Intro4u2u

What is Layer of Protection Analysis?

LOPA is a semi-quantitative risk analysis technique. It lies in between a HAZOP and a quantitative risk assessment (QRA) in terms of its rigorousness. This technique evaluates risks by orders of magnitude of the selected accident scenarios and builds on the information developed in qualitative hazard evaluation e.g. PHA
LOPA - Diagram
0
Layers of Protection
Plants are protected by various layers of protection Examples:

* Pressure Relief Devices
* Flare System
* Fire Suppression System
* Safety Instrumented System (SIS) or Emergency Shutdown System
* Automatic action safety interlock system
* Basic controls, critical alarms
* Community emergency response
* Inherently safe design features
* Operator intervention
* Plant emergency response

When should you use LOPA?

LOPA is applied when a scenario is too complex or the consequence is too severe for the HAZOP team to make a sound judgment based solely upon the qualitative information. On the other hand, it can screen scenarios as a precedent to a QRA.

LOPA helps you to answer the following questions about your facility:

* What’s the likelihood of undesired events / scenarios?
* What’s the risk associated with the scenarios?
* Are there sufficient risk mitigation measures?

Benefits of using LOPA

* Requires less time and resources than for a QRA but is more rigorous than HAZOP.
* Many process safety systems are over-engineered for safety with additional costs and have unnecessary complexity. LOPA helps focus the resources on the most critical safety systems.
* Acts as a decision making tool, helps make judgments quicker, resolves conflicts and provides a common base for discussing risks of a scenario.
* Removes subjectivity while providing clarity and consistency to risk assessment.
* Improves scenario identification by pairing of the cause and consequence from PHA studies.
* Helps to compare risks based on a common ground if it is used throughout a plant.
* Helps decide if the risk is As Low As Reasonably Possible (ALARP) for compliance to regulatory requirements or standards.
* Identifies operations, practices, systems and processes that do not have adequate safeguards.
* Provides basis for specification of IPLs as per ANSI/ISA S84.01, IEC 61508 and IEC 61511.
* Helps to decide which safeguards to focus on during operation, maintenance and related training.
* Support compliance with process safety regulations - including OSHA PSM 1910.119, Seveso II regulations, ANSI/ISA S84.01, IEC 61508 and IEC 61511.

The modern view of maintenance is that it is all about preserving the functions of physical assets. In other words, carrying out tasks that serve the central purpose of ensuring that our machines are capable of doing what the users want them to do, when they want them to do it. The possible maintenance policies can be grouped under four headings viz.

1. Corrective - wait until a failure occurs and then remedy the situation (restoring the asset to productive capability) as quickly as possible.

2. Preventive - believe that a regular maintenance attention will keep an otherwise troublesome failure mode at bay.

3. Predictive - rather than looking at a calendar and assessing what attention the equipment needs, we should examine the ‘vital signs’ and infer what the equipment is trying to tell us. The term ‘Condition Monitoring’ has come to mean using a piece of technology (most often a vibration analyser) to assess the health of our plant and equipment.

4. Detective - applies to the types of devices that only need to work when required and do not tell us when they are in the failed state e.g. a fire alarm or smoke detector. They generally require a periodic functional check to ascertain that they are still working.

Apart from detective maintenance, the central problem that companies have struggled with is how to make the choice between the other three. This has led to the increasing interest within industry in two strategies, which offer a path to long term continuous improvement rather than the promise of a quick fix. These are Reliability Centred Maintenance (RCM) and Total Productive Maintenance (TPM). The two strategies, although having similar names, actually have very different strengths. RCM has been fully described while TPM will now be discussed.

TPM is a manufacturing led initiative that emphasises the importance of people, a ‘can do’ and ‘continuous improvement’ philosophy and the importance of production and maintenance staff working together. It is presented as a key part of an overall manufacturing philosophy. In essence, TPM seeks to reshape the organisation to liberate its own potential.

