9. Weld Defects, Inspection and Quality Assurance

Introduction and Context

A welded structure is acceptable not because it appears sound, but because it has been produced under controlled conditions, inspected by suitable methods, and shown to satisfy specified service requirements. Weld quality therefore involves more than the detection of visible imperfections. It includes the selection of materials and procedures, control of fabrication variables, competence of personnel, reliability of inspection, and documented evidence that the completed joint is fit for purpose.

Weld defects may reduce strength, toughness, fatigue life, leak tightness, corrosion resistance, or dimensional accuracy. Their significance depends on their type, size, location, orientation, distribution, and the service conditions of the component. A small rounded pore in a static structure may be harmless, whereas a planar lack-of-fusion indication in a pressure vessel can be critical. Quality assurance consequently requires engineering judgement rather than the mechanical rejection of every indication.

This chapter examines weld discontinuities and defects, inspection methods, acceptance criteria, quality-management practices, and qualification documents. It also considers the relationship between welding procedure specifications, procedure qualification records, welder qualification, non-destructive testing, and modern digital inspection systems.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Distinguish between a weld discontinuity, an indication, and a rejectable defect.

2. Explain the mechanisms that produce common weld defects.

3. Select suitable visual, radiographic, ultrasonic, magnetic-particle, penetrant, and advanced inspection methods.

4. Interpret the main factors used in acceptance criteria.

5. Explain the purpose and relationship of WPS, PQR, and welder qualification records.

6. Develop a practical inspection and quality-assurance plan for a welded component.

7. Evaluate inspection limitations, risks, ethical responsibilities, and emerging digital practices.

From Discontinuity to Defect

A discontinuity is an interruption in the normal structure of a weld or base material. It becomes a defect only when it exceeds the limits established by a code, standard, drawing, specification, or engineering assessment. This distinction is essential. Inspection detects indications; acceptance criteria determine whether those indications are permissible.

An inspection indication may be real or false. A real indication corresponds to a physical feature such as porosity, slag, or a crack. A false indication results from surface contamination, poor coupling, equipment malfunction, geometric effects, or incorrect interpretation. The inspector must therefore establish the validity, location, size, and character of an indication before making a disposition.

Acceptance criteria are not universal. They are selected according to joint design, material, loading, failure consequence, fabrication process, and governing standard. Typical criteria address crack prohibition, limits on pore size and density, incomplete fusion, incomplete penetration, undercut, overlap, reinforcement, underfill, misalignment, and dimensional deviation. A sound quality system identifies these criteria before fabrication begins.

How Weld Defects Form

Cracks

Cracks are planar separations and are generally among the most serious weld discontinuities because their sharp tips produce high stress concentration. Hot cracks form during solidification when a weak, liquid-rich film is unable to accommodate contraction. They are promoted by unsuitable composition, excessive restraint, poor bead shape, and high impurity levels. Cold or hydrogen-assisted cracking may occur after cooling, particularly in hardenable steels when diffusible hydrogen, susceptible microstructure, and tensile stress coexist. Crater cracks develop when the weld pool is terminated abruptly without filling the crater.

Crack prevention involves material control, low-hydrogen practice, preheating, suitable heat input, correct joint restraint, proper termination technique, and post-weld heat treatment when specified. Cracks are commonly unacceptable regardless of their apparent size, although the governing code remains controlling.

Porosity

Porosity consists of gas cavities trapped in solidifying weld metal. It may result from moisture, oil, paint, rust, inadequate shielding gas, excessive gas flow, turbulence, long arc length, contaminated filler metal, or poor protection from wind. Porosity can be scattered, clustered, or aligned. The effect depends on its size, distribution, and loading direction. Prevention begins with clean surfaces, dry consumables, correct gas flow, protected work areas, and stable welding parameters.

