Dissimilar Welding

Dissimilar Welding


Welcome to the lesson on dissimilar welding.
In this lesson, we would be defining what we mean by Dissimilar Welding.
We will look at various aspects of dissimilar welding, namely the metallurgical, the thermal,
the fluid flow the mixing aspects. And then we will also in the end highlight some of
the issues where research has to be done. Let us first define dissimilar welding as
follows; we will be using the same terminology as we have been using till now in various
lessons in this course. Let us say we have a domain, in which we would
like to join two types of materials. And if we have the two pieces of the same material,
and if you have, for example, the fusion zone in which we added a filler which is heterogeneous,
namely it is of a different material let us say filler B, then here you have B plus A,
and so this is the situation where you have A plus B melt which is in contact with a base
material with A. So you could think of this as one very simple situation of dissimilar
welding. And then as we proceed further you may have
a situation, where you have the second part not as the same of the first part. And you
may have B and then filler can be also different from that. So we will in which case you would
have for example, melt which will be A plus B plus C melt, and in addition to being in
contact with A, it is also in contact with B. And in a situation which is somewhere in
between namely autogenous welding between these two, you would have a situation like
that. So you can see that these are the various scenarios that would be considered under the
broad subject of dissimilar welding, namely a melt that is different composition from
the two base materials. And a melt that would be having for example,
filler also, so you could have basically three different materials coming in a weldment together
that is the reason why you would call it as a dissimilar welding. Now this definition
can then be seen to extend the discussion in two even other related processes. For example, let us take brazing and surface
alloying, and we have cladding, and even a newly emerging subject called additive manufacturing.
In all these things, you would be basically encountering situation that is very close
to dissimilar welding discussion as follows. In the case of brazing, you normally have
a requirement that the filler, this filler is going to be different; it is going to be
molten when A and B are not molten yet, so this is the situation where you have a dissimilar
joining that is possible. Alloying is a situation, where you have on
the surface of a material, a molten pool that is created and you want to add a particular
element into it. So you would like to add B and an element B into a base material A,
so that the surface layer would have for example, enhanced corrosion resistance or enhanced
abrasion resistance etcetera. So you have situations where you want to alloy only the
surface and that would also constitute a geometry which would be similar to what we have discussed
in the dissimilar joining. Cladding is the situation, where you are going
to have a weld over lay; in such a situation, where you want to deposit a material B on
the top of a surface of material A, so that you would like to have a surface that is having
better properties again for perhaps oxidation or corrosion or abrasion resistance etcetera.
You normally would not be able to distinguish between alloying and cladding as two separate
processes, because most of the ways by which you can deposit these are the same processes
called as weld over lay. So, you would have a situation actually in
this manner, where you have some amount of alloying and some amount of cladding. And
more alloying you would be coming into this situation, and more cladding you will be coming
to this situation. And what is the parameter that would distinguish between these two,
it is basically a parameter called as dilution that is basically how much of thickness of the material has been alloyed when you want to actually clad
and that parameter you could call it as the extent of dilution during this processes.
Additive manufacturing is a process whereby you may want to build a part, and this part
may be then made by sequential deposition. And in this case, I want to make, for example,
a cylindrical tube on the surface of a plate, and this is made by layer by layer deposition
where each layer is basically like a circle, and you build the material on top. So, you
can see that additive manufacturing encompasses in each layer what is happening for example,
in a cladding process. So, in that sense, you can see that in a variety of other alloyed
processes also, you would have a scenario where dissimilar welding is being discussed
and so the wide applicability of this subject is now evident. We would go further by looking at various
aspects of dissimilar welding as follows. So dissimilar welding is practiced in multiple
geometries, we would look at a very simplified geometry as follows. You normally would encounter
butt geometry or lap geometry. And you will also encounter this lap geometry in the different
geometries in these manners also. So you will have situations where you want
to join a pipe to another pipe or plate to another plate or a thick plate to another
thick plate, in all these geometries you can actually visualize what would happen in a
small region and then you would see that you can always think of that as just a pair of
two materials that are being joined. So you can actually take the case of butt geometry
and then analyze the process and then see that the physics is happening in a very similar
manner in the rest of the geometries also. So we would limit our discussion to butt geometry
for this lesson, because that would illustrate adequately what we need to understand in dissimilar
joining. Now, what are the various aspects of dissimilar
joining that we are going to cover there as follows. Basically, what is happening is that
you have basically a heat source that is going to be present to melt. So you would have this
is going to affect heating; and then we would have melting, and then you would of mixing
of those molten regions. And after the mixing is happening, then when the heat source is
moving away you would have solidification of the mixed region that is in between the
two different materials, so you would have basically solidification.