The modern business world is a rapidly changing environment, so the last thing a company needs if it is to compete in the global marketplace is to get in its own way because of the way in which it approaches the business of looking after its income generating physical assets. So, TPM is concerned with the fundamental rethink of business processes to achieve improvements in cost, quality, speed etc. It encourages radical changes, such as;

*           flatter organisational structures - fewer managers, empowered teams,
*           multi-skilled workforce,
*           rigorous reappraisal of the way things are done - often with the goal of simplification.

It also places these changes within a culture of betterment underpinned by continuous improvement monitored through the use of appropriate measurement. The principal measure is known as the Overall Equipment Effectiveness (OEE). This figure ties the ’six big losses’ :

1. Equipment Downtime
2. Engineering Adjustment
3. Minor Stoppages
4. Unplanned Breaks
5. Time spent making reject product
6. Waste

to three measurables:

Availability (Time), Performance (Speed) & Yield (Quality).

When the losses from Time X Speed X Quality are multiplied together, the resulting OEE figure shows the performance of any equipment or product line.

TPM sites are encouraged to both set goals for OEE and measure deviations from these. Problem solving groups then seek to eliminate difficulties and enhance performance.
TPM achievements

Many TPM sites have made excellent progress in a number of areas. These include:

*           better understanding of the performance of their equipment (what they are achieving in OEE terms and what the reasons are for non-achievement),
*           better understanding of equipment criticality and where it is worth deploying improvement effort and potential benefits,
*           improved teamwork and a less adversarial approach between Production and Maintenance,
*           improved procedures for changeovers and set-ups, carrying out frequent maintenance tasks, better training of operators and maintainers, which all lead to reduced costs and better service,
*           general increased enthusiasm from involvement of the workforce.

However the central paradox of the whole TPM Process is that, given that TPM is supposed to be about doing better maintenance, why do proponents end up with (largely) the same discredited schedules that they had already (albeit now being done by different people)? This is the central paradox - yes, the organisation is more empowered, and re-shaped to allow us to carry out maintenance in the modern arena, but we’re still left with the problem of what maintenance should be done.

The RCM process was evolved within the civil aviation industry to fulfil this precise need. In fact, the definition of RCM is “a process used to determine the maintenance requirements of physical assets in their present operating context”. In essence, we have two objectives; determine the maintenance requirements of the physical assets within their current operating context, and then ensure that these requirements are met as cheaply and effectively as possible.

RCM is better at delivering objective one; TPM focuses on objective two.

Can the techniques be deployed together?

The answer depends on what ‘brand’ of RCM is being considered. The ‘hired gun’ or ‘magic box’ approach will never be compatible with TPM. RCM must be performed by the organisation itself. The sole focus must be to teach organisations to analyse their own assets. In this way, empowered teams remain empowered, ownership is retained and enhanced and companies begin to win the asset management battle.

Generally, solidification cracking can occur when:
•the weld metal has a high carbon or impurity element content
•the depth-to-width ratio of the solidifying weld bead is large
•disruption of the heat flow condition occurs, e.g. stop/start condition
The cracks can be wide and open to the surface like shrinkage voids or sub-surface and possibly narrow.
Solidification cracking is most likely to occur in compositions which result in a wide freezing temperature range. In steels this is commonly created by a higher than normal content of carbon and impurity elements such as sulphur and phosphorus. These elements segregate during solidification, so that intergranular liquid films remain after the bulk of the weld has solidified. The thermal shrinkage of the cooling weld bead can cause these to rupture and form a crack.
48
49
It is important that the welding fabricator does not weld on or near metal surfaces covered with scale or which have been contaminated with oil or grease. Scale can have a high sulphur content, and oils and greases can supply both carbon and sulphur. Contamination with lead, or other low melting point metals such as copper, tin or zinc should also be avoided.
A common form of solidification cracking is referred to as centre-line cracking. The cracks lie on the solidification centre-line where solidification fronts from opposite sides of the weld pool meet. The most unfavourable case is where weld beads are made with a large depth-to-width ratio, and end-to-end contact between opposing solidifying columnar grains results over a large proportion of the depth of the weld bead.
Generally, high depth-to-width ratio weld beads are produced when attempting to deposit weld metal at a high rate, using a process such as Submerged Arc Welding with a single electrode wire and a high welding current. If the same size of weld bead can be made using the process with two or more electrode wires operating in tandem, then a modified solidification pattern is usually formed and solidification cracking avoided.