Slag and Tungsten Inclusions

Slag inclusions occur when non-metallic material becomes trapped between weld passes or between the weld and base metal. They are encouraged by insufficient cleaning, poor manipulation, an unfavorable groove angle, low current, or incorrect travel speed. Tungsten inclusions are associated mainly with gas tungsten arc welding and may result from electrode contact with the molten pool, excessive current, or improper grinding of the electrode. Both types may act as stress raisers, particularly when elongated or aligned.

Lack of Fusion and Incomplete Penetration

Lack of fusion is failure to achieve bonding between weld metal and the base metal or between adjacent weld beads. It may be caused by low heat input, excessive travel speed, incorrect electrode angle, inadequate cleaning, or an unsuitable joint configuration. Incomplete penetration occurs when the weld metal does not extend through the intended root region. Root face thickness, root gap, joint access, polarity, and welding technique strongly influence it.

These planar discontinuities are often more dangerous than rounded porosity because they can provide a crack-like path. Ultrasonic testing is frequently effective for their detection, although radiographic testing may also reveal them depending on orientation and geometry.

Undercut, Overlap, Underfill, and Excess Reinforcement

Undercut is a groove melted into the base metal adjacent to the weld toe or root and left unfilled. Excessive current, high travel speed, poor electrode angle, and unsuitable technique are common causes. Overlap occurs when weld metal flows over the base material without proper fusion. Underfill is a depression below the adjacent surface, while excessive reinforcement produces an abnormally convex weld profile. These conditions may reduce effective section, create stress concentration, interfere with fit-up, or impair fatigue performance.

Distortion and Residual Stress

Uneven heating and cooling produce thermal expansion, contraction, distortion, and residual stress. Angular distortion, longitudinal shrinkage, transverse shrinkage, buckling, and misalignment may affect assembly and service performance. Residual tensile stress can assist crack growth and stress-corrosion damage. Control methods include balanced welding sequences, reduced heat input, restraint planning, presetting, intermittent welding where permissible, and suitable fixturing. Excessive restraint, however, can increase cracking risk and should not be treated as a universal solution.

Inspection as an Engineering Process

Inspection should be planned rather than added after fabrication. A practical inspection sequence normally includes review of material certificates and consumables, joint preparation and fit-up inspection, monitoring of preheat and interpass temperature, observation of welding variables, visual inspection after welding, non-destructive examination, dimensional verification, repair control, and final documentation.

The choice of method depends on the expected discontinuity, component geometry, accessibility, material, surface condition, required sensitivity, cost, and consequences of failure. No single NDT method detects every type of discontinuity. Inspection planning should therefore match the physics of the method to the likely failure mechanism.

Visual Testing

Visual testing is the first and often most productive inspection method. It can identify incorrect weld size, poor profile, undercut, overlap, arc strikes, surface cracks, spatter, distortion, incomplete filling, and dimensional errors. Before welding, it verifies bevel angle, root gap, root face, alignment, cleanliness, tack welds, and consumable identification. During welding, it supports control of interpass cleaning, bead placement, and temperature. After welding, it confirms profile and surface condition.

Visual inspection requires adequate lighting, access, calibrated gauges where appropriate, clean surfaces, and qualified personnel. It cannot reliably detect internal defects and should not be regarded as proof of internal soundness.

Radiographic Testing

Radiographic testing uses penetrating X-rays or gamma radiation to create an image based on differences in attenuation. Volumetric discontinuities such as porosity and slag inclusions are often readily displayed. Digital radiography can replace film and provides image enhancement, electronic storage, and rapid review.

Radiography is less reliable for planar defects aligned parallel to the radiation beam. It also requires radiation protection, controlled access, suitable source positioning, and interpretation by competent personnel. Thickness, geometry, exposure quality, image sensitivity, and scattered radiation influence detectability. Radiographic acceptance should never be inferred solely from image appearance; the applicable code and image-quality requirements must be applied.

Ultrasonic Testing

Ultrasonic testing introduces high-frequency sound into the component and interprets reflected energy. It is especially valuable for detecting planar discontinuities and for examining thick sections. Conventional pulse-echo testing provides information about reflector position and response. Phased-array ultrasonic testing electronically steers and focuses multiple beams, improving coverage and enabling sector scans. Time-of-flight diffraction can be effective for sizing crack-like defects by analyzing diffracted waves from defect tips.

