Shell & tube heat exchanger estimator

Cooling / process water defaults are representative values around 30 C average bulk temperature; adjust for actual site water chemistry and treatment (fouling factor especially varies widely with water quality and velocity).

Hot stream in red, cold stream in blue, plotted against fraction of total duty transferred (0% = cold inlet / hot outlet end, 100% = cold outlet / hot inlet end). The pinch marker shows the point of closest approach across the whole profile, not just the two terminals — for a heat-curve service this is often away from either end. A flat segment on either line marks an isothermal (condensing/vaporizing) stretch. The duty badges on the Hot side and Cold side panels above update live as you edit flow, temperatures or Cp — use them to bring both sides into balance before moving on to geometry.

Typical U guide (W/m2.K): gas-gas 10-50, gas-liquid 15-70, liquid-liquid 300-1000, condensing steam 1000-4000, water-cooled condensers 600-1200. This "Overall U" is your working assumption used to size area, tube count, and shell diameter; the Film coefficients panel below computes U independently from correlations, using that shell diameter, so you can check the assumption and iterate by feeding the calculated value back in here. "Kettle (K) reboiler" only takes effect when the shell-side fluid (the one not on the tube side) is vaporizing/boiling — it replaces the bundle-set shell ID with one sized for vapor disengagement above the liquid pool, using the Souders-Brown constant K (default 0.107 m/s, a typical value for a horizontal kettle without a mesh pad; use a higher value, e.g. 0.15-0.20 m/s, if a demister pad is fitted).

Tube length options are the common shop-fabricated lengths (8/10/12/16/20/24 ft, i.e. 2.44/3.05/3.66/4.88/6.10/7.32 m); each chip is colored by the length/shell-ID ratio it produces at the current OD, layout, passes and area — green sits in the typical economic range (L/D 5-12), plain is workable but outside that sweet spot, and greyed is impractically stubby or slender for the current duty. Tube passes are colored by the resulting tube-side velocity against the usual 0.9-3.5 m/s liquid guide (green = within range, plain = outside it but still usable, grey = unusable, e.g. odd passes with a U-tube rear head, or velocity can't be computed). Both are a sizing screen, not a final selection — very long tubes may not fit the plot plan's bundle-pull space, and gas/vapor services can run well above the 3.5 m/s liquid guide, so confirm against the actual site layout and line-sizing criteria.

Tube positions are generated directly on the selected pitch/layout (triangular or square) within the bundle-diameter circle from the sizing section above, giving an independent geometric tube count alongside the correlation-based Nt, and letting the baffle-window tube fractions (NTIW/NTCC) below be counted rather than estimated. These feed a simplified Bell-Delaware baffle-configuration correction, Jc, applied as a multiplier on the Kern shell-side film coefficient below for sensible (single-phase, non-condensing/boiling) shell-side service — see the tubesheet layout drawing further down for the visual basis. This is a partial Bell-Delaware treatment: it does not include the Jl (baffle-to-shell and tube-to-baffle-hole leakage) or Jb (bundle-to-shell bypass) correction factors, both of which typically further reduce the ideal crossflow coefficient in a real exchanger, nor the associated shell-side pressure-drop method — treat the corrected coefficient as a refinement over plain Kern, not a substitute for a full Bell-Delaware or HTRI/vendor rating.

Methodology: tube-side and single-phase shell-side films use Dittus-Boelter (Nu = 0.023 Re^0.8 Pr^n, n = 0.4 heating / 0.3 cooling), with the shell-side Reynolds and Prandtl numbers built on Kern's equivalent diameter and crossflow area (from the pitch ratio and baffle spacing set above, applied to the shell ID already sized in the Thermal performance section). Condensing shell-side film uses Nusselt laminar filmwise condensation on a horizontal tube bundle (h = 0.725*(k_l^3 rho_l(rho_l-rho_v) g h_fg / (Nr mu_l dT Do))^0.25, Nr taken as sqrt(tube count)), with the film dT approximated from the zone's corrected mean temperature difference. Boiling shell-side film uses a Chen (1966) convective term (Martinelli two-phase multiplier F on a liquid-alone Dittus-Boelter coefficient, with the Chen suppression factor S) plus a Cooper (1984) nucleate pool-boiling term, with heat flux seeded from the manual-U area estimate above (not iterated to convergence) and reduced pressure taken as design pressure / critical pressure. The resistance network combines 1/hi, tube-side fouling, the tube wall (conductivity from the selected material), shell-side fouling, and 1/ho on an outside-area basis. If the phase-change fluid is assigned to the tube side rather than the shell side, the tube-side coefficient still falls back to single-phase Dittus-Boelter as an approximation — flagged below when that happens. Surface roughness for the Cooper correlation is fixed at 1.0 µm (typical commercial tubing); for the isothermal (non-curve) boiling case, vapor viscosity is not a separate input and is assumed at 0.01 cP (typical of a low-pressure vapor) and vapor quality at a representative midpoint of x = 0.5.

Tube-side: Darcy-Weisbach friction (f = 0.184 Re^-0.2 turbulent, 64/Re laminar, a Blasius-type approximation to Kern's original tube-side friction chart) over the total tube-side travel (length × passes), plus a 4-velocity-head allowance per pass for the 180° channel return between passes. Shell-side: Kern's method, ΔP = f Gs^2 Ds (Nb+1) / (2 ρ De), using the same shell-side mass velocity Gs, equivalent diameter De and Reynolds number as the shell-side film coefficient above, with f a curve fit approximating Kern's shell-side friction-factor chart and the (Nb+1) crossflow passes between baffles from the baffle count above; the (μ/μwall)^0.14 viscosity correction is neglected. Both are single-phase (Kern) formulations — for a condensing or boiling side, the liquid-phase properties are used as a screening approximation, which will overstate ΔP somewhat since vapor generation/condensation reduces the effective liquid travel and, for two-phase flow, a proper two-phase multiplier (e.g. Lockhart-Martinelli) is not applied. Nozzle entrance/exit and inlet/outlet header losses are not included on either side. No allowable ΔP is entered here — typical liquid-service guidelines are roughly 35-70 kPa (5-10 psi) per shell/channel, lower for gas or vapor service; confirm against the actual process/line-sizing specification.