A perfect butt weld joint, when subjected to an external force, provides a distribution of stress throughout its volume which is not significantly greater than that within the parent metal.
This is achieved as long as the following features apply:
•Welds should consist of solid metal throughout a cross section at least equal to that of the parent metal
•All parts of a weld should be fully fused to the parent metal
•Welds should have smoothly blended surfaces
If any of these requirements are not fulfilled then the weld is imperfect and the stress distribution through the joint is disrupted.
A weld imperfection (or discontinuity) is therefore any object or shape which is capable of creating a stress concentration within a welded construction. Note that in practice, all welds contain small imperfections. However, the majority of these are so small that they do not significantly affect the performance of the joint. If an imperfection is considered of a sufficient size to be detrimental to the structure (as defined by an acceptance standard), it is classified as a defect.
43
5.1.2 Significance
Stress concentrations are detrimental to the performance of a structure as they can:
•lower the load bearing capacity of the joint
•initiate brittle fracture
•nucleate a fatigue crack
•initiate stress corrosion cracking
The greater the stress concentration produced by an imperfection, the more likely it is to cause failure of the weld in service. Therefore, imperfections need to be assessed according to the severity of their stress concentration. This is achieved mainly by the geometric descriptions volumetric or planar.
Volumetric imperfections (those which have three significant dimensions) are less effective stress raisers and are usually accepted in limited quantity and size.
Planar imperfections (those which are essentially two-dimensional) are more effective stress raisers and are almost invariably treated as unacceptable.
5.2 Types of Imperfection
Weld imperfections can be grouped into five distinct types according to their nature and shape. It is important that an imperfection is correctly identified to allow the welding procedure to be suitably modified to prevent their re-occurance. Details of these types of imperfections are contained in the appendix at the end of this document.
Cracks
•Solidification Cracking
•HAZ Hydrogen Cracking
•Weld Metal Hydrogen Cracking
•Lamellar Tearing
Cracks are more significant than other types of imperfection, as their geometry produces a very large stress concentration at the crack tip, making them more likely to cause fracture. Note that this section is concerned only with cracks produced at the time of welding, not subsequent service cracking, such as fatigue or stress corrosion cracking. Cracks can occur in the weld metal or heat affected zone. Due to their severity of stress concentration, crack-like imperfections are usually classed as defects.
44
Lack of Solid Metal
•Porosity
•Worm Holes
•Crater Pipe
•Root Concavity
•Underfill
•Slag Inclusions
•Inter-Run Imperfections
These imperfections are formed when there is insufficient weld metal to completely fill the cross-section between the parent metal plates. They are volumetric (blunt) in shape, and as such are usually only associated with a reduction in the load bearing capacity of a weld.
Lack of Fusion
•Incomplete Root Penetration
•Lack of Sidewall Fusion
These imperfections occur when there is incomplete fusion between the parent metal and weld metal or between weld runs. They are essentially two-dimensional in shape and so are effective stress raisers within the material. Therefore it is important to control them as they can lead to cracking within the weld.
Lack of Smoothly Blended Surfaces
•Surface Porosity
•Excess Weld Metal (Reinforcement)
•Excessive Penetration
•Undercut
•Overlap
It is not immediately obvious that irregularities on the surface of the weld are serious imperfections. However, any sudden change in the contours of the surface produce local stress concentrations. This can especially lead to the formation of fatigue cracks (most commonly at weld toes).
Miscellaneous
•Misalignment
•Arc Strikes
•Spatter
Several miscellaneous imperfections do not conform to any particular category.
45
5.3 Practical Aspects
5.3.1 Tolerance
Weld imperfections can seriously reduce the integrity of a welded structure. Therefore, prior to service of a welded joint, it is necessary to locate them using NDE techniques, assess their significance, and take action to avoid their reoccurrence.
The acceptance of a certain size and type of defect for a given structure is normally expressed as the defect acceptance standard. This is usually incorporated in application standards or specifications.
All normal defect acceptance standards totally reject cracks. However, in exceptional circumstances, and subject to the agreement of all parties, cracks may be allowed to remain if it can be demonstrated beyond doubt that they will not lead to failure. This can be difficult to establish and usually involves fracture mechanics measurements and calculations.
It is important to note that the levels of acceptability vary between different applications, and in most cases vary between different standards for the same application. Consequently, when inspecting different jobs it is important to use the applicable standard or specification quoted in the contract.
5.3.2 Repair
Once unacceptable defects have been found, they have to be removed. If the defect is at the surface, the first consideration is whether it is of a type which is normally shallow enough to be repaired by superficial dressing. Superficial implies that, after removal of the defect, the remaining material thickness is sufficient not to require the addition of further weld metal.
If the defect is too deep, it must be removed by some means and new weld metal added to make up to size.
Replacing removed metal or weld repair (as in filling an excavation or re-making a weld joint) has to be done in accordance with an approved procedure. The rigor with which this procedure is qualified will depend on the application standard for the job. In some cases it will be acceptable to use a procedure qualified for making new joints whether filling an excavation or making a complete joint. If the level of reassurance required is higher, the qualification will have to be made using an exact simulation of a welded joint which is excavated and then refilled using a specified method. In either case, qualification inspection and testing will be required in accordance with the application standard.
46
6. Conclusion
The weldability of a steel broadly refers to the ease with which it can be joined by fusion welding without introducing significant defects or deterioration in properties. Thus, in considering the welding of carbon-manganese steels, key factors which influence their weldability have been explored. In addition, the formal approach to welding quality controls the variables affecting properties, and ensures the welder has the skills required to produce sound joints.
It is hoped you will now have an appreciation of welding carbon-manganese steels which will enable you to recognise good practice. More importantly, when problems are encountered, you should be able to observe whether the causes have been identified and that appropriate corrective actions are taken.

