Here's Why Weld Cracking is So Crucial

WHAT happens when a weld cracks? You fix it, right? Well, it's not that simple.

Jon S Lee, manager of welding materials and procedures for Houston-based Chicago Bridge & Iron Co, laid it all out in his presentation, “Why Weld Cracking is Such An Important Issue,” which kicked off the American Welding Society-sponsored weld cracking conference in Houston June 7-8.

Lee said the cost of a weld crack whether it's related to fabrication/erection or performance/reliability can be enormous.

Not just in the money spent on repairs, which include grinding the crack out, re-welding, performing re-heat treatment, and examining the work.

Not just in delivery penalties.

Not just in the reputation of the organization, which is bound to suffer when the word gets out which definitely will happen, Lee said, leading to a potential decline in future orders.

All those things can mean a cost of thousands of dollars even millions.

But what if there is injury or loss of life?

“You can't put a dollar figure on it,” Lee said, “but it's totally unacceptable.”

Lee said cracking associated with welding can be located in the weld metal, heat-affected zone, or base metal, and it can happen during, after, or even because of metallurgic conditions due to stress at a later time in service.

You Need Knowledge

He said it normally can be eliminated or minimized greatly through knowledge.

“You need a knowledge of base metal and the weld metal alloy system,” he said. “You need to know the types of cracking that can occur solidification cracking of the weld metal, delayed cracking in the weld metal or heat-affected zone, or other types along with the causes of cracking and the cures.

“Welding consumable procurement means putting limits on that so you get the right material and limit the chemistry or other issues so you have an alloy that is less susceptible to cracking problems. There are welding procedural controls: preheat and interpass temperature, welding heat input, interpass cleaning with some of the higher alloys where impurities could cause a cracking problem, and cracking tests for procedure adjustment.”

He gave two examples of prevention:

• Procurement of austenitic stainless steel electrodes E308, E316, E309, ER308 with a minimum amount of delta ferrite (a ferrite number typically in the 4-10% range) to help prevent hot short cracking in the deposited weld metal.

  “Most reputable manufacturers have some of it in there anyway,” he said, “but there are special situations where you're welding carbon steel to stainless steel and you may need more than what's normal. You may want to specify that to get a high enough ferrite content. I've heard many instances of users having that problem because of low ferrite.”

• Welding consumable handling, care, and storage to minimize moisture pickup to reduce the risk of hydrogen cracking in higher strength steels.

A Tank-Fabrication Example

He gave an example of identifying cracking in the tank-fabrication process:

• Cracking in an E6010 fillet weld metal in the repair of a tank (A36 base material) that has been in sulfur service.

“What happens there is you weld the electrodes and immediately upon completion of welding, you see cracking with branching down the center of the weld particularly when you're making a fillet weld,” he said. “Obviously, sulfur is a low-melting-point element.

“You identify cracking as solidification cracking probably due to the low-melting element sulfur and complete additional cleaning or use an electrode (E7018) that is more tolerant of low-melting-point impurities, particularly with manganese. To my knowledge, it ties up with sulfur and keeps you from having a solidification crack.”

He said an inspection and examination is necessary to understand and determine the most cost-effective Non-Destructive Examination (NDE) to find and identify cracks that may occur. The dependent items are: base metal/welding consumable and welding process used; joint configuration; surface or subsurface cracking; orientation of cracking (perpendicular or parallel); and amount of area requiring examination.

He said that understanding the best times to apply the NDE depends on the alloy system.

“Obviously, you always want to do the examination early,” he said. “In our business (at Chicago Bridge & Iron), we always want to apply our NDE as soon after welding as possible. We're looking for indicators early so we can solve the problem.”

As an example of inspection, he cited the examination of weld cracking visible on the surface of fillet welds completed with ERNiCr-3 filler metal with the submerged arc welding process joining A516-70 carbon steel to a nickel-base alloy. He said the preferred method is the liquid penetrant examination because the weld metal and nickel base material are nonmagnetic. Then it should be reviewed with a welding engineer for adjustment of welding parameters to decrease the amount of carbon steel dilution.

“Individuals and organizations must continue to update their knowledge in the causes and prevention of weld cracking,” he said. “Continued research and development is needed to further understand the causes and cures of cracking in welds.”

Shot Peening Can Improve Resistance to Fatigue

AS components fatigue, the life of weldments is directly affected by the residual and applied stresses, which can work together to cause a component and weldment to fail prematurely, leading to catastrophic failure.

