In this article we will discuss about the welding of metals:- 1. What is Welding? 2. Principles of Solid Phase Welding 3. Principles of Fusion (Liquid State) Welding 4. Weld Defects and Inspection.
What is Welding?:
Unlike the manufacturing processes employed to produce a single component, the joining processes are used to assemble different members to yield the desired complex configuration. Such a complex geometry is either too difficult or impossible to obtain by using only the manufacturing processes. The joining processes are so intimately related to the overall production system that these are also considered to form a class of manufacturing techniques.
The joining of different elements can be either temporary or permanent in nature. Also, the mechanism of bonding may be either mechanical or atomic. All joining processes involving atomic bonding are of a permanent nature. In mechanical bonding, the strength of the joint is less than the combined strength of the original members. In atomic bonding, however, the situation is not necessarily so.
Another criterion used for a classification of the joining processes is based on the composition of the joint. According to this scheme, all joining processes can be grouped into three different categories, namely – (i) autogeneous, (ii) homogeneous, and (iii) heterogeneous. In the processes belonging to (i), no filler material is added during joining. All types of solid phase welding and resistance welding are examples of this category.
ADVERTISEMENTS:
In the homogeneous joining processes, the filler material used to provide the joint is the same as the parent material. Arc, gas, and thermite welding belong to this category. In the processes of type (iii), a filler material different from the parent material is used. Soldering and brazing are two such joining processes. It may be noted that two materials which are insoluble in each other, such as iron and silver, can be joined by a heterogeneous process. This may be achieved by using a filler material (e.g., copper and tin) which is soluble in both the parent materials (i.e., iron and silver).
The bonding force between two metallic atoms decreases very sharply with the interatomic distance. When the distance is more than a few atomic spacings (i.e., a few angstroms), the interacting attractive force reduces to almost zero. But the force increases sharply and attains a very large value when the distance is reduced.
Thus, if it is possible to bring together two metallic surfaces so that nothing but the grain boundaries separate them, the two bodies will adhere with a very large force, resulting in what we call welding. However, in normal atmosphere, the metal surfaces are contaminated with layers of oxides and adsorbed gases. These layers are normally a few hundred angstroms thick. So, it is not possible to generate a strong attractive force when two metal surfaces are brought in contact.
But this difficulty can be eliminated when the contaminating layers are removed from the surfaces. Though this may appear as a problem for welding, it is rather fortunate to have these contaminating layers. For example, in the outer space, a major problem is not the welding but the un-welding of surfaces.
Principles of Solid Phase Welding:
ADVERTISEMENTS:
The solid state welding processes may be carried out both at the room temperature and at an elevated temperature without, of course, melting any part of the joining surfaces. For a better understanding of the quality of a solid phase joint, it is worthwhile to recapitulate the strength and cohesion of metals.
A defect-free crystal fails by a cleavage along a crystallographic plane where the inter-atomic force is the weakest. As a result, two new surfaces are produced, and the surface energy γ is defined as the work done in order to create these surfaces. The strength of a single crystal (σc) is found to be –
Where l (≫d) is the length of the crack. The failure of a polycrystalline ductile material is due to the movement of dislocations, resulting in plastic instability, and this takes place at a stress much lower than that given by equation (5.1).
ADVERTISEMENTS:
The bulk strength of a material is much lower than the bonding forces of the constituent atoms. So, a good welding does not require to achieve a strength equal to that between the adjacent lattice planes. Moreover, it should be remembered that at the room temperature, i.e., with negligible creep, even a plane of lattice misfit is not weaker than the bulk material. This information is important because a cold weld junction is essentially a plane of lattice misfit.
When two metal surfaces are brought into contact, the real contact takes place through a small area of asperities. This metallic bridging occurs even in the presence of adsorbed surface layers. The bridges so formed have the property of a true grain boundary, and hence are stronger than the bulk material. Some work hardening also takes place in the layers, just beneath the mating surfaces.
If the yield strength (or flow pressure) of the material is σy with the applied force as pe, then the fraction of the total area coming in contact, and thereby forming a weld, is simply (pe / σy). However, around the welding zones, there will be some areas which come in contact (without actual flow) where the stresses are still within the elastic range. The experimental results suggest that including this area, the total area of physical contact, with a moderate external pressure, can be taken as 2pe / σy.
When the applied load is removed, the two surfaces separate out only when the elastic forces trapped in the regions around the bridges are strong enough to break apart these metallic bridges. It is seen that the softer the material, the better the permanent adherence.