Ultrasonic interpretation is influenced by material structure, surface condition, probe angle, calibration, attenuation, coupling, geometry, and operator skill. Weld anisotropy and complex component shapes can produce misleading indications. Calibration blocks and documented scanning procedures are essential.

Magnetic-Particle Testing

Magnetic-particle testing is applicable to ferromagnetic materials. Magnetization creates leakage fields at surface and near-surface discontinuities; applied magnetic particles accumulate at these locations to form visible or fluorescent indications. The method is highly sensitive to cracks oriented approximately perpendicular to the magnetic field, so two magnetization directions may be necessary.

The surface must be reasonably clean, and the component must be magnetized adequately. The method is not suitable for non-ferromagnetic alloys and has limited capability for deep internal discontinuities. Demagnetization and removal of inspection media may also be required.

Liquid-Penetrant Testing

Liquid-penetrant testing detects surface-breaking discontinuities in both ferrous and non-ferrous, non-porous materials. The surface is cleaned, penetrant is applied, excess penetrant is removed, developer draws penetrant from discontinuities, and the indications are examined under suitable visible or ultraviolet illumination. The method can reveal fine cracks, laps, seams, and pinholes, but it cannot detect subsurface flaws. Cleaning and dwell time are critical, and excessive removal can erase indications.

Selecting an Inspection Method

Method selection can be organized through a decision process. First identify whether the suspected discontinuity is surface-breaking, near-surface, or internal. Next consider whether the material is ferromagnetic, whether access is available from one or both sides, and whether the component geometry affects beam or radiation paths. Finally assess required sensitivity, production rate, safety, cost, and the consequences of missed defects.

For example, penetrant testing is suitable for surface cracks in an aluminium weld, whereas magnetic-particle testing is unsuitable because aluminium is non-ferromagnetic. Magnetic-particle testing is efficient for surface and shallow subsurface cracking in a carbon-steel fillet weld. Ultrasonic testing is often preferred for planar flaws in a thick steel pressure boundary, while radiography may be advantageous when a permanent volumetric image is required.

WPS, PQR, and Welder Qualification

The Welding Procedure Specification (WPS) is the controlled instruction used to produce a weld. It normally identifies the base materials, thickness range, joint design, welding process, filler classification, position, polarity, current or voltage range, travel speed or heat-input limits, shielding gas, preheat, interpass temperature, post-weld heat treatment, and any special technique requirements.

The Procedure Qualification Record (PQR) records the actual variables used to weld and test a qualification coupon. Mechanical tests may include tensile, bend, impact, hardness, or macro-examination, depending on the governing code. The PQR provides evidence that a procedure can produce a joint with required properties. It is not simply a copy of the WPS: the WPS gives permitted production ranges, whereas the PQR records demonstrated values and test results.

Welder or welding operator qualification demonstrates that an individual can deposit sound weld metal using a specified process, position, material range, and joint type. Qualification is personal and does not replace procedure qualification. A qualified welder may still produce unacceptable welds if the WPS, fit-up, consumables, or supervision is inadequate.

A controlled system links the drawing, material certificates, WPS, PQR, welder identification, consumable batch, inspection reports, repair records, and final release documentation. Traceability allows a nonconformance to be investigated without relying on memory or informal statements.

Quality Assurance and Quality Control

Quality assurance is the planned and systematic part of a quality system intended to provide confidence that requirements will be met. Quality control comprises operational techniques used to verify conformity. In welding, assurance includes approved procedures, competent personnel, calibrated equipment, supplier control, document control, audits, and corrective action. Control includes fit-up checks, parameter monitoring, visual examination, NDT, dimensional checks, and repair verification.