This solidification of this melts is different from normal welding, because the solid part
over which it has to grow is not the same as what is the molten region. So, it is not
same as for example, the melt is not the same as the solid from the where grains can grow,
so that is the very different thing. And after the solidification is over you have
usually the metallurgical processes such as precipitation of a various phases that would
be coming because of the alloy that you have chosen and then different morphologies of
those phases as they evolve. And then the stresses because of any shrinkage of the different
phases, the two different extents as the cooling are happening. So these are all the various
stages how a dissimilar weld is going to form and come up to room temperature. And then
in this process, there can be many things that can go wrong at various stages, so the
success of a dissimilar weld it depends very strongly on each of this processes happening
the way we would like to have. So, let us look at the heating process during
dissimilar joining. For illustration, we would take butt geometry and put two materials side
by side, and then look at how the heating could be different during the dissimilar joining.
The first thing that we must know is the placement of the heat source earlier when you are joining
the same material, the placement is with respect to this center line and that would not be
playing a very significant role as long as the air gap between them does not alter the
thermal characteristics. However, in this case, now you have got two materials that
are different, so the placement of heat source is very important. So, the placement of heat
source can be in the following manner, it can be symmetric or asymmetric.
Why would you like to consider the case of asymmetric placement, it is for example, when
the melting point of A is for example, very, very different from that of B. So if a melting
point of A is a very large compare to that of B, then what would happen is that if you
were to place the heat source towards the B side then B would melt a lot, and before
even A would start melting, and you may have a situation where you have a braze welding
kind of a process. So, you may want to choose a variant of the dissimilar welding by the
choice of the materials that are at hand, so the heat source is not always symmetrically
replaced. If you want to do braze joining, you may want
to place it on the side where middle is going to be molten faster, so that that is what
is being brazed on to the relatively refractory metal which is on the A side. Otherwise, normally
if you want to look symmetric side, symmetric placement of the heat source, then we will
see what will happen. So, let us say symmetrically placed. If it was symmetrically placed then what would
be changing. Essentially, you have to see that you have got two different materials
in what way are the different. So very often if you were to take two different alloys that
are not very different in their composition, let us take for example, two different carbon
steels with a small change in the carbon content then the thermal properties of A and B are
not very different. So, as for as heating is concerned, then they are not dissimilar
adequately dissimilar enough; however, you may have a generic situation, where these
two can be having very different thermal properties. And the most important thermal property which
would distinguish the heating process between them is the thermal diffusivity. So, it is
basically ratio of thermal conductivity, the density and the heat capacity of that material.
What would happen when you have this property very different? Let us take the case of alpha
A, for example, more than alpha B. What happens when the thermal diffusivity of A is large,
what would happen is basically the same amount of heat is placed on both sides, but the heat
is placed on A side would actually be getting removed much faster. So, relatively the same
amount of heat is going away slower on B side and faster on A side. This would have an implication
that the location that is far away from the center line is going to get heated up faster
on the A side than B side, which would mean that is the gradients, so the thermal gradients
on the A side are going to be shallow on the B side is going to be steep. In other words,
you could think of this. So the material which as poorer thermal conductivity
is going to experience sharper temperature gradients that is because you have basically
different amount of the diffusivity of the heat on both sides and this is going to result.
Now, this is the first thing that would be changing the way the molten region is going
to form. Now it is not obvious that just because the thermal diffusivity is different, the
molten region is not going to be the same way. There is a discussion that is required
before we see which one would melt. And let us just take the following example.
