the two streams converge, and the log mean weights that fall-off correctly. For a 25 F and 135 F pair the arithmetic mean is 80 F and the log mean is 65 F; using the arithmetic mean would undersize the unit by a fifth.
Raw milk and cream 0.0005 to 0.001 h-ft2-F/Btu, pasteurized and skim toward the lower end, and higher for products heated above 170 F where protein denaturation accelerates. State the run length between CIP cycles and we will set the allowance to hold temperature through it.
It raises the LMTD and reduces the surface, but it also raises the wall temperature. On dairy, egg and protein products the wall should stay within 15 to 20 F of the product to avoid burn-on, which caps the utility temperature regardless of what the size calculation would prefer.
Related
Sizing & Selection
Flow, Temperatures & Regeneration
State the flow so it cannot be misread, choose the approach on purpose, and let the product heat itself.
+ Learn More
Sizing & Selection
Viscosity & Particulates
The two product properties that move a duty from plates to a wide gap to a tube.
+ Learn More
Sizing & Selection
Size Your Exchanger
Duty, temperature difference and the U-value your product will actually give you.
+ Learn More