A mature system emphasizes prevention. Root-cause analysis should ask why a defect occurred rather than merely identifying the person who made it. Useful causes may include inadequate joint design, unclear instructions, poor consumable storage, unstable power supply, rushed production, insufficient training, or inappropriate acceptance criteria. Corrective action should address the process cause and verify effectiveness.

Acceptance criteria should be agreed before inspection. If a customer, designer, fabricator, and inspector apply different standards after welding, disputes and unsafe concessions may result. Engineering acceptance may sometimes be based on fracture mechanics, fitness-for-service assessment, or structural reliability rather than a simple visual limit; such decisions require authorized engineering evaluation and complete records.

Applied Case: Recurrent Lack of Fusion in a Pressure-Piping Project

Background

This illustrative case concerns a carbon-steel process-piping project involving full-penetration circumferential welds. Radiographic examination showed recurring linear indications near the sidewall in several welds.

Situation and Analysis

The fabricator initially increased welding current, assuming that insufficient heat was the cause. The indications continued. A review showed that the actual root cause was a narrow groove angle combined with poor electrode manipulation and inadequate removal of oxide between passes. The WPS permitted a range of groove dimensions, but the fit-up inspection form did not require measurement of the included angle. In addition, the repair procedure did not require repeat examination of adjacent areas after excavation.

Approach and Solution

The project team first quarantined affected welds and mapped the indications. Ultrasonic examination was used to characterize their planar nature, while visual inspection confirmed restricted sidewall access. The joint design was revised within the qualified range, fit-up gauges were introduced, and welders received focused practice on sidewall fusion. Interpass cleaning became a documented hold point. The WPS was clarified regarding travel speed, electrode angle, and maximum bead size. Repaired welds were excavated under controlled conditions and re-examined using the specified method before acceptance.

Outcome

The recurrence rate fell after the process changes, and the remaining welds were accepted against the project specification. The investigation demonstrated that increasing a single parameter without understanding defect mechanism is an unreliable corrective action. It also showed that inspection data can reveal weaknesses in procedure design and quality planning.

Learning Points

The important lessons are that planar indications require mechanism-based analysis, fit-up is a quality variable, WPS ranges must be practical and controlled, and repair work needs its own inspection and traceability. Quality assurance is strongest when inspection findings are converted into preventive process improvements.

Limitations, Risks, and Common Mistakes

Frequent mistakes include treating visual acceptance as complete inspection, selecting radiography for every defect type, ignoring orientation effects in NDT, using expired or damp electrodes, failing to calibrate equipment, accepting ambiguous indications without resolution, and repairing defects without identifying their cause. Another serious error is confusing a qualified procedure with a qualified welder, or assuming that a welder qualified on one material and position is automatically qualified for all work.

NDT results are probabilistic. Detection depends on equipment, technique, personnel, geometry, and discontinuity characteristics. A report stating “no indications detected” does not mean that no discontinuity exists; it means that none was detected within the capability and scope of the examination. This limitation should be recognized in risk-based inspection and structural integrity decisions.

Professional, Regulatory, and Ethical Considerations

Welding inspectors and engineers have a duty to report findings accurately and independently. Suppressing an indication, altering a report, using an unqualified operator, or accepting a repair without authorization can endanger workers and the public. Records must be secure, legible, traceable, and protected from retrospective manipulation.

Radiographic work requires strict radiation-safety controls. Chemical penetrants, developers, cleaning agents, and welding fumes require appropriate ventilation, personal protection, storage, and disposal. Digital images and automated inspection data should have access control and audit trails. Regulatory compliance is not a substitute for engineering responsibility: codes establish minimum requirements, while the responsible organization must ensure that the completed structure is suitable for its intended service.

Digital Inspection and Emerging Practice

Digital radiography, phased-array ultrasonics, total focusing methods, robotic scanning, three-dimensional surface mapping, and cloud-based inspection records are improving coverage and traceability. Automated image analysis can assist with indication detection and classification, but it should support—not replace—qualified human judgement. Algorithms may be affected by training-data bias, unusual geometries, poor image quality, or software configuration errors.