That failure can be due to fatigue or to stress corrosion cracking, and if the load exerted is dynamic, it's simply called corrosion fatigue.

What can be done about it? Controlled Shot Peening. It's a way to improve resistance to fatigue, stress corrosion cracking, or corrosion fatigue.

“Shot peening has been around for 50 or 60 years,” said Terry Tate, technical services manager for the Houston, Texas-based Metal Improvement Co Inc. “From the car you drive to the airplane you fly in, you can rest assured there's a lot of parts that have been shot peened. Feel safe. Don't fly in a plane that is not shot peened.”

Shot peening is a cold working process in which the surface of a part is bombarded with small spherical media called shot. Each piece of shot striking the material acts as a tiny peening hammer, imparting to the surface a small indentation or dimple. In order for the dimple to be created, the surface fibers of the material must be yielded in tension. Below the surface, the fibers try to restore the surface to its original shape, thereby producing below the dimple a hemisphere of cold-worked material highly stressed in compression.

Overlapping dimples develop a uniform layer of residual compressive stress in the metal. It is well known that cracks will not initiate or propagate in a compressively stressed zone. Since nearly all fatigue and stress corrosion failures originate at the surface of a part, compressive stresses induced by shot peening provide considerable increases in part life. The maximum compressive residual stress produced at or under the surface of a part by shot peening is at least as great as one half the yield strength of the material being peened. Many materials will also increase in surface hardness due to the cold working effect of shot peening.

Shot Peening is Not Blasting

He said he is continually frustrated by the perception that shot peening is the same thing as blasting.

“Controls are the most important part of shot peening,” he said. “We do not erode the metal. As we impact the metal, we cause dimples. And as we overlap those dimples, we can get a uniform compressive stress. It's essential that we overlap. As we hit the surface, the part yields or stretches. And as the part wants to restore back to its normal size, it goes into compression.

“Blasting is not what we do. We're highly, highly controlled. We work on everything from tools to F-16 Tomcat parts. We're governed and regulated, and we regulate what we do continually.”

Tate said the requirements for stress corrosion cracking are tensile stress at the surface (residual or applied), susceptible metal, corrosive environment, and time lapse.

He said the applied sources of stress are centrifugal loading, cycling, vibration, pressure, thermal cycling, thermal expansion, quenching, bolting, and dead load. The residual sources of stress are welding, shearing, punching, cutting, bending, crimping, riveting, machining (lathe/mill/drill), heat-treating, EDM, laser/wire cutting, and grinding.

He said shot fit is very important finding the right size to meet the geometry of the part.

“If we don't get to the bottom of it, we're not going to get the compressive layer,” he said. “Then there's a place where stress corrosion can happen. It's simple, but sometimes overlooked.”

Shot peening media are controlled and classified. Acceptable shapes are not necessarily spheres, but all the corners are rounded. Marginal shapes include shots that are nodulated, hollow, and elongated (diameter to length ratio greater than 1:2). Unacceptable shapes have broken or sharp corners.

He said the PEENSCAN process is used to measure uniformity and extent of coverage on a shot peened part and is approved by many aerospace, automotive, and industrial manufacturers.

DYESCAN fluorescent tracer liquids, used in the PEENSCAN process, are brushed, sprayed, or dipped on to a part and allowed to dry. This forms a fluorescent elastic coating, which is removed at a rate proportional to the percentage of shot peening coverage. With examination under UV (black) light, the PEENSCAN process provides a practical method of determining peening coverage in terms of the amount and uniformity of fluorescent tracer removal. The PEENSCAN process has been found to be clearly superior to inspection using a 10x glass for determination of peening coverage.

The tracer liquid coating responds to all intensity ranges. Low angle of shot impingement or low shot peening coverage will not remove all of the tracer coating. Uneven peening and hot-spot concentration can be seen easily by patterns on the remaining coating.

“These controls tell us that when we do our process, it's repeatable, time and time again,” he said. “We can repeat that process, and you can be certain you're getting the depth of compression consistently. But it's all about controls.”

How To Deal with Hydrogen-Assisted Cracking

ARC welding of higher-strength structural steels frequently produces hydrogen-assisted cracking (HAC), or cold cracking, in the heat-affected zones and even in the weld metal.

How should this be dealt with?

Paul Konkol, principal engineer for Concurrent Technologies Corp in Johnstown, Pennsylvania, said HAC can be minimized by proper selection of steel, the use of low-hydrogen welding consumables and welding procedures, and weld preheat. Because preheat can increase costs and reduce productivity, new steels and welding consumables are being developed that require little or no preheat.