4.2 Bend Tests
Bend tests apply strain to all parts of the welded joint, and are useful for the exploitation and detection of defects and embrittlement. When bend tests are specified for welder approval testing, the intention is to assess the ability of the welder to make sound joints. The bend test also reveals small cracks and embrittlement which arise from the impurities within the steel. The phenomena of burning or the related overheating, which occur close to the fusion line, are likely causes of bend test failure when impure steels are welded. The welder could be unnecessarily rejected due to bad material!
4.3 Impact Tests
The number of impact tests required and the test temperature depends on the standard being followed. The typical locations shown are intended to sample the toughness of all relevant zones at heat inputs around 3kJ/mm. At higher heat inputs, the HAZ will be wider and other locations will be relevant. Note that heat input is an example of an essential variable which cannot be increased without qualification testing when impact properties are specified.
When high heat inputs are used, loss of toughness is most likely in the CGHAZ of the steel. Here, close to the fusion boundary, almost all the carbide, nitride and sulphide particles are dissolved. Impurities (such as
41
sulphur and phosphorus) involved in burning and overheating are particularly detrimental in coarse structures, but have been significantly reduced in modern, clean, steels. Soft, low strength ferrite around the prior austenite grain boundaries also reduces impact toughness and control of the transformation (to give a fine ferritic structure) becomes especially important at high heat inputs.
Some special titanium refined steels contain very stable nitride particles which restrict grain growth, and thereby reduce the grain size and width of the CGHAZ. Other low carbon special steels (developed along the lines of weld metals) contain oxide particles, which promote nucleation of fine and therefore tough intragranular ferrite (which is similar to the acicular ferrite of weld metals).
Strain ageing caused by welding, forming or thermal correction of distortion can also cause loss of toughness at almost any location.

