Follow Us
0
Login
0
Login

Indirect Water Heater Design From First principles

Home Indirect Water Heater Design From First principles

Indirect Water Heater Design From First principles

By /
D
Design Hub India
Process Design Series · Module 04
Live Calculator Est. 60 min v1.0
DOC-DHI-PD-04
REV. 1 · 2026
OPEN COURSEWARE
Process Design Series
Module Four of Twelve

Indirect Water
Bath Heater Design
From First Principles.

A guided eight-step worksheet that teaches the why behind every number. Code-anchored to API 12K, ASME B31.3, and NIGC IGS-M-PM-104. Built for working engineers and the next generation learning the craft.

08
Design Steps
12+
Knowledge Checks
04
Codes Referenced
PDF↓
Export Report
01

The Design Basis.
What Are We Solving For?

Every calculation begins with a clear-headed statement of what’s flowing, what temperature it enters, and what temperature it must leave. Get this wrong and the rest is cosmetic. Read the well test report, the upstream conditions, the downstream constraint — and write down the numbers that matter.

API 12K § 1.1 · DJG Review · Datasheet Annex A

What an IWBH actually does

An Indirect Water Bath Heater sits on the producing flowline between wellhead and pipeline. Wellstream fluid flows through a coil immersed in a water bath; the bath itself is heated by a submerged firetube fired with natural gas. The bath acts as a thermal buffer — the firetube never touches the process fluid, which is why this design is preferred for hydrocarbon service over direct-fired heaters.

Three reasons engineers specify them: (1) prevent hydrate formation in cold gas streams, (2) break the produced emulsion before separation, and (3) heat fuel gas or flowline contents to prevent wax deposition. Mustang Pad is reason (2) — the cold North Slope wellstream needs to hit 140°F so the oil and water can disengage downstream.

Process Inputs
Editable · Auto-recalc

Flow Rates

BPD
Stock tank, max instantaneous
BPD
Produced water
MMSCFD
At 14.7 psia, 60°F

Fluid Properties

—
SGo = 141.5/(131.5+°API)
—
lb/lbmol
BTU/lb·°F
Heavy crude ≈ 0.45–0.50
BTU/lb·°F
BTU/lb·°F
Rich HC ≈ 0.50–0.60

Temperatures

°F
Winter pad design low
°F
Per DJG Rev 1 · emulsion break
°F
Below boiling, typ. 160–190
Q1.01 — Why is the bath temperature held below 200°F (well below water’s boiling point)?
A To save fuel gas — heating water to boiling wastes energy
B Per API 12K Cl. 3.11 the heater bath must be maintained below the boiling point — boiling would cause bath loss, scaling, and uncontrolled steam pressure in a vessel rated only for ≤1 psig
C The coil material (A106 Gr.B) cannot withstand temperatures above 212°F
D Insurance regulations cap bath temperature at 180°F
Step 1 of 8
02

Volumetric to Mass Flow.
The Conversion Engineers Always Forget.

Field rates come in barrels and standard cubic feet — but the heat duty equation needs lb/hr. Two unit conversions, three lines of arithmetic, but the small constants (5.615 ft³/bbl, 62.4 lb/ft³, ideal gas at 14.7 psia & 520°R) are where students lose marks.

API 12K Annex D · Industry Standard Conversions

Three conversions you must memorize

Liquid: One US oil barrel equals 5.615 ft³. Water at standard conditions weighs 62.4 lb/ft³. So for any liquid: ṁ (lb/hr) = Q (BPD) × 5.615 × SG × 62.4 / 24.

Gas: Standard conditions in the US oil patch are 14.7 psia and 60°F (520°R). Apply ideal gas: ρstd = (14.7 × MW) / (10.73 × 520). Then ṁg (lb/hr) = Qg (MMSCFD) × 10⁶ × ρstd / 24.

Sanity check: A 35°API crude (SG ≈ 0.85) at 5,000 BPD comes out near 62,000 lb/hr. Methane at 6 MMSCFD comes out near 10,500 lb/hr. If your numbers are 10× off, you’ve dropped a power of ten somewhere.