He said there are four factors affecting HAC, and all must be present together:

• Stress. This involves joint restraint, stress concentrator, and weld metal yield strength.

• Hydrogen. Because of moisture in the electrode coating and flux, “it's very important to maintain your welding rods and have them stored properly.” Lubricants on the wire (oil, drawing compounds, and rust) and paint contaminants need to be reduced.

• Microstructure. It's affected by chemical composition (“high carbon, high-alloy steels are more likely to have cold cracking”) and weld cooling rate.

• Low temperature. HAC occurs below 200° F.

Diffusible Hydrogen Test

Konkol said it's very difficult to measure hydrogen in steel. In order to determine how much hydrogen a particular electrode will produce, the American Welding Society has developed a standard technique called “The AWS Diffusible Hydrogen Test.” It involves welding a 6" test specimen, taking it out of the jig and putting it in ice water, cleaning it, and storing it. Then it is put in a mass spectrometer and heated to 200° F.

Konkol listed the beneficial effects of preheat:

• Reduced weld cooling rate. There is lower hardness and increased time for hydrogen diffusion. “Hydrogen is very mobile and goes up exponentially as the temperature increases,” he said. “So the more preheat you have, the faster the hydrogen moves away from the weld metal.”

• Reduced residual stress.

The beneficial effects of postheat:

• Reduced weld cooling rate of last pass.

• Maintains weldment HAC (cracking) temperature of 200° F.

Among the methods to determine the required preheat: weldability tests; steel producer recommendations on thickness; fabrication code tables relating to thickness and energy input; The Welding Institute's guidelines on thickness and energy input; AWS D1.1 Structural Welding Code, relating to thickness Pcm range, hydrogen range, and restraint range; and the Yurioka Chart Method on thickness, energy input, CEN, Pcm, CE, hydrogen, yield strength, and joint type.

The Welding Institute of Canada Test is designed to simulate restraint conditions, using two plates of steel attached to a strong back, with a simple weld put in a 60-degree groove and then inspected.

He said existing tests and methods may provide incorrect preheat temperatures. Correlation with actual weldments made under worse-case conditions is recommended.

High-strength steels and electrodes have been or are being developed for welding at ambient temperature without HAC.

‘Solidification Cracking’ a Complex Interaction

THE foundry industry traditionally has called them “hot tears.” Welders traditionally have called them “hot cracks.”

The best term is really “solidification cracks,” according to Dr Carl E Cross, a professor in the University of Montana's metallurgical engineering department.

That's because they occur in the weld metal during solidification specifically in the mushy zone, where both solid and liquid trail the weld pool. They normally are found at the grain boundaries, and they start and grow along the weld centerline, involving low-melting eutectic liquid films and the rupture of liquid film between grain metals. They are sensitive to alloy composition and the weld thermal cycle.

“If you're not sure you have one, look at it under high magnification on a scanning electronic microscope,” he said. “It looks like the fracture surface has been coated with a liquid. And that's exactly what it was.”

It is a complex interaction between thermal, mechanical, and alloy effects. The welding parameters are going to affect the strain and stress around the core, and alloy is going to affect the mechanical properties.

“As your weld pool is moving along, the region ahead of the pool is heating up and expanding,” Cross said. “The material around it keeps it from expanding, so it goes into compression. You get cells around the weld pool and as they're cooling down, they're trying to shrink, and again the material is preventing it from shrinking. So this material is going to stay in tension.

“Now you're fixturing, and your welding parameters affect the size and magnitude of the strain cells around the moving weld pool. On top of this, you have this two-phase region behind the weld pool that is solidifying. As you go from solid to liquid, you have a solidification shrinkage, which also tends to put that region in a state of tension.”

What does it all mean? Cross said if you look at a two-phased mushy zone, two solid grains are being pulled apart, and there's a liquid film between them. If those grains are not fed with liquid from the weld pool, the liquid goes into a state of tension and it becomes unstable. It will either cavitate or produce decohesion of oxides.

According to the Theory of U Feurer (1977), cracking is predicted to occur when the rate of grain separation (ROS) becomes greater than the rate of liquid back-filling (ROF).

Cross said sulfur will form a low-melting utectic with iron. At 512 degrees C with a cooling rate of 100 degrees C, the mushy zone will be 5 mm. “That's a very large mushy zone,” he said. “That's quite surprising that you could have such a large mushy zone.”

He said the acceptable range for sulfur to avoid solidification cracking is between 100 and 400 ppm. Below 100, problems arise with fluid flow and penetration. Above 400, solidification cracking is likely.

