Tubesheet Thermal Analysis: From Code Checks to FEA

11 minutes ago

11 min read

Perforated exchanger tubesheet on fabrication supports

Tubesheet thermal analysis means finding out how temperature differences between the tubes, the shell and the plate itself load the tubesheet, then checking that stress against allowable limits. For most exchangers you start with the ASME Part UHX rules. For startups, shutdowns and cycling duty, you add a transient temperature solution and a finite element stress check. This article shows where each method applies, where each one breaks down, and how to build a workflow you can defend in a design review.

TL;DR

Part UHX of ASME Section VIII, Division 1 handles fixed, U-tube and floating-head tubesheets with an equivalent solid plate model. It is the right first pass for steady-state design.
Thermal stress comes from differential expansion and from transient gradients. Fixed tubesheets feel the first most strongly. Every type feels the second during startup and shutdown.
The equivalent plate is least reliable in the transition between the perforated area and the rim. Cycling duty often needs detailed FEA and a Division 2 Part 5 style fatigue and ratcheting check.
Get the temperature field right first. A stress model with the wrong thermal boundary conditions gives precise, wrong answers.
Our free and subscription thermal tools can help you build and sanity-check that temperature field before you move to structural FEA.

What makes a tubesheet thermally critical
Step one: the ASME Part UHX check
Step two: know where the equivalent plate stops being reliable
Step three: transient thermal stress and fatigue
Step four: design by analysis when rules are not enough
A workflow that holds up in review
Modeling choices in the FEA itself
Perspective: what we see go wrong
Where our toolkit fits
Common pitfalls, in short
FAQ
Sources

What makes a tubesheet thermally critical

A tubesheet is a thick perforated plate that separates the tube side from the shell side. It is welded or rolled to hundreds of tubes and, in a fixed design, welded to the shell as well. That makes it a structural link between two parts that rarely sit at the same temperature.

Three things load it at once:

Pressure on the tube side and shell side, which bends the plate.
Differential axial expansion between tubes and shell. In a fixed tubesheet exchanger, the tubes and shell cannot move freely relative to each other, so the mismatch loads the plate and the tube joints. As one 2022 hazard analysis of fixed-tubesheet units notes, different thermal expansions can produce significant thermal stress and strain because the plate is welded to both shell and tube bundle. Treat that as the headline mechanism for fixed designs.
Temperature gradients inside the plate, radial and through the thickness, which develop when hot fluid reaches one face before the other.

The third item is the one people skip. It is easy to assume that U-tube and floating-head units are free of thermal problems because the bundle can grow. A 2017 paper on thermal stress at the tubesheet in floating-head or U-tube exchangers reports that the stress in those designs is still significant. Free growth of the bundle does not remove the gradient through the plate.

Step one: the ASME Part UHX check

ASME Section VIII, Division 1, Part UHX is the standardized framework for fixed, U-tube and floating-head tubesheets. It uses the equivalent solid plate method: the perforated region is treated as a solid plate with effective elastic constants, which makes the stress calculation tractable by hand or in a spreadsheet.

In practice, UHX asks you to run several load cases. You combine tube-side pressure, shell-side pressure and differential thermal expansion, then evaluate each combination, including cases where only one pressure acts. For a fixed tubesheet, the thermal case needs the metal temperatures of the tubes and shell, so you need credible numbers before you start.

Inputs you need before running UHX

Input	Why it matters	Common mistake
Mean metal temperature of tubes	Sets tube free growth	Using fluid inlet temperature instead of mean wall temperature
Mean metal temperature of shell	Sets shell free growth	Ignoring shell-side fouling or insulation effects
Material properties at temperature	Modulus and expansion coefficient both vary	Using room-temperature values
Tubesheet thickness and ligament pattern	Sets effective constants	Carrying a drawing pitch that differs from the as-built layout
Operating, startup and upset scenarios	Each can govern	Checking only normal operation

The temperature rows are the weak link. UHX tells you how to turn temperatures into stress. It does not tell you what those temperatures are. That is a heat transfer problem, and it deserves its own calculation. For a refresher on the method side, see our heat transfer model guide.

Step two: know where the equivalent plate stops being reliable

The equivalent solid plate is a simplification, and its authors never claimed otherwise. Work on homogeneous tubesheet models for liquid-metal reactor exchangers found severe radial thermal and stress gradients in the transition region between the perforations and the rim. In that region, stresses predicted by the equivalent model can deviate by 20% or more from the real behavior, according to practitioner warnings summarized from that work. Treat that figure as an indication of scale, not a universal correction factor, since it depends on geometry and loading.

