UCS-66 and MDMT: Impact Test Exemption, Step by Step

MDMT, minimum design metal temperature, is the coldest temperature at which a vessel may see significant pressure. It exists for a specific physical reason: carbon steel gets brittle when it is cold. A material that behaves ductile and forgiving at ambient temperature can fracture suddenly, with almost no warning, once it drops below a certain temperature threshold. UCS-66 is the procedure that decides whether your material needs Charpy impact testing to prove it has enough toughness at that cold temperature, or whether it is exempt from testing altogether.

The procedure, in brief, runs through a short sequence of steps. You assign the material to a toughness curve, A through D, based on its specification and heat treatment. You read the exemption temperature for the governing thickness off Fig. UCS-66. Then, if your part is stressed below its full allowable, you take a warming credit per Fig. UCS-66.1 that can push the exemption temperature low enough to eliminate the impact test entirely, even on a vessel headed for a genuinely cold service.

Most routine vessels are designed with this exemption path in mind from the start, because an unplanned impact test is not a small thing. It costs schedule, it costs money, and worse, it often surfaces late, after material has already been ordered and fabrication has already begun. Understanding UCS-66 well enough to plan around it, rather than discover it, is one of the more valuable skills in vessel design.

The reason all of this matters comes down to how carbon steel actually fails. A ductile material absorbs energy before it breaks, deforming and stretching, giving plenty of warning before final fracture. A brittle material does neither. It can crack suddenly, propagate that crack at close to the speed of sound through the material, and fail catastrophically with almost no visible warning beforehand. Every carbon and low-alloy steel has a transition temperature range where it shifts from one behavior to the other, and MDMT exists to keep the vessel operating safely above that transition, with enough margin that a cold snap, a startup condition, or an equipment upset does not accidentally push the material into brittle territory. UCS-66 is simply the Code's way of proving, on paper, that the material and thickness combination you chose sits safely on the ductile side of that line at the temperature you actually need.

Step 1, assign the material to a curve

Fig. UCS-66's general notes assign materials to Curves A through D by specification and heat treatment, running from least tough, Curve A, to toughest, Curve D. The classic example that shows up in almost every low-temperature job: SA-516-70 lands on Curve B as-rolled, but improves to Curve D once it is normalized.

That jump from Curve B to Curve D is the same chemistry, essentially the same plate, but a dramatically better exemption temperature simply because of how it was heat treated. This is exactly why "SA-516-70N," the N standing for normalized, appears on so many low-temperature jobs. The normalizing heat treatment is not free, and it adds cost and lead time to the plate order, but on a cold-service vessel it can be the difference between passing the exemption cleanly and needing an impact test program that adds weeks to the schedule.

Step 2, determine the governing thickness

Exemption temperature is read at the governing thickness per UCS-66(a). For butt joints, this is generally the nominal thickness of the thickest welded joint at the component, with additional rules covering corner joints, nonwelded parts, and joints between dissimilar thicknesses.

Getting this step wrong is the most common UCS-66 error engineers make, and it is worth slowing down for. The governing thickness at a nozzle-to-shell junction is not automatically the shell thickness, even though that is the intuitive first guess. Depending on the joint configuration, the nozzle wall, a reinforcing pad, or the weld itself can govern instead. Picking the wrong thickness at this step quietly changes every result that follows, since the exemption temperature you read off the chart depends directly on which thickness you plugged in.

Step 3, read the exemption temperature

Enter Fig. UCS-66, or its tabular equivalent, with the curve assignment and the governing thickness. Out comes the temperature at or above which no impact testing is required for that component. The relationship runs in two clear directions: thicker material is worse, meaning it needs a warmer exemption temperature, and a better curve buys you a colder exemption at the same thickness.

A half-inch Curve B component reads in the vicinity of minus 7°F, while the same half-inch thickness on Curve D reads tens of degrees colder. That gap between curves at the same thickness is exactly the value a normalizing heat treatment buys you, and it is often large enough on its own to clear an exemption that would otherwise fail.

Step 4, take the stress-ratio credit (Fig. UCS-66.1)

Here is the step that does the most practical work in real vessel design. If the component's governing stress is below its full allowable, because the plate is thicker than strictly required, or because the low-MDMT condition happens to coincide with reduced pressure, UCS-66(b) lets you reduce the exemption temperature further using Fig. UCS-66.1.

The credit is driven by the ratio of required strength to available strength, commonly expressed as $t_r E^* / (t_n - c)$. At low ratios, meaning the component is carrying only a small fraction of what it is actually capable of carrying, the credit is large. Below a certain threshold ratio, the part becomes exempt to very low temperatures outright, regardless of its base curve.