ADVERTISEMENTS:
In the solid phase welding processes, the four important factors are – (i) surface deformation, (ii) surface films, (iii) recrystallization, and (iv) diffusion.
The surface deformation that takes place during welding is difficult to measure. As such, in pressure welding, the bulk deformation is used as an index of the surface deformation and is expressed as –
The strength of a welded junction increases with increasing bulk deformation. Moreover, no weldment takes place below a certain critical deformation. The amount of deformation necessary for obtaining a specific strength decreases with increasing temperature. A strong weld may be made with only 10% deformation if the working temperature is quite close to the melting point of the material. The ratio of the oxide hardness and the parent metal hardness also effectively governs the amount of necessary deformation.
ADVERTISEMENTS:
The greatest hurdle in solid phase welding is posed by the surface oxide layers and oil films. The liquid films can be removed by heating in hot welding, and by means of scratch brushing in cold welding. The oxide films can also be reduced to a certain extent by scratch brushing. Moreover, these oxide layers (being hard and brittle) fracture when the pressure is applied.
A lateral movement is very useful (as in ultrasonic welding) since this tends to roll together the fragmented oxide layer into a relatively thick agglomerate. This results in a more metal- to-metal contact area. An excessive oxide contamination is always harmful, resulting in a poor joint efficiency.
A solid phase welding done at the room temperature does not allow recrystallization and grain growth at the interface. This reduces the ductility of the joint to some extent. An increase in working temperature not only increases the ductility but also eliminates some other defects. The phenomenon of diffusion has an important bearing on the performance of a solid phase weld. The shape and the size of the voids at the interface are modified considerably depending on the amount of diffusion.
Principles of Fusion (Liquid State) Welding:
In a fusion welding process, the material around the joint is melted in both the parts to be joined. If necessary, a molten filler material is also added. Thus, a fusion welding process may be either autogeneous or homogeneous. Metallurgically, there are three distinct zones in a welded part, namely – (i) the fusion zone, (ii) the heat affected unmelted zone around the fusion zone, and (iii) the unaffected original part.
The most important factors governing a fusion welding process are:
(i) The characteristics of the heat source,
(ii) The nature of deposition of the filler material in the fusion zone, known as the weld pool,
(iii) The heat flow characteristics in the joint,
(iv) The gas metal or slag metal reactions in the fusion zone, and
(v) The cooling of the fusion zone with the associated contraction, residual stresses, and metallurgical changes.
1. Heat Source:
A heat source, suitable for welding, should release the heat in a sharply defined, isolated zone. Moreover, the heat should be produced at a high temperature and at a high rate. The most common sources of heat include – (i) the electric arc (as in various arc weldings), (ii) the chemical flame (as in gas welding), (iii) an exothermic chemical reaction (as in thermite welding), and (iv) an electric resistance heating (as in electro slag and other resistance welding processes). The general characteristics of these heat sources are now discussed.
Emission and Ionization of Electric Arc:
First of all, let us see how an electric arc is created and maintained between two electrodes of opposite polarity. Figure 5.1 schematically shows an electric circuit used for arc welding where the work is the positive electrode (called the anode) and the electrode rod is the negative electrode (called the cathode). Initially, a good contact is made between the electrode and the work. Thereafter, the electrode is withdrawn. As a result, the metallic bridges start breaking, thus increasing the current density per bridge.
Finally, the current density rises to such a high value that the bridges start boiling. Under such conditions, the electrons come out of both the surfaces by a process known as thermionic emission. Obviously, the electrons (having the negative charge) coming out of the anode are pulled back, whereas those coming out of the cathode are also attracted towards the anode.
The rate at which the electrons are emitted from a hot surface is given by –
with e = charge of an electron, k = Boltzmann’s constant, and ɸ (when measured in electron volts) as the thermionic work function. ɸ, in fact, represents the kinetic energy necessary to ‘boil’ out an electron. The values of ɸ for some common metals are shown in Table 5.2. It is obvious from equation (5.3) that a low value of together with a high value of θ, makes the emission of electrons easier. Once started, the arc itself becomes a source of ions through a process of ionization.
These ions are attracted by the cathode and the resulting collisions keep the cathode hot. The total current in the arc is carried by two sets of electrons. The first set, emitted by the cathode, is called primary electrons, and the second set, known as secondary electrons, is produced as a result of the ionization of the arc gap. With tungsten and carbon electrodes, the primary electrons carry most of the current, whereas with copper or aluminium electrodes, the secondary electrons carry most of the current.
