News

Jun 21, 2023

Causes and cures for ductile spool component failures

When a chemical processer or an oil refiner works through the long, arduous

When a chemical processer or an oil refiner works through the long, arduous process of specifying and overseeing the construction of a processing plant or a refinery, the process is a bit like putting together a million-piece jigsaw puzzle. Every length of pipe, every valve, every flange, and every spool has a specific job to do—often under challenging temperatures and pressures—so that every other part can do its job and the plant can work as intended.

Occasionally, things go wrong. Whether a failure is minor, major, or catastrophic, an investigation ensues to determine the cause. Was the correct item installed? Was it installed correctly? If so, the next order of business is to evaluate the item. Why did it fail?

It's not just a matter of replacing a part. A catastrophic failure can result in an injury or even loss of life. Even in a case in which no injury occurred, the next consideration is downtime. Regardless of the plant's size or production capacity, a system that is shut down to deal with a failure doesn't produce a penny until it's up and running again.

Some users of carbon steels in the North American oil and gas industry have had to deal with such failures. Some spool components approved for use at temperatures as low as -20 degrees F (-29 degress C) have failed because of brittle fracture. Often the failures occurred during hydrostatic testing, cold startups, and sometimes during upset operational conditions. Regardless of when, the subsequent question is always the same: Why?

The question is one of ductility versus brittleness. Located along a single continuum, ductility refers to a material's ability to deform under tensile stress (its ability to stretch without breaking), whereas brittleness is its inability to do so. As a material's ductility increases, its likelihood to resist brittle fracturing decreases.

Components made from carbon steel—any ferrous material having 0.29 to 0.54 percent carbon and 0.60 to 1.65 percent manganese—are considered by ASME VIII Div. I and ASME B31.3 codes as inherently ductile and therefore resistant to brittle fracture. These include A105N flanges; A234 grades WPA, WPB, and WPC seamless fittings; A106N pipe (all grades); and A53 seamless pipe. However, some components rated for service down to -20 degrees F (-29 degrees C) have been found unsuitable for such applications. Some flanges made from A105 carbon steel, operating at less than 300 pounds per square inch (PSI), and some pipe made from A106 grade B, less than ½ in. thick, have been evaluated with a Charpy V-notch impact toughness test and found unfit for service at any minimum design metal temperature lower than 68 degrees F.

Failure investigations conducted by the Belgian Welding Institute indicated that some flanges exhibited large grain size. Further investigation found significant microstructural variation within one specific flange, indicating not just a lack of manufacturing consistency but also deficient heat treatment. Furthermore, a failure analysis conducted on an A350LF2 weld neck flange revealed that poor normalizing practice was a major contributor to the failure. Worse, the test report data listed in the accompanying certificate, EN 10204: 3.1.B, did not match the flange's tested characteristics.

In effect, although these components were within the range specified for chemical composition and mechanical properties and therefore considered ductile, they were susceptible to brittle fracture. Known to result in sudden, catastrophic failure, brittle fracture in newly procured pipe spool components is a potential hazard to equipment integrity, reliability, and process safety.

The Alberta Safety Authority, the agency that oversees pressure equipment safety in Alberta, Canada, issued an advisory in its information bulletin IB16-018: "This may be a concern since flanges made of SA-105 material are commonly exempted from impact testing per ASME Section VIII, Division 1 paragraphs UG-20(f), UCS-66, or ASME B31.3 paragraph 323 for temperature -29°C (-20°F) and greater."

In a blog post titled "Materials Degradation and Corrosion," Charles Becht of Becht Engineering stated, "All ASME B31.3 Figure 323.2.2A Curve B materials are considered to be potentially at risk, although the issue has not been found in pipe manufactured from plate material."

Figure 1 The presence or absence of boron has a considerable effect on austenitic grain size after heat-treating three alloys. Sample A was alloyed with protected boron, sample B was alloyed with unprotected boron, and sample C had no boron.These images originally appeared in European Commission: Technical Steel Research, "Physical metallurgy and the new generic steel grades: Optimisation of the influence of boron on the properties of steel" (Luxembourg: Office for Official Publications of the European Communities, 2007), p. 20.

Investigations revealed that brittle transgranular cleavage cracks were caused by two failure mechanisms—chemistry and poor heat-treatment (normalizing) practices. Both practices, likely originating from cost-cutting efforts of some steel manufacturers, have led to modification of steel chemistry and large, coarse-grain, ferritic-pearlitic microstructures.

The key element is manganese. Manganese promotes finer grain sizes, either as-rolled or normalized. As the grain size decreases (whether it concerns ferrite, bainite, or pearlite), yield strength increases and impact properties improve. An additional benefit is increased pearlite content.

In an attempt to offset the high cost of low-carbon ferromanganese required in the steel manufacturing process, some steel producers intentionally reduced the manganese content to meet the absolute minimum percentage requirement as specified by ASTM.

This isn't inherently wrong, but it can upset the balance between manganese and carbon. If manganese were reduced to the minimum, the manganese-to-carbon ratio would be 0.6 to 0.29, or 2.1 to 1; if they were at their maximums, the ratio would be 1.65 to 0.54, or 3.1 to 1. If both elements were in the middle of the allowable ranges, the ratio would be 2.7 to 1. By reducing the manganese content down to the minimum percentage allowed, it's possible (although unlikely) to come up with a ratio of 1.65 percent manganese content to 0.29 percent carbon content, or 1.1 to 1. Trouble can start when this ratio is anything smaller than 5 to 1.

When the ratio is less than this threshold, the material is known to have poor low-temperature impact properties. Some of the failed steels had ratios as low as 1.8 to 1, resulting in poor toughness, which in turn led to failures during hydrostatic testing or upset operational conditions.