This is the step that saves vessels that would otherwise need testing. A shell ordered one plate size up, purely because that is what the mill had in stock, frequently picks up enough ratio credit on its own to clear a minus 20°F MDMT without any impact testing at all. It is worth checking this credit before assuming a marginal exemption result is final, because the difference between failing and passing the exemption is often sitting unclaimed in a stress ratio nobody calculated.

Step 5, check the blanket exemptions

Two more doors exist before you need to order Charpy specimens. UG-20(f) exempts certain thin, common-material vessels that meet a specific set of conditions, independent of the curve-and-thickness method entirely. UCS-68(c) offers a further credit when postweld heat treatment is performed even though it was not otherwise required by the design. Both are worth checking on any job that lands close to the exemption line, since either one can tip a marginal result the right way without changing the material or the thickness at all.

A walk-through, in words

Take the vessel from the UG-37 worked example: SA-516-70 as-rolled, Curve B, shell 0.500 in nominal, required thickness 0.227 in corroded, MDMT specified at minus 20°F at 150 psi.

1. Curve B, governing thickness 0.500 in, gives an exemption temperature near minus 7°F. The specified MDMT of minus 20°F is colder than that, so the component is not exempt yet on curve and thickness alone.
2. Stress ratio works out to roughly 0.227 divided by 0.375, or about 0.61, which reads off Fig. UCS-66.1 as a credit of roughly 39°F.
3. Applying that credit moves the adjusted exemption temperature to roughly minus 46°F, comfortably colder than the specified minus 20°F. The component is now exempt, and no impact test is required.
4. This same sequence repeats for every component and every weld joint on the vessel, not just the shell. The vessel's overall MDMT is set by whichever component comes out warmest after its own credit is applied. One thick forging, checked last and found to be the weak point, can govern the whole nameplate even after every other component sailed through its exemption.

Working through this example makes something clear that is easy to miss in the abstract description of the method: the exemption is not a single pass or fail applied to the whole vessel. It is a component-by-component result, and the vessel only earns its stated MDMT once every single component has individually cleared its own exemption at that temperature.

What are the common pitfalls?

• Stopping at the shell. Every component, heads, nozzles, flanges, pads, gets its own UCS-66 evaluation, and the warmest result among all of them governs the nameplate MDMT. Checking only the shell and assuming everything else follows is a frequent and avoidable mistake.
• Standard flanges. B16.5 flanges carry their own exemption treatment entirely separate from the plate curves, so do not read a flange's exemption temperature off the same Fig. UCS-66 chart used for shell and head material.
• Wrong governing thickness at joints. At welded junctions, the rules look at the joint configuration itself, not simply the thickness of whichever part seems most obvious.
• Forgetting the coincident pressure. MDMT is a temperature at a specific pressure, not a temperature in isolation. The ratio credit exists precisely because low-temperature excursions in real operations often happen at reduced pressure, such as during startup or an upset, and the Code lets that reduced stress work in your favor.
• When exemption fails. The remaining options are impact-tested material per UG-84, moving to a better curve through normalizing, picking up more thickness ratio credit, or claiming PWHT credit. These are usually listed here in the order that is cheapest to evaluate first, not necessarily the order that is cheapest to actually buy, since a normalizing heat treatment or an impact test program each carry real cost once you commit to them.

FAQ

What exactly goes on the nameplate? MDMT at its coincident pressure, stamped alongside MAWP. See What Is MAWP for how that companion number is established and why the two are always read together.

Is UCS-66 used for stainless vessels? No. UCS-66 is the carbon and low-alloy steel procedure specifically. High-alloy materials, including most stainless grades, follow UHA-51's own separate toughness rules instead.

Can MDMT be colder than the exemption temperature without testing? Only through the credits described above: the Fig. UCS-66.1 ratio reduction, the PWHT credit, or one of the blanket exemptions. Outside of those paths, the material gets impact tested per UG-84, with no shortcut around it.

Who checks all this? Your design checker reviews it first, then the Authorized Inspector checks it again against the calc package, as stage 6 of the full workflow. Both reviews matter, since a curve assignment or governing thickness error caught early is a redline, while the same error caught at final AI review can mean reworking a calc package that was otherwise complete.

Does a colder MDMT always mean a more expensive vessel? Often, but not automatically. A colder MDMT pushes the required exemption temperature colder too, which can force a better curve, a thicker plate for stress-ratio credit, or an impact test program, all of which add cost. But because the stress-ratio credit rewards components that are already carrying less than their full allowable, a vessel that was going to be built slightly heavier anyway, for corrosion allowance or standard plate thickness reasons, can sometimes clear a genuinely cold MDMT without any extra cost at all. This is why checking the credit early, before assuming a cold service automatically means an impact test, is worth the extra few minutes of calculation.

---

DeepMechanix evaluates UCS-66 per component with the ratio credit applied and cited. See the product.