An electron of charge e, moving in an electric field of gradient E (volt / distance), experiences a force of magnitude eE. In other words, it accelerates at a rate of eE / m, where m is its mass. So, if it travels through a distance d before colliding with another particle (a neutral atom or another electron), it has a kinetic energy eEd. This kinetic energy is nothing but heat and manifests itself through increased temperature.
The inter particle collisions, taking place in the gap between the electrodes, give rise to a process called thermal ionization. Normally, these collisions are elastic and both the momentum and kinetic energy are conserved. However, occasionally a collision is such that an electron may be completely knocked out from a neutral atom, producing a free electron and a positively-charged ion. Such a collision is, of course, not elastic in nature.
The ions thus produced are attracted towards the cathode, as already explained in the foregoing paragraph. The free electrons (earlier termed as the secondary electrons) help the arc to remain electrically conductive. A definite amount of energy is required to produce ionization in a given atom or molecule. This energy (in electron volts) is numerically equal to the ionization potential (in volts). The ionization potentials for different metal vapours are also shown in Table 5.2.
Most of the ion-producing collisions in an arc are between hot, neutral atoms and molecules. To maintain the conductivity of the arc, only a small fraction of the atoms need to be hot enough to ionize, whereas the rest of the arc should be hot enough to supply the fast atoms. For most common gases and vapours at the atmospheric pressure, the arc temperature is of the order of 6000°C.
Arc structure, characteristics, and power structurally, we can distinguish five different zones in an electric arc.
These are as follows:
(i) Cathode spot – This is a relatively very small area on the cathode surface, emitting the electrons.
(ii) Cathode space – It is a gaseous region adjacent to the cathode and has a thickness of the order of 10-3 cm. This region has the positive space charge, so a voltage drop is necessary as the electrons are to be pulled across this region.
(iii) Arc column – This is the visible portion of the arc consisting of plasma (hot ionized gas) where the voltage drop is not sharp.
(iv) Anode space – This, again, is a gaseous region (thickness ≈10-3 cm) and is adjacent to the anode surface where a sharp drop in the voltage takes place. This is because the electrons have to penetrate the anode surface after overcoming the repulsion of the thermionically emitted electrons from the anode surface.
(v) Anode spot – This is the area on the anode surface where the electrons are absorbed. This area is larger than the cathode spot.
The potential drop across an arc is schematically shown in Fig. 5.2. The voltage drop shown in this figure is for given spacing, current, and electrode materials. A change in the materials alters all the values. However, a change in the spacing and the current essentially changes only the drop in the arc column.
It has been experimentally found that, for given spacing (and, of course, electrode materials), the voltage reduces up to a current value of 50 amp (against the ohmic law of constant resistance) and increases thereafter, as shown in Fig. 5.3. This can be explained as follows. Up to 50 amp of current, the shape of the arc is almost cylindrical and the surface to volume ratio of a cylinder decreases with increasing radius.
Thus, a thick, high current arc loses less heat and essentially burns hotter. This results in a higher conductivity (and consequently lower resistance) as compared with a thin, low current arc. However, beyond 50 amp of current, the arc bulges out and the current path becomes more than the arc gap which again increases the resistance of the arc. Due to these two opposite effects, i.e., higher temperature and longer current path, the voltage drop remains constant over a wide range of the current values.
As a first approximation, we can assume the conductivity of the arc column to be independent of the arc length l. The electrode drops are also independent of the arc length. Hence, we can write the voltage drop across the entire arc as –
Where A is the electrode drop and Bl represents the column drop.
The voltage-current relationship of an arc (Fig. 5.3) determines the required characteristics of the power source. In Fig. 5.4, we consider two different characteristics of the power source. The curve AB represents a flat characteristic, whereas the curve CD represents a sharply drooping one. In this figure, two typical arc characteristics for two different arc lengths (say, l and l + Δl ) are also indicated (see the dashed lines).
The intersections of the characteristic of the source and that of the arc determine the operating points. It can be easily seen that the stable operating points are given by the intersections on the right-hand side (shown by the solid circles), and not by those on the left-hand side (shown by the solid squares), of the figure. This can be verified by considering a change, say, an increase, in the arc current.
