Isolation Raises Ring ΔP From 3.5 to 49.2 kPa

Close one ring span to repair a weld. Zero racks lose cooling. The whole zone duty now travels one way. On the current calculation that move takes the ring pressure drop from 3.5 kPa to 49.2 kPa, an increase of 45.7 kPa [MODELED]. Read against the loop total of 235 to 450 kPa built up in the same document, that increment is 10% to 19% of everything the secondary pump has to overcome [MODELED]. Recommendation R-13 turns it into a selection rule: pump head is chosen as rated-condition pressure drop plus a 50 kPa isolation detour allowance, and the allowance must be written explicitly into the unit specification rather than folded into a safety factor [TARGET].

Most of the Isolation ΔP Rise Is Doubled Velocity

Both rows use the same friction gradient and the same pipe. Normal operation splits flow both ways around the ring, which the hydraulic basis takes as an equivalent resistance of about a quarter of the one-way full length: 120 m of run, multiplied by 1.5 for fittings, divided by 4, gives 45 m equivalent. At 78 Pa/m that is 3.5 kPa.

Close one span and the split disappears. Length goes to 165 m equivalent and velocity doubles from 1.41 to 2.82 m/s, so the gradient rises by the velocity ratio squared with a small friction-factor correction — 78 × (2.82/1.41)² × 0.955 = 298 Pa/m — and 298 Pa/m over 165 m is 49.2 kPa [MODELED]. Splitting the two effects is worth doing: at the unchanged 78 Pa/m the longer path alone would give 12.9 kPa, so of the 45.7 kPa rise about 9.4 kPa is extra length and the remaining 36.3 kPa is velocity [MODELED]. Routing care can trim the first term and cannot touch the second.

That is why the factor of fourteen is geometry rather than friction. The pipe did not get rougher. The two-way split went away, the path lengthened, and the velocity doubled on the same bore.

The Friction Gradient Is Derived; Roughness Is Assumed

78 Pa/m is derived, not given. Stainless 316L at a roughness of 0.045 mm gives a relative roughness of 2.13 × 10⁻⁴, a Reynolds number of 261,700 and a Colebrook friction factor of about 0.0164; the Darcy expression with a density of 1,012 kg/m³ and 1.41 m/s over a 0.21158 m bore returns 78 Pa/m [MODELED]. The hydraulic basis grades the roughness itself an assumption while grading the derivation a fact, and that split is the honest part of the entry.

Density is borrowed from the same assumption set. The parametric engine records the working fluid's 1,012 kg/m³ and 3.921 kJ/(kg·K) as engineering assumptions held under item H-05, so both the friction gradient here and the flow that produced the velocity inherit an unconfirmed property pair. The same engine states that a uniform split is for comparison only. Without manufacturer curves, temperature-dependent properties, pipe Δp, valve authority, pump curves and a measured balance, it cannot size a pump, a pipe, or a unit operating point.

Isolation Sizes the Ring; a Vendor Snapshot Does Not

Neither row sets the pipe. Conclusion 22 states that DN200 satisfies both the normal and the detour velocity criteria at a 10 K temperature difference, but with only 6% margin at a liquid fraction of 0.95 and 1.3% at 1.00, and that the ring is therefore sized by the detour rather than by normal velocity [MODELED]. Remove segment isolation and DN150 would pass normal duty at 2.51 m/s; raise the liquid fraction or lower the temperature difference and the answer must jump straight to DN250 rather than find margin inside DN200.

So one design decision — that a weld can be repaired without stopping racks — sets the ring bore, then sets the pump head, and the 49.2 kPa is where the second of those two consequences is priced. An isolation capability which exists on paper but not in pump head produces insufficient flow to the racks on the detoured side. That is the failure mode of a half-purchased feature.

A vendor distribution-unit packet was independently recomputed on 864 ring pipe segments, segment by segment, against Darcy, Swamee–Jain and transition-region formulations. It reports a maximum pressure-drop difference of 6.67 × 10⁻¹⁶ kPa and a closed-loop residual of 2.12 × 10⁻¹⁵ kPa across 144 supply-and-return loops [MODELED]. The reviewer then states what that does and does not mean: it is a snapshot of arithmetic self-consistency under a fixed topology, an imposed equal-flow assumption and assumed local-loss coefficients. It is not a re-solution of a real network. It is not evidence that real branches balance themselves. It is not this plant's zone ring. The accompanying calculation program was not supplied, so changing an input in the spreadsheet does not re-solve anything. A companion file in the same packet claimed 1,152 ring rows [MODELED]. The 864-row sheet is what was actually present.

The same review records that dynamic performance was deferred to a post-contract physical test, and that a required minimum differential-pressure statement covering one-sided pump loss, relief, isolation, dead-head and fouling is still owed. An isolation detour is exactly the condition that statement would cover. Steady snapshot arithmetic cannot prove switching without flow interruption or water-hammer.

Limits and open items

Confirmed on the current record: single-span isolation raises the computed ring pressure drop from 3.5 kPa to 49.2 kPa; DN200 is sized by the detour velocity check. [FACT]

Target, not confirmed: pump head is rated-condition pressure drop plus a 50 kPa isolation detour allowance, written explicitly into the unit specification rather than folded into a safety factor. [TARGET]

Modelled and not frozen: 45.7 kPa, 10% to 19% of 235 to 450 kPa, 45 m and 165 m equivalent lengths, 78 and 298 Pa/m, the 9.4 / 36.3 kPa split, 6% and 1.3% velocity margins, and the 864-row residual pair. The 120 m circumference and the 1.5 fittings multiplier are placeholders until the route is drawn. [MODELED]

Held open, beside the figure it bounds: item 07A-CALC-13, ring hydraulic balancing and zone flow distribution including branch valve authority, is recorded as not started [HOLD]. If two spans have to be isolated at once — one for planned work and one for a fault, which is the maintenance-plus-failure case the capacity section does not yet check — the one-way detour assumption behind 165 m no longer holds, and the 50 kPa allowance is being asked to cover a case it was not computed for.

Nothing above reflects Engineer of Record review, supplier confirmation, commissioning acceptance or observed operation. No pump duty here is a selection or a procurement release.

Write the 50 kPa Isolation Allowance on the Pump Sheet

The 50 kPa is not a comfort margin. It is the price of being able to isolate a span. If that line is missing from the unit specification, the isolation valves can still be bought, and the weld can still be cut, and the racks on the long way around will not see the flow the velocity screen assumed they would.

© 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., “Isolation Raises Ring ΔP From 3.5 to 49.2 kPa,” https://www.kkdatasvc.com/lab/hydraulics-and-transients/isolation-raises-ring-delta-p-from-3-5-to-49-2-kpa/.