Welds are commonly produced by multi-run sequences such as that shown for a SAW. The weld beads deposited first are largely remelted by subsequent beads. Runs 1 to 7 were deposited from the first side of the double-V weld preparation before the plate was turned over and beads 8 to 14 deposited. Numbers 7 & 14 are thus examples of virgin weld beads which display a columnar grain structure. Elsewhere, in bead 6 (or 13), for example, the columnar structure has been reheated by the overlying bead 7 (or 14). Note that it is the dark etching temper zone which accounts for much of the obscured as-deposited columnar structure; only part of the reheated region microstructure has actually been changed.
Towards the root of the weld, a high proportion of the structure has been reheated and will to some extent have been refined, tempered and partially stress relieved. The toughness is likely to be poorest in the unrefined outer beads 7 and 14, and for this reason Charpy V-notch impact testpieces are often located in these areas. Welding sequences can be devised for welding difficult materials (especially in repair situations) which optimise reheating and refinement of the HAZ.
36
The proportion of columnar structure in a multi-run weld is reduced by an inefficient welding process such as GTAW, with a high heat input per unit mass of weld deposit, expressed in the units kJ/kg.
3.5 Distortion Effects
3.5.1 Residual Stress
When a weld bead is deposited, the fused metal and the part of the HAZ heated sufficiently to form austenite will readily deform under the constraint of the cooler (and therefore stronger) surrounding steel. As the hot metal cools, shrinkage occurs, and yielding takes place. The distribution of the residual stresses which result, particularly with multi-run bead sequences, can be quite complex. Tensile residual stress can have a significant influence on the nucleation of cracks, growth by fatigue and failure by brittle fracture.
Often residual stresses of up to yield magnitude are found in the outer runs (last to be deposited), especially at stop/start locations or local weld repair areas. However, stresses may be compressive in the first runs at the weld root.
3.5.2 Thermal Distortion
Generally, if the structure allows movement, distortion will occur during welding. It is possible to minimise this by balancing the weld beads, for example by simultaneous welding on each side of a stiffener to plate T-joint. 37
3.5.3 Restraint
If a structure is restrained during welding so that thermal distortion is prevented, large residual stresses will be produced. High restraint occurs at cruciform joints where major structural members cross, when welding a circular patch into a thick plate, and for the final or closure joints in a structure (especially for thick sections).
Careful attention is required for high restraint situations. The fit-up of such joints should be correct, as shrinkage and residual stresses are increased by excessively wide welds. Proper monitoring by non-destructive examination should be used.
It is unwise to allow stop/start positions to be located at locations of stress concentration (such as the end of a stiffener) or of high restraint (corners at plate intersections). Cracking is most likely to occur at these locations.
3.6 Post Weld Heat Treatment
In carbon-manganese structures fabricated by welding, removal of residual stresses is important if the risk of brittle fracture or fatigue is to be minimised. For thick section joints, application standards commonly specify Post Weld Heat Treatment (PWHT), which serves to both temper and to stress relieve the joints. Stress relief PWHT is preferably carried out at 600±20°C for 1 hour per 25mm of section thickness. Longer times at lower temperatures (down to an absolute minimum of 550°C) are often used for large structures, which must be heated and cooled very slowly to avoid distortion and residual stresses from uneven temperature distributions.
38
The range of PWHT temperature specified is dependent upon the variation of strength, toughness and degree of relaxation of residual stresses.
The strength of the weld is designed to be above some minimum value (specified minimum strength), based on the stress the fabrication is expected to carry. As the temperature of PWHT increases, the actual weld strength (as measured after the treatment) tends to decrease. The maximum allowable temperature is that where the finished strength falls as low as the specified minimum strength. This will usually be above the temperature at which there is no further significant reduction in residual stress. The finished toughness generally increases as a result of increased PWHT temperature.
Note that the welding consumables used should be chosen according to the PWHT temperature used. If normalising is carried out as a PWHT, a suitable consumable must be employed.