Microalloying—the practice of adding elements in tiny amounts—has been deployed extensively in the manufacture of low-carbon steels to increase the steel's strength, while grain size refinement techniques have been used to increase impact toughness. Elements added to carbon steels for the purpose of microalloying include titanium, vanadium, niobium, and boron. One example is ferroboron, a low-cost alloying agent compared to expensive proprietary alloy mixtures; it is designed to ensure required strengths and consistent results in the steels being processed.

Boron, in soluble form, has been found effective in strengthening the steel, uniformly and consistently, only when balanced with strong nitride and carbide formers such as titanium and niobium. This is to prevent formation of boron nitride or Fe23(C, B)6 precipitates.

While the standard boron content is recommended to be 0.0015 to 0.0030 percent, boron has been known to segregate and form localized areas of high concentrations, causing significant variation in grain size and, by extension, low impact toughness properties. As boron content exceeds 0.007 percent, a low-melting-point B-C-Fe eutectic (Fe2B/Fe3C/Fe) forms, resulting in steels with poor room-temperature toughness.

Boron promotes bainite formation. Deliberate additions of unprotected boron to commercially rolled carbon steel rods are known to promote the formation of coarser ferrite grain size (see Figure 1).

Austenitizing temperature increases are known to cause grain coarsening in boron-treated carbon steels. The amount of grain coarsening depends on two factors: the total boron content of the steel and the amount of boron present in the steel after any free nitrogen has been tied up as boron nitride precipitates. Therefore, the nitrogen content of the steel controls the amount of nitride formation and, in effect the degree of grain pinning, irrespective of the microalloy (titanium, niobium, or aluminum) nitride that is formed.

Titanium, vanadium, and niobium are grain refiners and aggressive oxygen scavengers, while vanadium, niobium, and aluminum also work as nitride formers. Nitrides or carbide precipitates in the matrix can result in a fine, ferrite-pearlite microstructure or may transform to bainite.

The drawback is the lack of research relative to carbon steels. Microalloying with titanium, vanadium, niobium, and boron is common in processing low-carbon and alloy steels with reliable results. This practice seems to have been applied to carbon steels to improve their mechanical properties, as required by ASTM and ASME standards, without proper study. Microalloying has been carried out without clean steelmaking practices (such as argon oxygen decarburization or vacuum oxygen decarburization) and without testing the final product adequately.

In summary, microalloying of carbon steels therefore calls for a very tight material balance between nitride formers and boron and a processing route (heat treatment) customized to that microalloying formulation.

In low-carbon steels (carbon content less than 0.06 percent), thermomechanical processing uses controlled rolling followed by accelerated cooling to produce high-strength, tough microalloyed steels. During thermomechanical controlled rolling, the mechanical properties depend on the applied deformation and coiling temperatures. Boron is known to increase the thermomechanical processing window significantly.

According to Walter J. Sperko, as noted in the "National Certified Pipe Welding Bureau Technical Bulletin May 2016," boron is known to cause directional recrystallization on the 100-crystal plane at 30 to 45 degrees to the pipe axis, resulting in very low toughness at 45 degrees to the pipe or fitting axis. This is the direction in which maximum shear loading occurs for pipe under pressure, and because of this crystal orientation alignment, axial or circumferential impact specimens will not identify the material as having low toughness.

Alloy steels are normally subjected to a full austenitizing quench-and-temper heat treatment to develop mechanical strengths. In the case of the failed flanges, it therefore appears that in an attempt to cut manufacturing costs, the process heat generated during forging was substituted for the required additional heating-quenching-tempering cycle.

Furthermore, ANSI B16.5 does not mandate heat treatment for ASTM A105 flanges below Class 300.

While the presence of more than 0.05 percent titanium in low-carbon steels containing boron has been found to lead to deteriorating toughness, variations in the welding metal's titanium and boron content have been known to cause a wide variation in the weld metal microstructure.

The National Association of Corrosion Engineers standard SP0472, Methods and Controls to Prevent In-Service Environmental Cracking of Carbon Steel Weldments in Corrosive Petroleum Refining Environments, cautions that welds of P1 materials—carbon steels—manufactured with the deliberate additions of microalloying elements such as titanium, vanadium, niobium, and boron may require additional preheat and higher postweld heat treatment temperatures to obtain the mandated hardness in the heat-affected zone. However, the heat treatment could adversely affect toughness values.

All carbon steel piping spool components used for transporting hazardous materials at temperatures less than 32 degrees F should be evaluated. This evaluation includes a preliminary fitness-for-service survey, to include positive material identification, hardness test, and in-situ metallography to determine if it can be put back into service or if it must be replaced. Note that X-ray fluorescence isn't adequate for determining the content of carbon and light elements such as boron, so the PMI needs to be carried out by an optical emission spectrometer.

As a preventive action to reduce the future potential for brittle fracture in carbon steel, fabricators should procure impact-tested carbon steel piping and fittings from accredited sources only. The material must be accompanied with a mill test report certified to at least EN10204, 3.1B specification, which identifies its heat and batch number, and the material standard (ASTM/ASME) to which it is manufactured.

A lesson learned from investigations conducted by the Belgian Welding institute is that to address the possibility of bogus test certification, the purchaser should insist on full manufacturing traceability to the heat number and impact test results from the original steel mill.

Additional requirements to be incorporated into the procurement process may include:

Naddir M. Patel, P. Eng., is a materials and metallurgical engineer for Sinclair Oil Corp., 100 Lincoln Ave., Sinclair, WY 82334, [email protected].

References available upon request.