File 07 still caps the project at 177 litres a minute. File 16 of the same packet states that approval of actual cabinet suitability is not required. Which of those two files is the interface? Raising the secondary operating flow from 53.1 to 67.3 cubic metres an hour is a 26.742 per cent increase, and the resistance that goes with it was neither measured nor quoted [MODELED]. It was obtained by squaring away from a single published flow point: five branches at 224.333 litres a minute each, against a published point of 177, giving 203.788 kPa. Working the ratio backwards puts that published anchor near 126.9 kPa [MODELED]. One point and one exponent are the entire basis for the new figure.
A Square-Law Extrapolation Is Not a Manufacturer Limit
A square law is the right shape for fully developed turbulent flow in a fixed geometry, and it is not a manufacturer guarantee. The facility hydraulic basis writes the same family as an assumption: pressure drop scales with volume flow to the power 1.9, inside an engineering band of 1.8 to 2.0. The supplier packet used two. The review states the narrow version of this twice: exceeding a published table point does not by itself mean exceeding an allowable limit, and it also does not demonstrate that the cabinet interface accepts the higher flow. Those two sentences are not a hedge. They are the boundary between a hydraulic extrapolation, which is the designer's to make and to own, and an equipment limit, which is not.
A facility hydraulic point of 176.9 L/min at one local rack duty sits next to the vendor's 177 L/min cap. They are two stories, not one constraint. The raised 224.333 L/min is 67.3 cubic metres an hour divided by five, not the facility formula.
Raised Flow Fails the Package's Own Velocity Screen
At 67.3 cubic metres an hour through a connection pipe of 102.26 millimetres bore the velocity is 2.276 metres a second, above the 2 metres a second screening value printed in the package's own pipe-sizing table [MODELED]. The review is exact about what that is and is not:
- it is not a breach of any mandatory standard, and no pipe size is directed to change;
- it is a claim the package can no longer make, because it cannot state that its screening criterion is satisfied while operating above it.
The review asks instead which pipe runs the criterion was written for, what the basis for the deviation is, and how a fixed high flow at every load compares with a modulated one once a real pump curve exists. None of the three is answerable inside the package as delivered. That 2 metres a second screen is not the facility isolation-velocity cap, and the two must not be merged.
The extrapolation is not the only one in the package built on a single anchor, which is what makes it a pattern rather than a lapse. The exchanger's constant overall coefficient is the second case, and the review's own account of where that constant comes from is cited here rather than worked again; the dry cooler behind it is selected at one ambient point and asked to hold at a hotter one. In each case a single supplied point is being asked to describe a curve, and in each case the curve is the thing the supplier would have to stand behind.
Two Files Disagree, and No Interface Matrix Adjudicates
Two documents disagree about whether the increase is permitted at all, and neither of them is the one that would settle it. One still writes the project as capped at the published branch point; another states that approval of actual cabinet suitability is not required. The first reads as a limit and the second as a disclaimer of the limit, and the interface matrix that would adjudicate between them, giving allowable flow, allowable pressure and permission to switch, has not been supplied.
What follows for the enquiry is a separation the review asks for in plain terms. The operating point and the independent 900 kW test point are two different things, and the exchanger selection flow may not be read as the cabinet's requirement. Holding them apart costs nothing at this stage and is expensive to unpick after an acceptance criterion has been written against the wrong one.
Until the interface matrix is supplied, this finding reaches the extrapolation and stops at the cabinet connection. The review declines to call the selection inadequate on the strength of a missing curve, and records the whole assessment as reading supplied files with no physical test run; the supplier's proposed equivalent-load test could pass at the raised flow and would still leave the allowable-flow and allowable-pressure columns empty.
That is also the limit of what the extrapolation can ever deliver. A squared resistance predicts a pressure drop at a flow; it says nothing about the temperature, flow, pressure and connection boundary of the equipment the flow has to pass through, and an equivalent-load test on a representative circuit does not supply that boundary either. No document in the package contains it.
Limits and open items
Confirmed: exceeding a published table point does not by itself mean exceeding an allowable limit; an equivalent-load test does not replace the cabinet temperature, flow, pressure and connection boundary. [FACT]
Supplier points, not a freeze: 53.1 to 67.3 cubic metres an hour; five branches at 224.333 litres a minute against a published 177; a 2 metres a second self-set screen. [TARGET]
Extrapolations: 26.742 per cent; 203.788 kPa from a recovered 126.9 kPa anchor; 2.276 metres a second through 102.26 millimetres bore. [MODELED]
Open: the interface matrix; a manufacturer pressure-drop curve; which of the two files governs the cap. [HOLD]
Nothing here has been reviewed or sealed by an Engineer of Record. Nothing describes how any installed system behaved.
Name the Document Allowed to Cap Branch Flow
The useful test is not whether 67.3 cubic metres an hour is high. Name the document an acceptance criterion is allowed to copy the branch flow from. If that name is missing, the square law is still doing the interface's job.
© 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., “Raised CDU Flow Was Square-Law Extrapolated From One Published Point,” https://www.kkdatasvc.com/lab/errata-and-open-questions/a-26-7-per-cent-flow-rise-squared-off-the-only-published-point/.
Discuss This Constraint. 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.