Thermal gradient across perforated tubesheet rim

That has a practical meaning. If the governing location in your UHX report is the rim or the edge of the perforated zone, and the margin is thin, you should not trust the hand calculation alone. A detailed model with the real hole pattern near the transition is worth building.

A second limit is time. UHX is a steady-state set of load cases. It does not follow a temperature front moving through the plate during a ramp. For that you need a transient thermal solution followed by a stress solution.

Step three: transient thermal stress and fatigue

If your exchanger starts and stops often, or follows a load, transient behavior can govern life. A 2018 ASME PVP paper on fixed tubesheets opens with the point that some fixed-tubesheet exchangers suffer a large number of startups and shutdowns, which cause large thermal stress in the plate. It analyzes how the rate of temperature change drives that stress.

The clearest recent example of cost comes from concentrating solar power. A 2024 study of molten-salt-to-steam exchangers found that typical ramping cycles can cut tubesheet life to roughly 10 years for evaporators and 25 years for superheaters, against a 30-year design target. That finding is specific to those plants and their cycling, but the mechanism applies anywhere the duty cycles hard: the plate sees a thermal gradient each time, and fatigue damage accumulates faster than a steady-state design life assumes.

Reported exchanger life compared with design target

The same research argues that designers can overpredict equipment value when they use a nominal design life and ignore accelerated fatigue from cyclic thermal stress. If you size or price equipment on life, this is a reason to run the transient case before you commit.

We cover the typical modeling errors in more depth in Don’t Overestimate Thermal Fatigue Life: 6 FE Pitfalls for Engineers.

Step four: design by analysis when rules are not enough

When UHX is overly conservative or does not cover the loading, ASME Section VIII, Division 2, Part 5 offers alternative rules. ASME’s course outline for Design by Analysis lists the building blocks: stress definitions, primary, secondary and peak stress limits, and stress classification. The course page states that the focus is the analytical methods in Part 5 and how to meet them using FEA.

Three checks matter most for a thermally loaded tubesheet:

Plastic collapse: can the plate carry the loads without gross yielding?
Ratcheting: does plastic strain accumulate cycle after cycle? Thermal gradients combined with pressure are the classic cause, and standard formulas do not detect it.
Fatigue: how many cycles of the thermal and pressure range can the plate and the tube-to-tubesheet joint take?

A high-density mesh alone does not make the analysis good. Material modeling and load combinations matter as much, because the answer depends on which cycles you include and how the material behaves at each temperature.

If you need a refresher on stress categories before you start, our pressure vessel stress analysis guide walks through formulas and FEA together.

A workflow that holds up in review

Here is the sequence we recommend for a tubesheet where thermal loading is a real concern.

Define the scenarios. List normal operation, startup, shutdown, trips and any upset where one fluid is lost. Give each a temperature-versus-time profile for both fluids.
Estimate metal temperatures. Use convection coefficients for each side, the wall conduction path, and fouling. Check the answer against a simple energy balance.
Run the UHX check. Use the mean metal temperatures to evaluate every required load case. Record which case governs and by how much.
Decide whether a detailed model is justified. Thin margin at the rim, high cycle count, large temperature differences or a non-standard layout all say yes.
Build the transient thermal model. Use the actual fluid temperature histories as boundary conditions. Check peak through-thickness gradient against hand estimates.
Map temperatures into the structural model. Keep the same geometry and, where possible, a compatible mesh.
Evaluate against Part 5 criteria where Division 2 applies, or against the criteria your client or code of construction specifies.
Stress-test the answer. Vary heat transfer coefficients and ramp rates. If life changes by a factor of three when you move a coefficient by 20%, say so in the report.

Getting the thermal boundary conditions right

The temperature field is where most tubesheet analyses go quietly wrong. A few habits help:

Use the metal temperature, not the bulk fluid temperature, for expansion calculations.
Remember that the tubesheet sees both fluids: one on each face. The inlet-side face usually leads the other during a ramp.
Account for the tube-side inlet region. Flow distribution near the channel cover can leave local hot or cold spots. A University of Pretoria master’s thesis on a feedwater heater tubesheet coupled a 1D network solver with CFD to get a plate temperature field without modeling every tube in 3D. That is a sensible approach when full conjugate CFD is too heavy.
Check units. A 60°C rise can produce roughly 144 MPa of constrained thermal stress in steel, as we show in our thermal stress calculation walkthrough. Mixing MPa, psi and kPa in a spreadsheet is a common source of errors.

A classic reference for the procedure is a 1983 thermal stress analysis procedure for heat exchanger tubesheets. It describes an axisymmetric finite element model of the tubesheet and shows how modest the model can be if the temperatures are good.