Data analytics can identify links between defect occurrence and variables such as welder, joint type, consumable batch, heat input, or environmental conditions. When used responsibly, this supports predictive quality control. The essential requirements remain calibration, validation, cybersecurity, competence, and clear accountability for the final decision.

Best-Practice Framework for Weld Quality

A robust welding quality programme can be summarized as follows:

• Define service requirements, codes, acceptance criteria, and inspection stages before fabrication.
• Qualify procedures and personnel appropriate to the actual production range.
• Control materials, consumables, storage, identification, and traceability.
• Inspect joint preparation, fit-up, preheat, interpass temperature, and welding variables.
• Use complementary NDT methods selected according to likely discontinuity and geometry.
• Ensure inspectors, equipment, procedures, and interpretation systems are competent and calibrated.
• Control nonconforming work through approved repair procedures and repeat examination.
• Analyze trends and root causes rather than treating each defect as an isolated event.
• Preserve objective records and release only work that meets defined requirements.

Chapter Summary

Weld quality is the result of controlled design, procedure qualification, competent execution, inspection, and documented acceptance. Discontinuities become defects only when they exceed specified limits, and their significance depends on type, geometry, location, loading, and service consequence. Cracks and planar flaws generally demand particular attention, while porosity, slag, profile errors, distortion, and residual stress require prevention and appropriate evaluation.

Visual testing is fundamental but limited to accessible conditions. Radiography is effective for many volumetric discontinuities; ultrasonics is valuable for thick sections and planar flaws; magnetic-particle testing detects surface and near-surface flaws in ferromagnetic materials; and penetrant testing reveals surface-breaking flaws in non-porous materials. WPS documents production requirements, PQR demonstrates procedure capability, and welder qualification demonstrates individual competence. Quality assurance emphasizes prevention, traceability, calibrated systems, ethical reporting, and continual improvement. Digital and automated NDT tools extend capability but require validation and professional oversight.

Key Terms

Discontinuity: An interruption or irregularity in the normal structure of a weld or material.

Defect: A discontinuity that exceeds specified acceptance limits.

Indication: A response produced by an inspection method that may represent a discontinuity.

Acceptance criterion: A specified limit used to determine conformity or rejection.

Porosity: Gas cavities trapped in solidified weld metal.

Slag inclusion: Non-metallic material entrapped in a weld.

Lack of fusion: Failure of weld metal to fuse with adjoining metal or a previous bead.

Incomplete penetration: Failure of weld metal to extend through the intended joint root.

Undercut: A groove melted into the base metal beside the weld and left unfilled.

Crater crack: A crack formed in the weld crater during improper termination.

Residual stress: Stress remaining in a component after the original thermal or mechanical cause is removed.

Visual testing: Examination using sight and suitable visual aids or gauges.

Radiographic testing: Examination using X-rays or gamma rays to produce an internal image.

Ultrasonic testing: Examination using high-frequency sound and reflected-wave responses.

Magnetic-particle testing: Detection of surface and near-surface flaws through magnetic leakage fields.

Liquid-penetrant testing: Detection of surface-breaking flaws using penetrant and developer.

WPS: Controlled instruction specifying production welding variables and requirements.

PQR: Record of actual qualification variables and test results demonstrating procedure capability.

Welder qualification: Evidence that an individual can produce acceptable welds within a defined range.

Quality assurance: Planned activities that provide confidence that quality requirements will be met.

Review Questions

Short Answer Questions

1. Distinguish between a discontinuity, an indication, and a defect.

2. Why are planar discontinuities often more critical than rounded porosity?

3. State two limitations of visual testing.

4. What is the purpose of a Procedure Qualification Record?

5. Why may two magnetization directions be required in magnetic-particle testing?

Long Answer Questions

1. Explain the causes, consequences, and prevention of hydrogen-assisted cracking.

2. Compare radiographic and ultrasonic testing for the examination of thick pressure-boundary welds.

3. Describe the contents and functions of a WPS, PQR, and welder qualification record.

4. Explain how a welding quality-assurance system prevents defects rather than merely detecting them.

5. Discuss the effects of distortion and residual stress and describe suitable control methods.

Analytical / Application Questions

1. An aluminium weld contains a suspected surface crack. Select an appropriate NDT method, justify the choice, and identify important process controls.