“If you're buying cheap steel from somebody and getting very high sulfur content, you're very likely to run into a hot-cracking problem,” he said.

He said a welding flux with a high Basicity Index promotes the removal of sulfur from the weld pool and elimination of hot cracking, because magnesium, manganese, and calcium tend to grab onto sulfur and form sulfites, taking the sulfur into the slag.

Aluminum-copper alloys tend to peak out at cracking (24mm) when their weight is 3% solute.

Cross said the Scheil Equation can be used to predict the amount of eutectic (and hence liquid film thickness) that's going to form. The higher the composition in alloy content, the more eutectic is formed. When the solute content is high, a lot of eutectic is formed, allowing the solidification shrinkage to be fed, preventing cracking.

He said if you're welding aluminum 6082 without filler wire, cracking will probably result. It takes the base metal and dilutes it, making it less crack-sensitive. With an aluminum-zinc-magnesium alloy like 7108 which is close to the top of the crack-susceptibility hill a 5000 filler can be used for dilution. Aluminum-copper-magnesium alloys such as 7075 and 2024 are not very weldable, while 5083 is.

Weldability tests can be used to evaluate the cracking susceptibility of various alloys: inverted T, circular patch, Varestraint, and Trans-Varestraint.

Non-Destructive Evaluations Detect Flaws

INITIAL ultrasonic examination weld scanning is typically generic in nature for good flaw-detection reasons, in that it's designed to optimize the probability of detection.

Sizing, however, usually requires more specialized techniques.

“We want good quality information,” said Ronald W Kruzic, corporate QA and NDE manager for Chicago Bridge & Iron Co in Houston, Texas, “because you're going to ask, ‘Am I going to immediately take my structure out of service and repair that flaw that's been detected? Can I wait a couple of months? Or perhaps the decision can actually be made from what was originally there from initial fabrication so I never have to take care of it, and I'll monitor it during the life of my structure.’”

The selection of a technique is dependent on:

• Flaw location.

• Orientation and size. “Are they large, small, perpendicular to the surface? Are they like a lamination parallel to the surface?”

• Acceptance criteria.

• Materials and types of weldments. “Some weldments are relatively easy to inspect. Others, like your austenitic-type welds, can get to be much more difficult.”

Techniques for Detection

Kruzic focused his discussion on the following techniques: refracted longitudinal waves, manual Time of Flight Defraction (TOFD), semi-automated TOFD, and Alternating Current Field Measurement (ACFM).

Refracted longitudinal waves use dual-element, high-angle longitudinal wave transducers and detect the reflected energy from the tip of the crack. The benefit is that the technique can be used on coarse-grain materials (ie, austenitic welds). The drawbacks are that it only gets one tip, must be repeated from the opposite surface for embedded flaws, and has focused probes, offering only a working limited range.

Manual TOFD is usually restricted to cracks connected to a surface so that a corner trap signal is utilized as well as the tip-diffracted signal.

Semi-automated TOFD uses two probes on opposite sides of the weld with an encoder for positional information. For most accurate data, both parallel (B-scan) and non-parallel (D-scan) scans can be performed.

“If all you're interested in is height and length, typically you can get away with just a D-scan inspection,” Kruzic said.

A B-scan takes the transducers and moves them to the center, giving positional information. Kruzic said the benefit is that it provides a high degree of accuracy, particularly if the purpose is to monitor flaws for growth. The drawback is that if the flaws are close to the plate surfaces, then the lateral wave that represents the top surface and the back-wall reflected signal that represents the bottom surface can interfere with signals and can actually blend together. He said the procedure can be modified, using different frequencies and different spacings to optimize the sensitivity.

ACFM is capable of detecting and sizing flaws open to the surface of the material. It works on both ferrous and nonferrous materials. Programming must be applicable to the material for accurate sizing. Like Magnetic Particle Testing, it requires two scans with the probe rotated 90 degrees.

He said that because it is an electromagnetic non-contacting technique, it can work through coatings. A probe induces an alternating current in the test piece and other sensors measure the resulting electromagnetic field close to the surface. Defects disrupt the flow of the fields around the edges of the flaw, providing information on the flaw's location and length.

Kruzic recommended that in order to achieve accurate, dependable results, the user needs to specify which Non-Destructive Test system is going to be applied not just the procedure, but also the equipment.

“You must decide on how the NDT system is validated, whether it's the manufacturer or subcontractor, owner/user, or third party,” he said. “The more critical the application, the more critical the validation should be.”