Welding creates heat affected zones (which can be seen on an etched macrosection) because the heat from welding has changed the microstructure and properties of the parent steel. The strength of carbon-manganese steels can be greatly increased by welding (as a result of the rapid cooling during transformation from austenite) but the toughness is often correspondingly degraded. The changes are a result of the thermal cycles to which the steel in each of the zones is subjected.
HAZ
10mm
The thermal cycles experienced in the HAZ can best be understood by referring to the temperature distribution around the welding arc, a source of heat which moves at a constant velocity of w mm/s. The heat distribution is represented by contours (lines) of equal temperature, called isotherms.
A better appreciation can be gained if the heat source is thought of as stationary, and the workpiece is moved at the velocity w. The heat distribution, represented by the isotherms, then appears to stay still. The heat distribution is said to be quasi-stationary.
The movement of the heat source along the plate is actually apparent in two ways. First, ripples form behind the weld pool showing that solid metal is being continually produced. Second, the isotherms in front of the weld pool are closely spaced (so there is a steep temperature gradient) whereas those
31
behind the weld pool are well spaced (the gradient is flatter). This effect has been observed previously in the shape of the weld pool, the trailing edge of which becomes elongated and more ‘V’ shaped as the travel speed is increased.
°C
1500
1200
900
600
300
At the fusion line (point 1) the temperature rises rapidly to the melting temperature and then fall as heat is dispersed to the surrounding plate. At greater distances (points 2-5) from the fusion line, the heating rates decrease, and peak temperatures are progressively lower, and are reached a little later.
The temperature to which the plate is heated before welding starts is called the preheat temperature and the interpass temperature is that to which the joint is cooled before subsequent weld beads are deposited. Raising the plate temperature by preheat can be used to slow the rate of cooling from 800 to 500°C to produce softer transformation products and to increase the time for hydrogen to disperse and escape during welding. Preheat has relatively little effect on the duration and zone width of the high temperature part of the thermal cycle.
The parts of the heat affected zone can be related to the maximum temperatures reached and the distances from the fusion line, which separates fused weld metal from the unmelted heat affected zone.
32
33
In the highest temperature zone (close to the fusion line), iron and manganese sulphides dissociate and the sulphur released segregates and concentrates at the boundaries of coarse austenite grains. For high sulphur steels (generally older-style materials), this can result in a phenomenon known as burning, where low melting temperature liquid films are formed as a result of the locally high concentration of sulphur and phosphorus impurities. Liquation cracking can then result when rupture of the intergranular liquid film is caused by shrinkage strains. Even when no cracking occurs, intergranular weakness from the presence of the impurities is likely to result in low ductility (overheating).
_______________
150μm
The coarse grained HAZ (CGHAZ) generally has the highest hardness, the greatest susceptibility to hydrogen cracking and the poorest toughness. During heating, the temperature reached not only exceeds that required for transformation to austenite, but may also be sufficient to dissolve the fine dispersion of carbide/nitride particles which limit grain growth. The austenite grains which result are therefore not only coarse but are also enriched in dissolved alloy content. The CGHAZ therefore has enhanced hardenability and higher hardness which, combined with the coarseness of structure, can result in loss of ductility and toughness.
_______________
100μm
The transformation products in the CGHAZ depend upon the steel composition and the cooling rate. When ferrite and carbide are formed during slow cooling, coarse plate morphologies can form, which give poor toughness. In the micrograph AC (aligned carbide) points to the inferior plate form while FN (ferrite non-aligned carbide) is better.
_______________
100μm
34
_______________
100μm
If the cooling rate is sufficiently high, martensite will be formed. In carbon-manganese steels, martensite is the hardest and most brittle structure to be formed from austenite during cooling, and is the most susceptible to hydrogen cracking.
_______________
100μm
Grain refinement occurs when the thermal cycle allows sufficient time at temperature (above AC3, the upper critical temperature) for the formation of austenite but does not permit grain growth. This fine grained region is of lower hardness than the CGHAZ.
The intercritical zone lies at the outer edge of the grain refined region. Here, the steel has only partially transformed to austenite at temperatures between AC3 and AC1 (the upper and lower critical temperatures on heating of the steel). The austenite grains are very small, and fine polygonal ferrite plus carbide structures are produced on transformation during cooling. This appears similar to the normalising heat treatment of steel. However, it is worth noting that the dark carbide areas in any part of the weld or HAZ may actually contain untransformed (retained) austenite or martensite in microalloyed or alloy steels.
_______________
100μm
Tempering produces sub-microscopic changes, accompanied by relaxation of internal stresses, which influence the etch response giving rise to the dark etching band outside the intercritical zone. A slight softening may occur due to over-tempering in this region.
35
In the parent steel (outside the visible HAZ) the microstructure is polygonal ferrite plus pearlite, and shows segregation banding. Wherever the steel is deformed (by rolling, bending, fitting, line heating or thermal strain during welding), strain ageing can occur if nitrogen is present in the ferrite structure. In aluminium-treated fine grained steels, the precipitation of aluminium nitride particles during normalising reduces the free nitrogen content, thereby reducing strain ageing.

