Once three independent paths close to one decimal, which of the remaining books is allowed to re-derive the coefficient? Three unit confusions are named at the head of the cross-check section as the errors it exists to exclude, and they are named because they are the cheap fatal ones in liquid cooling [TARGET]:
- specific heat per kilogram read as specific heat per litre;
- litres a minute read as cubic metres an hour, a factor of 60;
- gallons a minute read as litres a minute, a factor of 3.785.
The section then recomputes the flow basis three ways and declares in advance that any one of them failing to close voids the whole chapter. All three closed.
The three paths do not share an intermediate
The first goes through the volumetric heat capacity that the flow table itself uses. The second bypasses that constant entirely: mass flow of 300.6 kilograms a second from specific heat and temperature difference, divided by density, giving 1,069.3 cubic metres an hour against the table's 1,069, a deviation below 0.1 per cent traceable to rounding [MODELED]. The third leaves the metric system altogether and returns 46.8 gallons a minute for one cabinet against 46.7 converted from the metric result [MODELED]. Because the second path never touches the composite constant, it can disprove that constant rather than merely restate it, and that property is what makes the set a check instead of three paraphrases.
The chapter's own later self-check restates the same closures and then writes the limit: the recomputation proves internal consistency of formula and substitution. It does not prove that the fluid properties, the supplier's allowed temperature difference, the effective nameplate capacity or the 385 kW pump-heat assumption actually hold.
Closure Moves Residual Risk Onto Four Registered Blockers
It found nothing, and finding nothing is the output. Its value is that it converts an unbounded worry into a bounded one: once it closes, the residual uncertainty in the chapter sits entirely on four registered blockers, the liquid fraction, the temperature difference, the cabinet density and the fluid the nameplate was rated on, and not on the arithmetic. That relocation is what allows the flow basis to be released for the enquiry package and for pipe sizing while all four blockers stay open. A fourth comparison supports it from outside, against a flow intensity widely quoted for water at ten kelvin, and lands within the ratio the chapter derives elsewhere for the working fluid.
The unit-intensity comparison is the one that would catch a silent factor error, and it is worth stating what it can reach. At the recommended temperature difference the chapter's figure is 1.512 litres a minute per kilowatt against 1.447 for water on the same basis, and the ratio between them reproduces the fluid correction the chapter derives from its property table by a different route [MODELED]. A sixty-fold or a three-point-eight-fold slip would put the intensity nowhere near either figure, which is why the comparison is carried even though it proves nothing about this plant.
So the decision the section enables is one about sequence. Main pipe sizes may be drawn before the cabinet count is frozen, because the main sizes follow total heat while the terminal quantities follow the count, and the chapter separates those two explicitly. The cross-check is what makes the separation safe to act on, and it is the reason a pipe schedule can be issued against inputs that are still open. The same chapter recommends locking the unit-check book before every other calculation book, so later books cannot each re-derive the coefficient.
A separate parametric engine expands the same formula across a grid and reports liquid duty without pump heat. Its mass flow and volume at the same temperature difference are not this pipe-size flow. Quoting one as the other is a source error.
Closure Does Not Prove the Fluid Properties Hold
It is one fluid property table that all three paths draw their properties from, and that table is engineering-correlation data with the supplier's own property sheet not obtained. Three paths agreeing on a wrong density would close exactly as cleanly, and the chapter marks the property basis as an assumption to be replaced before construction drawings are issued.
The method also does not survive being scaled up, in the one place it was tried. A separate parametric engine expands the same formula across every distribution unit and every scenario, runs an audit over each, and then withholds the word: it records that substituting a formula back into itself is no longer to be called an independent engineering check. Two hundred and eighty-eight of those audits pass. Both golden scripts pass. Eleven of eleven negative tests pass. The pack still withholds engineering acceptance on every row.
That withdrawal is the sharper of the two limits, because it names the property the manual section relies on and the scaled one lacks. Three paths that share no intermediate can falsify a constant; two hundred and eighty-eight audits that all re-enter one formula cannot, however many of them pass. Nothing in the record says which of the two the four open blockers will eventually be closed against.
An independent check of a different supplier packet closed eight hundred and sixty-four ring rows to floating-point noise with the Python solver missing. That is snapshot arithmetic on a supplied network, not this coefficient, and not a solved plant.
Limits and open items
Confirmed: any one of the three paths failing to close voids the hydraulic chapter; the residual uncertainty after closure sits on the four registered blockers, not on the arithmetic. [FACT]
Definitional factors: 60 between litres a minute and cubic metres an hour; 3.785 between gallons a minute and litres a minute. [TARGET]
Closed recomputations: 300.6 kilograms a second; 1,069.3 against 1,069 cubic metres an hour; 46.8 gallons a minute against 46.7; 1.512 litres a minute per kilowatt against 1.447 for water. [MODELED]
Open: the purchased-fluid property sheet; which book later calculations are forbidden from re-deriving the coefficient in; which of the four blockers the next arithmetic waits on. [HOLD]
Nothing here has been reviewed or sealed by an Engineer of Record. Nothing describes how any installed system behaved.
Name the Book Forbidden From Re-Deriving the Coefficient
The useful test is not whether 1,069 cubic metres an hour closes. Name the calculation book that is forbidden from re-deriving 1.512 litres a minute per kilowatt. If that name is missing, the three-path close has not yet done the work it was written to do.
© 2026 K&K Data Service Inc. All rights reserved. Reproduction or republication is permitted only with clear attribution to K&K Data Service Inc. and a working hyperlink to the canonical URL of this article. Excerpts must preserve the technical context, maturity labels, assumptions, and limitations. No excerpt may imply project approval, field validation, certification, or endorsement that the original article does not state.
Source: K&K Data Service Inc., “Three Hydraulic Flow Paths Closed; Residual Risk Left the Arithmetic,” https://www.kkdatasvc.com/lab/hydraulic-calculation-detail/three-paths-closed-to-one-decimal-and-the-risk-moved-off-the-arithmetic/.
Request the Public Calculation Note. If your project record shows a different result, or the same failure domain under another name, we want to see it. Email inquiry@kkdatasvc.com or use the contact page.