EQUATION 2.1
ṁliq = Q × 5.615 × (SG × 62.4) ÷ 24
ṁgas = Qg × 106 × (14.7 × MW)/(10.73 × 520) ÷ 24
Q = volumetric rate, BPD (liquid) or MMSCFD (gas) · SG = specific gravity · MW = molecular weight, lb/lbmol · 5.615 = ft³/bbl · 62.4 = lb/ft³ for water at std · 10.73 = ideal gas constant in psia·ft³/(lbmol·°R) · 520 = 60°F + 460
Q2.01 — A junior engineer reports the gas mass flow as 173,000 lb/hr for 6 MMSCFD of MW=26.2 gas. What’s wrong?
A The molecular weight is too low — should be ~50
B The standard conditions are wrong — should use 0°C and 1 atm
C They forgot to divide by 24 hours — they reported lb/day instead of lb/hr
D Nothing — that’s the correct value
Q2.02 — For a gas with MW = 18 (mostly methane), what is the standard density at 14.7 psia and 60°F?
A 0.0123 lb/ft³
B 0.0474 lb/ft³ — calculated as (14.7 × 18) ÷ (10.73 × 520)
C 0.789 lb/ft³
D 1.225 lb/ft³ — same as air
03

Heat Duty.
The One Number That Sizes Everything.

Q = ṁ·Cp·ΔT for each phase. Sum them up, divide by burner efficiency, and you have the firebox rating that drives every downstream sizing decision. API 12K’s Annex D gives two routes — explicit per-phase, or the empirical emulsion form. Smart engineers run both and use the conservative.

API 12K Annex D § D.3 · Eq. (D.2), (D.3), (D.4)

Two methods, one answer (almost)

Method A — Explicit per-phase. Treat oil, water, and gas as independent streams. Each absorbs Q = ṁ × Cp × ΔT. Sum the three. This is fundamentally correct and works for any oil API.

Method B — Annex D.3 emulsion formula. API 12K offers a shortcut for liquid-only: Q = W × [6.44 + 8.14·X] × ΔT, where W is total liquid in BPD and X is water volume fraction. The 6.44 and 8.14 constants are derived assuming 35°API oil (Cp = 0.52) and water (Cp = 1.0). For other crudes, treat it as an approximation.

Method C — Conservative. Take max(A, B), then divide by burner efficiency η to get gross fired duty. Typical η for an indirect heater with natural draft burner: 0.70–0.80, often 0.75.

EQUATION 3.1 · Method A
Qtotal = (ṁo·Cpo + ṁw·Cpw + ṁg·Cpg) × (T2 − T1)
EQUATION 3.2 · Method B (API 12K Eq. D.3)
Qliq = W × [6.44 + 8.14·X] × (T2 − T1)
W = oil + water, BPD · X = Qw / (Qo + Qw) · gas added separately as ṁg·Cpg·ΔT
EQUATION 3.3 · Fired Duty
Qfired = Qtotal ÷ ηburner
—
App. C Fig. C.1, typically 0.70–0.80
§
About the “5 MMBTU/hr myth” API 12K Table 4 lists preferred firebox ratings up to 5 MMBTU/hr — this is a preferred sizes list, not a scope ceiling. Cl. 5.3 explicitly permits ratings outside the table “unless otherwise agreed upon between the purchaser and manufacturer”. NIGC IGS-M-PM-104 Cl. 4.5.2.11 reads it correctly: above 5 MMBTU/hr → two firetubes mandatory, but the heater is still within scope. Don’t reject a design just because it’s above the table.
Q3.01 — Your fired duty calculates to 8.45 MMBTU/hr. Which response is correct per API 12K?
A Reject the design — API 12K caps firebox rating at 5 MMBTU/hr
B Round down to the next preferred size (5 MMBTU/hr) and accept higher outlet temp drift
C Above 5 MMBTU/hr requires purchaser/manufacturer agreement (Cl. 5.3) and at least two firetubes (NIGC Cl. 4.5.2.11) — proceed with custom rating
D Switch to a direct-fired heater since IWBHs cannot exceed 5 MMBTU/hr
Q3.02 — Why does Method A typically give a higher duty than Method B for heavy crude (≤25°API)?
A Method B’s 6.44 + 8.14·X coefficients assume 35°API oil with Cp = 0.52; heavy crude has lower Cp (~0.45–0.48), so Method B over-estimates oil contribution
B Method A includes gas duty; Method B doesn’t include any gas at all
C Method B uses metric units internally and the rounding error compounds
D They always give exactly the same answer — the question’s premise is wrong
04

The Coil.
Where B31.3 Meets the Wellstream.