Simplified selection logic based on shell/tube-side fouling, need for a removable bundle, gas pressure-drop sensitivity, and thermal expansion (front head / shell type / rear head). The pick list above lets you override any individual letter directly; picking Kettle (K) or U-tube (U) also updates the "Shell configuration" / "Bundle construction" selectors above so the mechanical sizing stays in sync, while the other letters are label-level choices only (this tool's LMTD/area method doesn't distinguish F/G/H/X shells or P/T/W/C/D/N heads, so those stay greyed out). U-tube still requires an even tube-pass count and is unavailable if shell-side fouling calls for a pullable bundle, since a U-bend can't be mechanically cleaned like a straight tube. Confirm against project specification and client TEMA class (B/C/R) and the TEMA Standards nomenclature.

Velocity limit uses the API RP 14E erosional-velocity form, Vmax = 122/sqrt(density in kg/m3), applied to both liquid and gas/vapor nozzles as a first-pass screen. For a heat-curve side, inlet/outlet densities come from the first and last curve points (mixture specific volume basis). Confirm against line sizing criteria and, for two-phase or flashing service, a momentum (rho V^2) check.

Nozzle count and velocity are adjusted for the recommended TEMA configuration above, not sized in isolation: (1) a divided-flow shell (J) keeps the shell-side inlet as a single nozzle at the shell's mid-length, carrying the full shell-side flow, and splits the shell-side outlet into two nozzles, one near each end, each carrying half the flow after it divides internally and travels in both directions along the shell — this halved flow path (and halved per-nozzle/per-path velocity) is what gives a J-shell its lower shell-side pressure drop versus a single long E-shell pass; see the sketch below for the arrangement. (2) any single-nozzle calculated ID that comes out impractically large relative to the shell (greater than 500 mm/20 in, or over 60% of the shell ID) is likewise split into two nozzles of half the flow each, a common shop/fabrication practice rather than a strict TEMA rule; (3) for a kettle (K) shell, the shell-side outlet is the vapor-product nozzle and its velocity is capped at the Souders-Brown disengagement limit from the shell-sizing calculation above (rather than the API RP 14E value alone) whenever that is the tighter constraint, since an oversized/high-velocity vapor nozzle can re-entrain liquid droplets carried up from the pool. Tube-side (channel) nozzles are left single unless the oversize check above triggers a split. TEMA's published Standards (front head / shell type / rear head letters and the associated nomenclature diagrams) govern the configuration naming used throughout this tool; see the TEMA Standards FAQ for the official nomenclature reference — note it is an FAQ on interpreting the Standards, not a substitute for the full Standards or a detailed nozzle-sizing specification, so confirm final nozzle count/size/orientation against the full TEMA Standards and vendor practice.

A proportionally-scaled schematic elevation, not a fabrication drawing — front head, shell, and rear head shapes follow the recommended/selected TEMA letters above (removable cover vs. bonnet vs. U-bend vs. floating head), the shell is drawn to the estimated ID and tan-tan tube length, and nozzles are placed roughly where that TEMA configuration would put them (channel end for tube-side, shell body for shell-side, top of the kettle vapor dome for a K-shell vapor outlet). Actual nozzle orientation, support saddle location, and head details are set during detailed mechanical design.

Segmental baffles drawn as vertical plates spanning the shell ID except for the cut segment set above, alternating top and bottom so consecutive baffles are 180° apart — the standard single-segmental arrangement this tool's Kern-method calculations assume. Baffle count and spacing come directly from the Baffle & tube-layout geometry panel above; actual baffle thickness, tie-rod/spacer detail, and orientation (cut horizontal vs. vertical, which affects deposition and venting) are set during detailed mechanical design.

Tube positions generated on the selected pitch and layout pattern (triangular/30° or square/90°), kept within the bundle-diameter circle sized in the Thermal performance section above. The dashed line marks the baffle-cut window boundary used for the NTIW/NTCC counts and Jc correction above — tubes above/below it (in either baffle window) are shown in a lighter shade. This is a schematic, uniform-pitch layout for estimating purposes, not a certified tube count/layout drawing: it does not remove tubes for pass-partition lanes, impingement plates, tie rods, sealing strips, or U-bend clearance, all of which a vendor's real tubesheet layout would account for and which typically reduce the actual installable tube count somewhat below the number shown here.

Tubes are grouped into the selected number of tube-side passes using the standard radial partition convention: one straight diametral plate for 2 passes, or equal pie-slice sectors radiating from the tubesheet center for 4/6/8 passes, with the pass sequence running around the sectors so the inlet and outlet end up in adjacent sectors near the top — matching how a real multi-pass channel's two nozzles are mounted side by side. This is still schematic, not a vendor partition-plate design: actual plate placement also balances flow area between passes, allows for partition-plate thickness and minimum width, and follows the channel nozzle orientation actually specified, which can shift sector boundaries away from perfectly equal angles. Use this drawing to sanity-check the pass count and general arrangement against the tube count, not as a fabrication partition-plate layout.