Modeling choices in the FEA itself

Solid plate or explicit holes

An equivalent solid plate with effective constants is cheap and works well far from the rim. Modeling the actual holes is costly but gives local stress concentration at ligaments. A common compromise is a submodel: use the global solid plate model to find the critical region, then refine it with the real hole pattern in that area only.

Axisymmetric or 3D

Axisymmetric models handle round plates with uniform loading well and run fast. They miss pass partition lanes, nozzle effects and uneven flow. Use 3D when partitions or non-uniform temperatures dominate.

Tubes and joints

Tube-to-tubesheet joint loading adds another layer. If joint integrity is your concern, the tube axial load from the thermal case is the key output of the structural model. Capture the tube stiffness correctly, because it sets how much load the tubes take from the plate.

Geometry capture

When you analyze an existing unit, you often start from a drawing that does not match the as-built condition. A quick scan or model of a nearby component, such as a 3D model of a distribution manifold built from a short video, shows what geometry capture can look like for a part you cannot get drawings for. For a tubesheet, you still need verified dimensions for the hole pattern and thickness.

Perspective: what we see go wrong

We build thermal tools, so we look at tubesheet problems from the temperature side. Our experience is that teams put weeks into mesh refinement of a stress model while the thermal boundary conditions come from a single guessed heat transfer coefficient. The stress result inherits every error in that guess, and it inherits it silently.

Our advice is to do the cheap thermal work first and do it honestly. Estimate conduction through the wall, convection on both sides, and radiation where temperatures are high enough to matter. Make a bounding case with high and low coefficients. If both bounds pass, you can stop. If they straddle your limit, that is where detailed CFD or a coupled model earns its cost.

We also think the line between a code check and a full analysis is drawn too rigidly. UHX is a good tool for what it was built to do. It does not replace judgment about cycling duty, and nobody should read a passing UHX report as a fatigue assessment.

This is why we publish engineering tools with free basic features. A student or a small team should be able to run a defensible first-pass thermal estimate without a license barrier.

— Joel

Where our toolkit fits

The Jewlz Engineering Toolkit is an Excel-based tool that combines conduction, convection and radiation analysis in one place, with variable material properties, a wide temperature range and a built-in property database. For tubesheet work, that is the temperature-field step: estimating metal temperatures, wall gradients and heat transfer for the scenarios you will feed into a stress check. It is not a replacement for a UHX calculation or a structural FEA package, and we would not suggest otherwise.

You can explore the Jewlz Engineering Toolkit and the thermal analysis tools. If you are sizing the exchanger itself, see our heat exchanger optimization software. For the vessel-side context, our pressure vessel thermal analysis workflow covers the FE steps in more detail.

Common pitfalls, in short

Treating a UHX pass as a fatigue pass. UHX is a steady-state design check. Cycling duty needs its own assessment.
Room-temperature properties. Modulus and expansion coefficient change with temperature, and so do allowable stresses.
Ignoring the rim transition. That is where the equivalent plate is weakest.
One scenario only. Startup, shutdown and upset often govern over steady operation.
Skipping validation. Check the FE model against a hand case, such as a simple thick cylinder, before trusting it on the real plate. Our thick wall cylinder temperature guide shows one such check.
Joint efficiency confusion. Weld joint factors apply to shell and head calculations, not to tubesheet ligament efficiency. If that distinction is hazy, read our note on ASME joint efficiency.

FAQ

Do I need FEA for every tubesheet?

No. For many steady-duty exchangers, a Part UHX calculation is the accepted route. FEA becomes worth the effort when margins are thin, when startups and shutdowns are frequent, or when the layout falls outside what UHX covers. Your code of construction and the owner’s specification decide what is required.

Is thermal stress only a problem in fixed tubesheet exchangers?

Fixed designs feel differential expansion most, because tubes and shell cannot move independently. U-tube and floating-head units are not exempt, though. Research on those designs shows tubesheet thermal stress there can still be significant, mainly from temperature gradients in the plate.

What is ratcheting, and why does it matter for tubesheets?

Ratcheting is the progressive accumulation of plastic strain over repeated cycles. It can occur when a steady load combines with a cyclic thermal gradient. Standard formulas will not detect it, so Division 2 Part 5 style analysis is used to check for it.

How accurate is the equivalent solid plate model?

It is good for the bulk of the perforated plate and unreliable near the rim transition. Practitioners report that stress there can differ from model predictions by 20% or more. If your governing location is near that region, add a detailed model with the real hole geometry.

Can I do a first-pass tubesheet temperature estimate in a spreadsheet?

Yes, and it is good practice. Estimate wall temperatures from convection on each side and conduction through the metal, then bound the answer with high and low coefficients. Use the result as input for UHX and to decide whether a transient FE model is needed.