The coil is the only pressure boundary in the heater that handles full process pressure. API 12K Cl. 5.1.1 anchors the design to ASME B31.3 with two equations — required wall thickness for a chosen design pressure, or maximum allowable pressure for a chosen schedule.

API 12K § 5.1.1 Eq. (1) & (2) · ASME B31.3 · Table 1 (S) · Table 2 (P)

Reading the formula

The coil sees full upstream pressure of the choke (or the well shut-in pressure if no choke is upstream). Material is seamless A106 Gr.B (S = 20,000 psi) or Gr.C (S = 23,300 psi) per API 12K Table 1. Mill tolerance shaves the actual wall to tm = 0.875·T. Y is a weak-pipe correction — for small-bore high-pressure pipe (T < D/6) it equals 0.4.

NIGC IGS-M-PM-104 mandates seamless Sch. 80 minimum and ANSI 600 weld-neck flanges regardless of what B31.3 calculates. This is a prescriptive overlay — read your project spec.

EQUATION 4.1 · API 12K Eq. (1)
tr = P·D ÷ [2·(S + P·Y)] + C
tr = required wall, in · P = design pressure, psig · D = pipe OD, in · S = allowable stress (Table 1) · Y = 0.4 if T < D/6, else d/(d+D) · C = corrosion + erosion + thread allowance
EQUATION 4.2 · API 12K Eq. (2)
Pmax = 2·S·(tm − C) ÷ [D − 2·Y·(tm − C)]
Solve for P given a chosen pipe schedule. Round up to next 10 psig per Footnote 2. Cross-check against Table 2.
Coil Inputs
B31.3 / API 12K Cl. 5.1.1
NPS
—
NIGC mandates ≥ Sch.80
psi
psig
Choke upstream or well SI pressure
in
Typical 1/16″ to 1/8″
in
Sand service: 0.05–0.125
Q4.01 — For 3″ XS A106 Gr.B with C=0, API 12K Table 2 lists Pmax = 3,200 psig. Your calculation gives 2,560 psig. What’s the most likely cause?
A The mill tolerance factor (0.875) is wrong — should be 0.95
B You included a corrosion allowance (Ca > 0); Table 2 assumes C = 0. Set Ca = 0 to reproduce the table value
C The Y coefficient is wrong — 3″ XS has T < D/6 so Y must be 0.4, but you used d/(d+D)
D You used the wrong S — should be 23,300 not 20,000 for Grade B
Q4.02 — Per API 12K Cl. 5.1.5, what happens if your coil design pressure exceeds the Pmax from Eq. (2)?
A You can use a higher schedule and proceed under the same monogram
B Increase the joint efficiency E to 1.15 to compensate
C Apply a derating factor of 0.67 and retest
D Heater coils for pressures greater than Eq. (2) shall NOT be furnished under this specification (Cl. 5.1.5) — design is outside API 12K scope, must be done as a custom B31.3 piping spec
05

LMTD & Coil Area.
How Big Does the Coil Need to Be?

The pressure check made sure the coil won’t burst. The area calculation makes sure it actually heats the fluid. Q = U·A·LMTD — three quantities, two of them already known, solve for A.

API 12K Annex D § D.2, D.4, D.5, D.6

The bath behaves as a constant-temperature reservoir

Because the bath is a large pool of water held at near-uniform temperature, the bath-side ΔT is essentially zero across the coil. So the LMTD reduces to the temperature approach between the bath and the process fluid at each end. The greater approach (GTD) is at the cold inlet, the lesser approach (LTD) at the hot outlet.