When austenite (FCC crystal structure) is cooled, a phase transformation takes place to form ferrite (BCC crystal structure). Carbon, which is highly soluble in austenite, must now be rejected as iron carbide or accommodated by the formation of martensite, as its solubility is much lower in ferrite.
28
The effect of cooling rate upon the transformation products formed can be illustrated by a Continuous Cooling Transformation (CCT) diagram. Note that practical cooling rates are conventionally measured between temperatures of 800 and 500°C.
Composition is the dominant factor which determines the microstructure and properties for a given cooling rate. Increasing manganese and other alloy elements typically suppresses the high temperature transformation products like polygonal ferrite and allows higher heat inputs to be used. However, increasing alloy content risks the formation of martensite at low heat input.
29
On very slow cooling (1°C/s), coarse polygonal ferrite grains are produced with islands rich in carbon, which form a pearlitic structure.
At a cooling rate of 10°C/s, the polygonal ferrite grains are fine, and signs of intragranular ferrite nucleation to form coarse acicular ferrite are apparent. This acicular ferrite is nucleated by the dispersion of fine oxide particles throughout the structure, and is beneficial to the toughness of the weld metal.
At a cooling rate of 50°C/s, some polygonal ferrite is present, together with laths of sideplate ferrite. There is a large proportion of fine acicular ferrite in the microstructure. This is a typical structure for a carbon-manganese weld using normal welding conditions.
At very fast cooling rates (1250°C/s), all other transformations are avoided and martensite is formed. It is the hardest structure and the BCC structure of ferrite is distorted to accommodate the carbon.