At the points shown by the solid circles, such an increase causes an increase in the voltage which, in turn, decreases the generator (source) current. Thus, any disturbance is automatically opposed, and the operating points return to their original values. At the points shown by the solid squares, just the opposite phenomenon takes place, i.e., any disturbance moves the operating points continuously away from their original locations. The changes in the arc current for the two power sources for a given change in the arc length (from l to l + Δl) are also indicated in Fig. 5.4.
In manual arc welding, an inadvertent change in the arc length is inevitable.
However, this should not cause a large change in the welding current. This obviously makes the sharply drooping characteristic desirable for manual arc welding. With a flat characteristic, for a big change in the arc length, there may not be any point of intersection between the arc and the source characteristics and the arc may blow out. For an efficient striking of the arc, it is necessary that the open circuit voltage of the source be high above the operating voltage.
Moreover, it is necessary to have a quick response of the source (low time constant) since the welding process itself is unsteady. The power source should be such that it is not damaged by short circuiting for an appreciable length of time.
In a semiautomatic arc welding process, the arc is maintained between the work piece and a wire which is driven forward at a constant speed as it melts away from the tip. An increase in the arc length increases the voltage [equation (5.5)] and, as a result, the current falls. The melting rate being dependent on the current, an increase in the arc length causes a decrease in the melting rate.
If the reduction in the current due to an increase in the arc length is significant, the melting rate decreases considerably so that the arc length returns to its original value. Hence, in this case, for a stable operation, a flat characteristic of the power source (Fig. 5.4) is desirable.
The power of an arc varies with its length and there is an optimum length for which the arc power is maximum. This optimum arc length (lopt) can be determined as follows. For a given length, say, l1 first the arc voltage (V) is determined from equation (5.5). Then, from the source characteristic (Fig. 5.4), the arc current (l) is determined for this value of the arc voltage.
The product of these two, i.e., VI, gives the power (P1) for the given arc length l1. This procedure can be repeated for various values of the arc length and a plot of the arc power (P) versus the arc length (l) can be obtained (Fig. 5.5). Now, the optimum arc length lopt can be easily determined from this figure. Since the electrode drop is utilized with a higher efficiency than the column drop, the actual optimum length is a bit shorter than the optimum obtained from Fig. 5.5.
Every half cycle of a 50-Hz alternating current (ac) takes 0.01 sec, whereas an arc takes only about 0.001 sec to reach the equilibrium state. Due to this quick response, both Fig. 5.3 and equation (5.5) are equally applicable for every half cycle of an ac arc as well. It should be remembered, however, that an ac arc must reignite itself after every crossing of the zero current instant. Reignition requires a voltage higher than the normal arc voltage.
The process of reignition of an arc is facilitated by the presence of ions having a low ionization potential. So, the electrodes for an ac arc welding are coated with potassium silicate binders, whereas those used for a dc arc welding are normally coated with sodium silicate. From Table 5.2, it is readily seen that potassium has a lower ionization potential as compared with sodium.
Chemical Heat Source:
Acetylene (C2H2) is the most common chemical heat source and is used in a chemical gas flame. In the presence of excess oxygen, it burns according to the reaction –
Once the amount of heat liberated (ΔH) is known, we can roughly estimate the maximum flame temperature with the assumption of an adiabatic flame. This means that the entire ΔH leaves the flame only through the heating of the reaction products. Care must be taken to subtract the latent heat from ΔH if any of the reaction products undergoes a change of phase. The entire reaction is assumed to be completed at the room temperature (say, θR) Then, the flame temperature (θf) can be computed from the equation –
where ΔH = 0.242 x 106 kJ/ kg of atomic weight of the contained oxygen. The adiabatic temperature is calculated to be of the order of 3000°C. Reaction (5.9) is utilized in what is known as thermite welding.
Contact Resistance Heat Source:
The electrical resistance heating, too is a heat source. This may be done either by utilizing the contact resistance of the interfaces (as in various resistance welding processes) or by utilizing the resistance of a molten flux and slag (as in electro slag welding).
When two metallic surfaces are brought into contact, only a small fraction of the apparent area is in actual metal-to-metal contact. When a current is sent through such an interface, all of it is carried by these tiny metallic bridges. The oxide layers in contact carry no current. As a result, the current flow is constricted, as indicated in Fig. 5.6. Due to this constriction, the resistance to the flow of current increases, and this increment is termed as contact resistance.