The overall heat transfer coefficient U is the trickiest number. App. D.4 explicitly says it’s “established by the manufacturer based on laboratory and field experience” — meaning there’s no clean formula for general service. For estimation: U ≈ 50–70 BTU/hr·ft²·°F for mixed oil/water/gas in a still water bath, U ≈ 150–250 for clean gas service (NIGC’s tabulated values back-calculate to ~200). Glycol addition reduces U by 15–25%.

EQUATION 5.1 · LMTD (Annex D Eq. D.5)
LMTD = (GTD − LTD) ÷ ln(GTD ÷ LTD)
GTD = Tbath − T1 · LTD = Tbath − T2
EQUATION 5.2 · Coil Area (Annex D Eq. D.1)
A = Q ÷ (U × LMTD)
A = required coil OD surface area, ft² · Q = total process duty (not fired), BTU/hr · U = overall coefficient, BTU/hr·ft²·°F
EQUATION 5.3 · Total Coil Pipe Length
L = A ÷ (π × OD)
Where OD is in feet. Coil area is the OUTSIDE surface area per API 12K Cl. 3.3.
BTU/hr·ft²·°F
Mixed oil/water/gas: 40–70. Clean gas: 150–250. Glycol bath: × 0.80
in
!
Velocity check (NIGC Cl. 4.4.9) Maximum allowable gas velocity in coil = 20 m/s (66 ft/s). API 12K Annex B uses 80 ft/s as the design reference. Always check the gas-side velocity in your coil against erosion criteria (API RP 14E for two-phase) — high velocity in the coil causes acoustic vibration, erosion, and pressure drop.
Q5.01 — Why is the LMTD formula simplified for an indirect water bath heater (compared to a counter-current shell-and-tube)?
A The bath is at lower temperature than the process fluid
B Indirect heaters always use parallel flow, never counter-current
C The bath behaves as a constant-temperature reservoir (large thermal mass, well-mixed by firetube convection), so there is no bath-side temperature change to integrate over
D The correction factor F is always 1.0 for water bath heaters by definition
Q5.02 — A Mustang Pad calc gives required coil area = 1,440 ft² with U=55. Your colleague claims that NIGC’s table gives only 92 m² (~990 ft²) for similar duty — so 1,440 must be wrong. What’s actually going on?
A Your LMTD calculation is wrong — recheck the natural log
B NIGC’s table is for clean gas service (city-gate stations) where U ≈ 200 BTU/hr·ft²·°F. Mustang’s mixed oil/water/gas service has U ≈ 55, so 1,440 ft² is correct for the harder service
C NIGC uses a different LMTD formula — the corrected NIGC area is actually 1,440 ft²
D Your colleague is right — switch to NIGC’s table
06

Firetube.
The Heat Flux & Heat Density Twins.

Two checks, both prescriptive. Heat flux (BTU/hr per ft² of firetube surface) controls bath-side film temperature and prevents hot spots. Heat density (BTU/hr per in² of firetube cross-section) controls flame-to-tube clearance and bounds the burner size. API 12K Cl. 5.4 and 5.5 — code-mandated limits, not suggestions.

API 12K § 5.4 (Heat Flux ≤ 12,000) · § 5.5 (Heat Density ≤ 15,000)

Why both limits exist

Heat flux is the average rate at which heat crosses the firetube wall, per unit of OD surface area. Above 12,000 BTU/hr·ft² (glycol/water bath), the water film on the bath side starts to nucleate boil locally — creating vapor pockets that insulate the firetube and let it overheat. Fresh water bath can take more flux because it boils at a fixed temperature without the depression glycol introduces.

Heat density is the heat release per unit of firetube internal cross-section — directly related to flame combustion intensity. Above 15,000 BTU/hr·in² for natural draft burners, the flame becomes too long for a typical firetube U-tube and starts impinging on the return bend. Flame impingement = local hot spots = firetube failure within months.

