Legacy US Property Data Meeting an SI Simulation
Thermophysical property data has a long memory. Correlations fitted decades ago in US-customary units are still embedded in spreadsheets, in-house codes and vendor tables, while the process simulator, the CFD case and the journal template around them all expect SI. Rankine and kelvin share a zero point, so reconciling the two is the cleanest conversion in the whole family — a single division, no offset, nothing to get the sign wrong on.
K = °R × 5/9, equivalently °R ÷ 1.8. Worked example: 671.67°R divided by 1.8 gives exactly 373.15 K, the boiling point of water at one atmosphere.Where Two Unit Sets Have to Agree
Inherited Correlations
Simulator Input Files
Shared Datasets
Dividing Out the Degree-Size Difference
Enter the Rankine value
Take it straight from the table row. Neither scale goes negative, so a minus sign appearing anywhere in this workflow is a signal that something upstream is wrong.
Read the kelvin figure
The result appears as you type. Dividing by 1.8 frequently produces a repeating tail, and up to four decimal places are kept rather than rounded off.
Reverse to feed the legacy code
Swapping the sides sends an SI temperature back up to Rankine, which is what an inherited correlation needs before it will return a sensible property.
Paste it into the case file
Copying returns digits with no unit and no thousands spacing, so a value dropped into an input deck or a data column will parse first time.
Both boxes accept typing, so a round trip can be done in a couple of seconds as a sanity check. Celsius and Fahrenheit are in the same menus when a result has to be shown to someone who thinks in gauge temperatures.
Reference Points for Checking a Unit Set
The fastest way to prove that a spreadsheet, a solver or a colleague's table is interpreting temperature the way you expect is to push a known pair through it. Feed in 491.67 and a correctly configured Rankine-to-kelvin path returns exactly 273.15; anything else means the unit declaration and the data have parted company.
| Rankine | Kelvin | Why it is a useful check |
|---|---|---|
| 400°R | 222.2222 K | Cryogenic auxiliary stream; exercises the repeating tail |
| 491.67°R | 273.15 K | Ice point; the single best round-trip test value |
| 519.67°R | 288.7056 K | US standard-day ambient used in performance corrections |
| 671.67°R | 373.15 K | Atmospheric saturation of water, exact on both scales |
| 1000°R | 555.5556 K | Hot process stream; a round number on one side only |
| 1500°R | 833.3333 K | Fired-heater tube skin territory |
| 2000°R | 1 111.1111 K | Radiant-section gas; shows the thousands spacing |
A Single Clean Division
With a shared zero point there is no constant to add or subtract, which removes the most common source of sign errors between unit systems.
Repeating Tails Preserved
Values such as 555.5556 K keep four places, which is enough to survive being fed back through the reverse conversion without visible drift.
Round-Trip Verification
Typing a value into one box and reading it back through the other is a two-second test of whether a table has already been converted once.
Gauge Scales Alongside
Selecting Celsius or Fahrenheit in a menu turns an absolute state point into the reading a plant instrument would show, without leaving the page.
Unit-System Questions
Why does a round Rankine number give a repeating kelvin value?
Because 5/9 is not a terminating decimal. A tidy 1000°R becomes 555.5556 K once rounded, and the tail continues indefinitely. Keep several places while a value is passing between calculation steps, and round only in the final table.
Do properties quoted per degree convert by the same factor?
The temperature part of them does, but in the opposite sense. A quantity expressed per Rankine degree becomes a larger number per kelvin, since a kelvin is 1.8 times wider. Energy and mass units usually need converting in the same expression, so treat the whole compound unit rather than patching the temperature alone.
How do I confirm a solver is reading my units correctly?
Send a known pair through it. If 491.67 on the input side does not emerge as 273.15, the unit declaration in the case file and the numbers in the data do not match. That one test catches a surprising share of quiet unit errors before they reach a result.
Which scale should a shared dataset be stored in?
Store the original, and record which one it is in the column header rather than in an accompanying note. Converting a dataset before handing it on means the recipient inherits your rounding, and a table converted twice by different people almost never agrees with itself.
Is Rankine still encountered in current work?
Yes, mostly through inheritance. Aerospace, power generation and process industries in the United States carry decades of correlations, acceptance codes and validated spreadsheets built around it, and rewriting all of that is rarely worth the risk. New work tends to be SI while the archive stays where it was.
No comments yet. Be the first to comment!