An estimate of the order of magnitude of this resistance can be obtained by idealizing the constricted current flow (as shown in Fig. 5.7) with the following assumptions:
(i) All bridges are of uniform size and spherical shape with radius r1.
(ii) All bridges are uniformly spaced at a distance 2r2 apart.
(iii) The constriction effect due to each bridge is restricted within a concentric sphere of radius r2.
(iv) Each bridge is of zero resistance.
Now, if there are n bridges per unit area, then the contact resistance per unit area can be calculated from the resistance of each spherical constriction and considering n such paths in parallel. Each spherical constriction consists of two identical hemispheres in series. The resistance of each hemispherical constriction is given by –
Where ρ is the resistivity of the material, (r2 — r1) is the length of the current
path, and S is the geometric mean area of the two hemispheres of radii r2 and r1, respectively. Thus –
In the absence of the interface, the resistance of the same path is negligible. Thus, the contact resistance per unit area can be taken as that given by equation (5.12). Experiments show that the assumptions leading to equation (5.12) do not cause an error more than 15%. So, the contact resistance per unit area can, finally, be taken as –
The rate of heat generated by this contact resistance with an applied voltage V is V2 / Rc per unit area. However, after a very short time (≈0.001 sec), the contact resistance drops to about (1 / 10)-th of its original value. This is mainly due to the softening of the material as the temperature increases. As the material softens, the value of the quantity (nr1) used in equation (5.13) increases. This effect is more predominant than the increase of the bulk resistivity (ρ) with the temperature.
2. Modes of Metal Transfer in Arc Welding:
The depth of penetration, the stability of the weld pool, and the amount of spatter loss depend, to a large extent, on the mode of metal transfer from the consumable electrodes. Various forces cause the transfer of metal into the weld pool. The mode of transfer depends on the intersection of these forces and governs the ability of welding in various positions. The major forces which take part in this process are those due to – (i) gravity, (ii) surface tension, (iii) electromagnetic interaction, and (iv) hydrodynamic action of plasma.
The force of gravity may be a retaining or a detaching force, depending on whether the electrode is pointing upward or downward. But the surface tension always tends to retain the liquid drop at the tip of the electrode. This force depends on the radius of the electrode, the capillarity constant, and the density of the liquid metal. The electromagnetic force, known as the Lorentz force, is set up due to the interaction of the electric current with its own magnetic field.
This force acts in the direction of the current when the cross-section of the conductor is increasing in the direction of the current. Similarly, the force acts in the direction opposite to that of the current if the cross-section of the conductor is reducing in the direction of the current. Figure 5.8 explains how this force accelerates the process of separation of a droplet which has started to separate out. The hydrostatic pressure is created due to the magnetic force.
At a high current density, this pressure elongates the liquid drop and also adds to its stiffness. As a result, the liquid drop is projected along the line of the electrode, independent of gravity.
The plasma of the arc also causes the drop to be projected towards the work piece, whereas a high evaporation rate from the surface of the drop tends to move it in the opposite direction.
All these forces interact in a complicated manner and give rise to two broad classes of metal transfer, namely, free flight transfer and short circuit transfer. In the former, the liquid drop travels freely in the arc space, i.e., gets completely detached from the electrode before contacting the work piece. The free flight transfer may be (see Fig. 5.9) – (i) gravitational, (ii) projected, and (iii) repelled.
When the transfer is gravitational, the predominant force is that of gravity and the molten drop falls almost vertically from the electrode into the weld pool. If the electromagnetic force, the gas jet, and the hydrostatic pressure are predominant, then the drop is given an initial acceleration towards the weld pool, and thus projected into it independent of gravity. If the resulting force directs the drop away from the weld pool, then the repelled transfer occurs.
This situation is encountered when CO2 is used as the shielding gas, particularly at low and moderate currents. Obviously, the gravitational transfer is not very reliable, and the repelled transfer is undesirable since it causes too much of a spatter loss. The projected transfer is seen in oxide coated carbon steel electrodes where a strong gas jet is set up.
In the short circuit transfer, the liquid drop at the tip of the electrode gets in contact with the weld pool before being detached from the electrode. Thus, the arc is momentarily short circuited. However, due to the surface tension and the electromagnetic force, the drop is pulled into the weld pool and the contact with the electrode is broken. This re-establishes the arc.
Figure 5.10 schematically shows a short circuit transfer. Here, the spatter loss is minimum and can be achieved by controlling the gap and the other welding variables. This type of transfer, being independent of gravity, is suitable for overhead welding purposes.