EQUATION 6.1 · Heat Flux
q = Qfired ÷ Afiretube   ≤   12,000 BTU/hr·ft²
Afiretube = π × ODft × Lft for U-tube (both legs combined)
EQUATION 6.2 · Heat Density (Cl. 5.5 worked example)
HD = Qfired ÷ (Across-section × η)   ≤   15,000 BTU/hr·in²
Across-section = π/4 × IDin² · η is the burner combustion efficiency
Firetube Inputs
API 12K Cl. 5.4 / 5.5
in
Larger OD → lower flux for given duty
in
NIGC Cl. 4.8.1 minimums
ft
U-tube total length = 2 × this
no.
NIGC Cl. 4.5.2.11
Q6.01 — Your firetube heat flux calculates to 18,000 BTU/hr·ft². Worst-case consequence if you proceed?
A Burner stops igniting after 200 hours
B Bath water boils away within the first week of operation
C Localized nucleate boiling on the bath-side firetube surface forms an insulating vapor layer; firetube wall overheats, hardness rises, and stress-corrosion failures occur within months. Heat flux is a code-mandated upper bound for a reason
D The fired duty rating drops by 30%, but operation continues normally
Q6.02 — Per API 12K Cl. 5.4, the heat flux limit may be increased above 12,000 BTU/hr·ft² for which condition?
A Forced-draft burner instead of natural draft
B Fresh water bath (no glycol) — Cl. 5.4 explicitly says “may be increased for fresh water bath applications”
C Vertical orientation of the firetube
D Multi-pass serpentine firetube geometry
07

Shell.
Atmospheric Tank, Not a Pressure Vessel.

The single most-misunderstood part of API 12K. The IWBH shell is not designed to ASME Section VIII. It is designed to AISC structural rules with a 1.5 psi (or full-of-water) pressure case. Internal operating pressure ≤ 1 psig. Period.

API 12K § 5.2.1 (≤1 psig) · § 5.2.3 (3/16″ plate min) · § 5.2.4 (AISC Manual)

Why this matters

If you (or your vendor) design the shell to ASME VIII Div.1 with thick walls, you’ve over-designed by 5–10× — wasting steel, weight, and shipping cost. If you under-design and the pressure-vacuum vent gets blocked, the shell can collapse on cooling or balloon on heating. The vent (Cl. 3.5) and water saver / expansion tank (NIGC Cl. 4.2) are the safety devices, not the shell wall thickness.

NIGC overlay. NIGC IGS-M-PM-104 mandates a separate expansion tank on top of the shell sized for 100% of bath thermal expansion (Cl. 4.2.2), with a 6% reservoir (Cl. 4.2.3) plus low-level alarm (LLA) and low-level shutdown (LLS) per Cl. 4.6.7.3. Mandatory 50 mm fiberglass insulation with 0.8 mm aluminum cladding (Cl. 4.1.4).

EQUATION 7.1 · AISC Allowable Stress
σallow,bend = 0.66 × Fy  |  σallow,tens = 0.60 × Fy  |  τallow = 0.40 × Fy
Per API 12K Cl. 5.2.4 — note these are AISC structural allowables, not ASME stress tables. Fy = 36,000 psi for A36 / A283-C / A285-C.
EQUATION 7.2 · Hoop Stress (Cylindrical Shell, 1.5 psi case)
σhoop = P × R ÷ t   ≤   0.60 × Fy
P = max(1.5 psi, full-water head). R = D/2. t ≥ 3/16″ per Cl. 5.2.3 minimum.
Shell Inputs
API 12K Cl. 5.2 + NIGC overlay
in
Std diameters: 36/42/48/54/60/72
ft
Match firetube length
in
Cl. 5.2.3 minimum 3/16″ (0.1875)
psi
+
NIGC overlay items to specify (best practice) Mandatory expansion tank (sized for 100% thermal expansion + 6% reservoir) · 50 mm fiberglass insulation + 0.8 mm aluminum cladding · 2″ Cl.150 gate-valve drain · LLA + LLS bath level switches · 100% RT on all T-joint welds · Pressure-vacuum vent on fill connection · Permanent lifting lugs sized for 75% of empty assembly weight.
Q7.01 — A vendor proposes a 1.25″-thick shell wall designed to ASME VIII Div.1 UG-27 for 1,440 psig. What’s the best response?
A Accept it — thicker is always safer
B Reject — vendor must use ASME VIII Div.2 instead
C Reject — API 12K Cl. 5.2.1 limits shell pressure to 1 psig and Cl. 5.2.4 mandates AISC structural design. The vendor has confused the coil (which is a pressure boundary at 1,440 psig) with the shell (atmospheric tank). The pressure-vacuum vent on the fill connection is the protection
D Accept but require 100% RT on all welds
Q7.02 — Per API 12K Annex E.9, what mechanical issue must be addressed for the firetube specifically?
A Thermal expansion at the burner end
B Vibration from natural draft chimney effect
C Sulfur corrosion of the stack inlet
D Buoyancy — when immersed in bath water (62.4 lb/ft³ × tube displaced volume), the firetube wants to float and must be mechanically restrained
08