Gases can enter the molten metal from the arc atmosphere, as well as from slag-metal reactions during welding. It is beneficial to reduce the content of these gases to a minimum, as they can adversely affect the properties of the weld.
•Oxygen
If an inert shielding gas is used, the oxygen content of the weld will be very low. Most processes used for the bulk welding of steels, however, involve the use of an oxidising gas (typically carbon dioxide, CO2) or oxide-containing slags, such as lime (CaO), manganese oxide (MnO), silica (SiO2) and titania (rutile, TiO2). These decompose to produce oxygen, which dissolves in the molten metal.
Reactions which occur between the molten metal and slag result in a weld metal oxygen content of typically 0.05wt.%, which is then precipitated during solidification as deoxidation products. These are commonly manganese silicate (2MnO.SiO2) and take the form of glassy spherical droplets 0.2-2mm in diameter, occupying a volume fraction of around 0.5% of the weld metal. Deoxidation products reduce toughness by nucleating ductile mode fracture, but are also beneficial as they can nucleate the desirable intragranular acicular ferrite microstructure.
•Carbon Monoxide
Carbon monoxide can also be evolved as a deoxidation product during solidification of the weld metal (where carbon and oxygen combine to form this gas). As a result, porosity can occur when a high oxygen weld metal is diluted by a high carbon parent plate. This can be prevented by using a consumable with a higher silicon content, which
26
lowers the dissolved oxygen content.
27
•Argon
Porosity and wormholes can often result from gas which has expanded on heating and then become frozen-in during solidification. With GTAW it is not uncommon for argon gas to be introduced with cold filler wire, and then trapped.
•Nitrogen
Nitrogen is normally present in steel weld metals at a higher concentration (typically 0.01%) than for parent steels. Weld metal aluminium contents are generally low, and so free nitrogen is likely to be present, leading to strain ageing. Post weld heat treatment at around 600°C is beneficial in limiting strain ageing if the silicon content is greater than 0.25%, as the free nitrogen is effectively fixed by the formation of silicon nitrides.
Porosity can also result from the reduction in solubility on solidification at high nitrogen concentrations (above about 0.03%N). This can be at the start of welding (start porosity), or from the use of an excessively long arc when welding with basic coated electrodes.
•Hydrogen
Hydrogen is always mobile in the microstructure, as it cannot be fixed as a compound. Therefore, the best practice is to avoid its introduction from water vapour (moisture), grease or other organic compounds. Hydrogen is soluble in ferritic iron above 200°C but is much less soluble at ambient temperature. However, it is still mobile at this low temperature. Therefore, if the weld metal stress and hardness are sufficiently high, hydrogen will migrate to any local weakness or stress concentration and hydrogen cracking may result. Preheat is employed to keep the weld above the temperature at which cracking occurs and allows an opportunity for the hydrogen to disperse and escape. Note that basic electrodes will typically result in a low hydrogen content.

3. The Welding Sequence
3.1 Weld Metal
As weld metal is molten when deposited, there is no prior history to the material’s microstructure, so it is primarily determined by composition and cooling rate.
3.1.1 Weld Pool
When the molten material enters the weld pool it is diluted by mixing with the melted parent plate. Although the undiluted composition produced by the welding consumable may be optimised and pure, the parent plate may be lower in alloying elements (such as manganese), or introduce an undesirable level of impurities such as sulphur and phosphorus into the weld pool. Prepared surfaces may also be contaminated with scale and rust, and the ability to cope with these will depend upon the welding consumables.
3.1.2 Solidification
Solidification of the weld metal takes place by columnar grain growth at the trailing edge of the weld pool. Ripples on the weld bead (or segregation bands revealed by etching of a section through the weld) result from alternating slow and rapid growth of the solid metal. The columnar grains grow perpendicular to these indications of successive positions of the growth front, at the solidification temperature (about 1530°C for mild steel). At the sides of the weld pool, the heat flows to the cold plate, and the grains grow in the opposite direction towards the centre of the weld. Towards the centre of a low speed weld, the grains sweep round to follow the heat source.
The columnar grains of the weld bead grow from the coarse grained region of the HAZ. Therefore, the grain size of the weld may be influenced by that of the parent steel.
23
In a faster weld, the pool trails further behind the arc and becomes V-shaped. As a consequence, the columnar grains are straighter and impinge at the weld centreline. At a higher rate of heat input (I.V), the weld pool becomes larger.
In the presence of impurities (such as sulphur), the centre-line of the weld can contain a low-melting temperature intergranular film, and is therefore susceptible to solidification cracking. A deeply penetrating weld bead (as shown for the SAW example) will also favour development of a central plane of weakness.
The origin of this low-melting temperature intergranular film can be understood by referring to the Fe-S phase diagram. Sulphur segregates during solidification, which can result in a high sulphur concentration within the last liquid to solidify. This can lower the solidification temperature of this last liquid to below 1000°C (compared with over 1500°C for pure iron).
Other impurities (such as phosphorus) act in a similar but less severe manner. Manganese is beneficial as it will segregate with the sulphur and lower its concentration in solution by the precipitation of manganese sulphide, hence
24
restricting the depression of melting point.
25
An increase in the size of the weld pool, caused by raising the rate of heat input (I.V), will promote cracking. However, solidification cracking will still not normally occur unless the smooth progress of the weld is interrupted. Such disruptions occur at the start and stop locations as shown.