3. Heat Flow Characteristics:
A study of heat flow characteristics can provide an estimate of the minimum heat input rate required to form a weld of a given width. Moreover, recognition of the major variables controlling the thermal cycle (i.e., the heating and the cooling rate of the heat affected zone) is essential for a successful fusion welding. In the fusion welding processes, the heat source is moving, except in spot welding where the source is stationary. Once the steady state is reached, even with a moving heat source, the temperature distribution relative to the source becomes stationary.
The most convenient way of analyzing such a problem with a moving source is to assume the source as stationary and the work piece as moving in the opposite direction with the same velocity as that of the moving source. This speed is called the welding speed. Two different types of heat sources can be considered. In most cases, the heat is liberated in a small zone which is idealized as a point source, and the heat flow from the source is three-dimensional.
In a few cases, e.g., in butt welding of relatively thin plates, the heat is liberated along a line and the heat source is idealized as a line source. In such situations, the heat flow is two-dimensional. These two types of heat sources are explained in Fig. 5.11. For an elaborate analysis of the temperature distribution in the work piece under various situations, see the literature.
The available results include those of infinite, semi-infinite, and finite medium, each with point and line sources. Of these results, the most useful is the one which gives the minimum heat input rate necessary for maintaining a given width of the weld. For a three-dimensional heat source, this is given as –
Where h = plate thickness. It is clearly seen from equations (5.14) and (5.15) that the most important parameter is vw/α.
It should be noted that the theoretical results fail to accommodate many practical considerations, e.g., inhomogeneous conducting medium (liquid within the weld pool and solid outside), and absorption and rejection of the latent heat at the forward and the rear edges, respectively, of the weld pool. However, equations (5.14) and (5.15) are still useful for providing a good estimate.
In arc welding with short circuit metal transfer, the heat input rate is easily seen to be –
If the heat input rate given by equation (5.16) falls short of that given by equation (5.14) or equation (5.15) (as the case may be), a lack of side fusion occurs.
In fusion welding, the other important heat flow variables are the cooling rate and the thermal cycle. A mathematical analysis leads to the following conclusions which are in accordance with the practical experience.
The cooling rate increases with increasing weld speed, and for a given weld speed, the cooling rate increases with decreasing size of the weld pool. For example, in electro slag welding, since the weld pool is large and the welding speed is very slow, the cooling rate is seen to be very low. On the other hand, in automatic tungsten-inert gas welding, the process is operated at a very high speed with a small weld pool, and this result in a very fast rate of cooling.
The thermal cycle at any point in the medium is mainly governed by its distance from the heat source. Obviously, with increasing distance from the source, the maximum temperature is lower and the temperature lags further behind the source. Figure 5.13 shows the variation of temperature with time at different distances from source.
4. Gas Metal Reaction:
The absorption of gas in the weld pool from the arc or the flame plays an important role in most fusion welding processes. This is due to the possibility of a reaction between the gas and the liquid metal in the weld pool. The chances of such a reaction are enhanced by the high temperature of the gas and the metal.
There can be two different types of reactions. In one type, the gas may just get dissolved in the liquid metal. In the second type, on the other hand, the gas and liquid metal may react chemically to form a stable compound. In such a case, the situation may be considerably different, depending on the degree of solubility of the reaction product in the weld pool.
As long as the reaction product is soluble, it does not prevent the formation of a weld pool. However, it may result in an embrittlement of the welded joint. An insoluble reaction product produces either surface scales or slags, and thus physically interferes with the formation of the weld pool. In this case, either the excess gas to the weld pool is prevented or a flux is used to dissolve and disperse the reaction product.
When the gas gets dissolved in the liquid weld pool, there is obviously no hindrance towards the formation of the weld pool. However, as the solubility decreases on cooling, degasification starts and, with suitable nuclei, bubbles may form. If these bubbles are trapped, then the quality of the weld is very poor. Even otherwise, this degasification makes the joint porous. This defect is very common in a metal whose oxides are easily reducible by hydrogen, and can be avoided by the addition of a suitable de-oxidant in the filler metal.
Another important gas metal reaction is the diffusion of the gas into the parent metal from the weld pool. When the temperature of the thermal cycle is high, this diffusion process may be quite fast. The diffusion of hydrogen into the heat affected zone may, again, cause an embrittlement of the welded joint.