The Calculation Report.
Submit to Client.

You’ve worked through eight design steps, six formulas, and twelve knowledge checks. The output below is your client-deliverable calculation report. Review it, export to PDF, attach to your IFR transmittal.

Document No. WSS-IWBH-CALC-001 · Rev 1 · Issued For Review

IWBH Design Report.

Indirect Water Bath Heater · API 12K Compliant Calculation
Document No.
WSS-IWBH-001
Revision
Rev 1
Status
IFR
Heater Tag
H-101
Date
—
Score
—/12
0/12
Knowledge Check Score
—
Performance Grade

1.0 — Process Basis

2.0 — Heat Duty Calculation

3.0 — Coil Pressure Design

4.0 — Coil Heat Transfer Area

5.0 — Firetube Sizing

6.0 — Shell Structural

7.0 — Standards Compliance Statement

API 12K
Specification for Indirect Type Oil-Field Heaters — primary specification, all clauses
ASME B31.3
Process Piping Code — coil wall thickness per API 12K Cl. 5.1.1 Eq. (1) & (2)
AISC Manual
Steel Construction Manual — shell structural allowable stresses per Cl. 5.2.4
NIGC IGS-M-PM-104(2)
Indirect Water Bath Heater specification — referenced for prescriptive overlay items (expansion tank, instruments, NDE)
NACE MR0175
If H₂S service — material hardness control per API 12K Cl. 4.1
API 521
Pressure-relieving Systems — bath thermal relief PSV sizing (open item, not in scope of this calc)

8.0 — Open Items for Detailed Engineering

  • Burner fuel gas consumption & orifice sizing (vendor scope, App. C)
  • Bath thermal relief PSV sizing per API 521 (fire case + thermal expansion)
  • Stack height per Cl. 5.6 — depends on site elevation and ambient draft
  • Glycol-water bath fraction (NIGC Annex I) for freeze protection at 40°F design low
  • NACE MR0175 confirmation if H₂S present
  • Hydrostatic test of coil to 1.5 × Pmwp per Cl. 6.2.9
  • NDE plan: 100% RT all coil butt welds (NIGC Cl. 4.4.12), 100% T-joint RT shell welds (NIGC Cl. 4.1.8)
  • Vendor general arrangement & performance datasheet review
Workbook complete · 8 / 8
Design Hub India
Open courseware in oil & gas process engineering. Built by working engineers. Anchored to the codes that govern the field.

Series Modules

  • 01 · Two-Phase Separator
  • 02 · Three-Phase Separator
  • 03 · Sand Management Cyclone
  • 04 · IWBH
  • 05 · Pressure Vessels (UG-27)
  • 06 · PSV Sizing (API 520/521)

Codes Referenced

  • API 12K · 12J · 6A · 521
  • ASME VIII Div.1 · B31.3 · IX
  • NACE MR0175
  • NIGC IGS-M-PM-104
  • AISC Manual
  • API RP 14E · 14C
© 2026 Design Hub India · Educational use
v1.0 · DOC-DHI-PD-04 · IWBH-Designer

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top