2. Repeated slag inclusions are found in multipass welds despite acceptable visual appearance. Develop a root-cause investigation and corrective-action plan.

3. A thick carbon-steel butt weld contains a linear indication detected by ultrasonic testing but not clearly shown by radiography. Explain why the results may differ and recommend an acceptance decision process.

22. Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a defect from a discontinuity?

A. A defect is always visible, whereas a discontinuity is internal.

B. A discontinuity becomes a defect when it exceeds specified acceptance limits.

C. A defect is produced only during welding, whereas a discontinuity is produced during design.

D. A discontinuity is always rejectable, whereas a defect may be acceptable.

Correct Answer: B. A discontinuity becomes a defect when it exceeds specified acceptance limits.

Question 2

Which combination is necessary for hydrogen-assisted cracking in susceptible steel?

A. High oxygen, low restraint, and austenitic microstructure

B. Diffusible hydrogen, susceptible microstructure, and tensile stress

C. High shielding-gas flow, low carbon content, and compressive stress

D. Porosity, undercut, and excessive reinforcement

Correct Answer: B. Diffusible hydrogen, susceptible microstructure, and tensile stress

Question 3

Which NDT method is generally most suitable for detecting surface-breaking flaws in an aluminium weld?

A. Magnetic-particle testing

B. Liquid-penetrant testing

C. Demagnetization testing

D. Conventional magnetic flux testing

Correct Answer: B. Liquid-penetrant testing

Question 4

A major limitation of radiographic testing is that it may fail to reveal a planar flaw when the flaw is:

A. Large and volumetric

B. Filled with gas

C. Aligned unfavorably with the radiation beam

D. Located in a uniform-thickness weld

Correct Answer: C. Aligned unfavorably with the radiation beam

Question 5

Which document records the actual variables and test results used to qualify a welding procedure?

A. WPS

B. PQR

C. Welder continuity log only

D. Inspection release note

Correct Answer: B. PQR

Question 6

Why might two magnetic-particle examinations be performed with different field directions?

A. To examine non-ferromagnetic materials

B. To eliminate the need for surface cleaning

C. To improve sensitivity to flaws with different orientations

D. To replace calibration of the equipment

Correct Answer: C. To improve sensitivity to flaws with different orientations

Question 7

In the illustrative piping case, the most effective corrective action was to:

A. Increase current without changing the joint or inspection plan

B. Accept the indications because they were linear but small

C. Control groove geometry, fit-up, cleaning, technique, and repair examination

D. Replace ultrasonic testing with visual inspection only

Correct Answer: C. Control groove geometry, fit-up, cleaning, technique, and repair examination

Question 8

Which statement about ultrasonic testing is most accurate?

A. It is unaffected by material structure and geometry.

B. It is especially useful for planar flaws but depends strongly on calibration and interpretation.

C. It detects only surface-breaking discontinuities.

D. It is applicable only to ferromagnetic materials.

Correct Answer: B. It is especially useful for planar flaws but depends strongly on calibration and interpretation.

Question 9

Which activity is primarily quality assurance rather than immediate quality control?

A. Measuring weld reinforcement after completion

B. Reviewing a radiographic image for porosity

C. Establishing approved procedures, competence requirements, and document control

D. Checking root gap before welding

Correct Answer: C. Establishing approved procedures, competence requirements, and document control

Question 10

What is the most appropriate interpretation of a report stating that no indications were detected?

A. The component contains no discontinuities of any kind.

B. The weld is automatically acceptable under every code.

C. No reportable indication was detected within the method's defined capability and scope.

D. All internal planar flaws have been eliminated.

Correct Answer: C. No reportable indication was detected within the method's defined capability and scope.