5. Cooling of Fusion Weld:
The three important effects intimately connected with the cooling of a fusion weld are:
(i) Contraction,
(ii) Residual stress, and
(iii) Metallurgical phase transformation.
All these effects significantly control the quality of a weld.
(i) Contraction:
During the freezing of the weld pool, a decrease in the volume takes place. Moreover, the direction of freezing, and thus the effect of contraction, depends on the type of joint, as explained in Fig. 5.14. Figure 5.14a shows the solidification of a groove weld. Here, the solidification front moves simultaneously from the bottom upwards and from the sides inwards.
Further, the molten top portion always makes up for the contraction in the inner layers and piping occurs only in the surface layers. Figure 5.14b shows the solidification in a corner joint where more cold metal is near the surface of the weld pool. Thus, the top of the weld pool freezes faster and a long piping throughout the joint may occur.
(ii) Residual Stress:
During the fusion welding of plates, as the weld pool contracts on cooling, this contraction is resisted by the rest of the plates (which have not melted). As such, a tensile stress is generated in the weld, and this is balanced by the compressive stress in the parent metal. Figure 5.15 shows a typical distribution of these stresses in a plate weld. This residual stress may result in the cracking of a brittle material and is not important as far as a ductile material is concerned.
(iii) Metallurgical Changes:
These changes are due to the heating and subsequent cooling of the weld and the heat affected zones of the parent materials. Such changes significantly affect the quality of the weld.
The wide variety of changes that may take place depend on various factors, e.g.:
(a) The nature of the material, i.e., single-phase, two-phase,
(b) The nature of the prior heat treatment, if any, and
(c) The nature of the prior cold working.
We now consider typical examples of these changes.
Let us consider the fusion welding of two pieces of a single-phase material, which have been cold worked to yield a desired grain orientation. These cold worked grains result in a high strength and low ductility. However, on fusion welding, a random grain growth again takes place within the melt boundary, which, in turn, results in a low strength. Within the heat affected zone, the grains become coarse due to heat input (annealing), and a partial recrystallization also occurs.
In either case, the strength falls much below that of the parent material. With increasing distance from the melt boundary, the grains become finer until the heat unaffected zone with elongated grains is reached. All these changes are shown in Fig. 5.16.
Let us now consider a two-phase material which derives its strength mostly from precipitation hardening. In this case, the strength within the melt boundary is again too low. But, in the immediately adjacent heat affected zone, the thermal cycle results in heating and quenching followed by further aging. This aging process recovers some of the strength. The material beyond this zone is only over aged due to the heat of welding and becomes harder with the loss of
The two examples we have considered clearly demonstrate that various types of metallurgical changes are possible during welding, particularly for complex alloys. These changes are governed by the non-equilibrium metallurgy of such alloys, and must be clearly understood to yield a satisfactory fusion weld. Also, a decision on the post welding heat treatment to be given must be taken to restore the desirable characteristics of the joint.
Weld Defects and Inspection:
In a fusion weld, the defects often found include the lack of fusion, lack of penetration, inclusion of slag or oxide, presence of cracks, porosity, and undercut or excessive penetration (i.e., bad profile). These defects are shown diagrammatically for butt welds in Fig. 5.41. Such defects do not significantly alter the static strength of a welded joint under ductile conditions.
However, the defects have serious consequences if the material runs a risk of brittle fracture or the joint is subjected to a fatigue loading. The presence of a crack always enhances the probability of a brittle fracture. Similarly, a lack of fusion causes a sharp discontinuity which, in turn, diminishes the fatigue strength.
Various standard tests are conducted to ensure the acceptability of the strength and quality of a welded joint. The strength is checked by the tensile and bend tests. Figure 5.42 schematically shows three different bend tests, namely, free bend, guided bend, and controlled bend.
The bend test is conducted with two specimens; in one specimen, the outside face of the weld is put to tension and in the other specimen, the root is subjected to tension. The bend test is difficult to conduct on very thick plates. In such a situation, the side bend test is performed with a slice, 3-6 mm thick, cut at right angles to the plate surface and to the weld axis.
The side bend test reveals a lack of side fusion very clearly but is less sensitive to a root defect than the normal bend test. The impact strength of a joint is tested by the usual methods after providing a notch in the joint.
Nondestructive inspection techniques are used to locate the internal defects and minute cracks in a welded joint. For the details on such techniques, viz., radiographic examination, ultrasonic testing, magnetic particle test, dye-penetrant testing.