So, as you can see that here as the heat is being applied the heat is accumulated on B
side more than on A side. So, if we have a situation that the melting point of A versus
the melting point of B; case 1, if they were very close by, if they were very close by,
then you would see that on the B side the heat is not going away fast so it is getting
heated up faster at the central line. And therefore, you would see that the melting
may initiate on the B side. Now, let us take the case 2; let us take A
to be less. Now if A were to be having a lesser melting point, it is not obvious that A would
melt first. The reason is as follows; there is a competition between the heating rate
and the melting point. If the melting point difference is small, if it is small, by small,
what we mean is couple of 100 degree centigrade, and then you would see that if the heating
rate on the B side is faster, then B would still melt though the melting point of B is
higher. So, it is not obvious that what is lower melting would actually melt earlier;
it could also be that what is higher melting can melt earlier when its thermal diffusivity
is quite poor. The competition between these two can be resolved
only numerically when we simulate the process of heating and melting on both sides by giving
the same amount of heat. And then seeing which would melt fast, because this is a competition
of how much of heat is removed versus how much of heat is available for it to rise upwards. So, the once the materials have reached to
the melting point, then you would see that there will be a competition of the melt formation
also on both sides. And we would see that unlike in similar welding, you may not have
the same way of melting on the both sides, by that what we mean is as follows. Let us
take one of the cases for our analysis. Let us take the case of whatever you have written,
so that a melting point if A is slightly lower, but melts later. You can have that situation
and what would happen that is illustrated as follows. You can have a situation here,
then that the contours are going to be widely separated here and narrowly separated here.
These are all basically temperature contours. And you would see that the gap between them
is a measure of the temperature gradient, and wider gap would mean shallow gradient
and lesser gap means steeper gradient. So if you have A and B, and alpha of B is less
which would mean that you have the heat being removed bit slowly on the B side, so you would
see that heat is accumulated on B side. So, it would actually mean this would actually
mean this kind of a situation would mean that B melts first. If B were to melt first, then
you would see that the molten region would appear like this. The first region that would
melt is like this. And the heat is still being applied symmetrically on top.
Now you can see that for the A side, you have got two modes by which melt can initiate.
For the A side to start melting, it has two modes; one is direct heat from the source,
and another way of heat this is first direct heat what is coming down from the heat source
that is one way of making the A side melt. There is another way that is from the liquid
that as formed on the B side so that is basically convective heating from B side. So, I can
actually explain that by exaggerating this geometry to indicate this phenomenon, basically
let us make this very big so A is not molten B is molten and we have got on the B side,
so some kind of an advection is always there in the melt pool.
So you can see that the heat is being brought towards A side by the advection within the
molten pool, so as it would grow and you can see that heat coming laterally is also aiding
in the melting of A. So apart from what is heat is coming from the top, you also have
lateral heat coming in and therefore, from the A side the melting is actually in two
fold. And because it is a molten B that is coming
in contact of A, then you can say that the way A would melt is not by melting directly,
it also is by dissolving into B. So, you also have possibility that this may indicate possibility
of dissolution, so that it can start melting and mixing. So you have got this kind of a
twofold mechanism for the A side if the thermal conductivity of A is high because of which
it would melt later on. Now, once the molten regions are available
how do they mix? So, this is possible for us to analyze as follows, the mixing process
can be analyzed as follows. So, mixing is only when we actually have both the sides
molten to some extents, so the initial point for mixing let us take the situation, where
basically you have got some amount of A and some more amount of B molten. This entire
thing is basically liquid. Now you know that due to various driving forces
namely the buoyancy, the Marangoni convection, and also depending upon the heat source, the
electromagnetic forces, there are various driving forces because of which the molten
pool would be having an advection and that advection would tell whether or not A and
B would mix. So, how any two materials would mix, we can
analyze as follows. You can actually think of mixing as a process of stretching and folding
followed by diffusion what I mean by that is as follows. If you have a region, let us
say a small amount of B is being advected into the A side then you would see that this
is going to be stretched. So what I mean by stretching is that like this; and because
of the advection, this layer is then going to be folded so you would then see that.
Now you see that the thickness over which this is A, and this is B, this is b, this
is A. The thickness over which the diffusion as to take place is now shrunk, and if this
were to proceed further, you may have a situation like this, so that the thickness over which
the diffusion needs to take place between A and B is shrunk as the stretching and folding
is going on and this stretching is mainly because of the velocity gradients that are
present in the pool. And the folding is because of change of velocity directions. And this
process would mean that at some point the length scale may be small enough length scales
may reach dimensions which are small enough small enough for diffusion.
Small enough for diffusion in the sense, you know already that the solutal diffusivity
is generally very poor it is about three to four orders of magnitudes smaller than thermal
diffusivity. So, normally we do not want to think of ability of mixing between A and B
over such a large length that is not the fact. But when the layers of B are going to penetrate
into A and then are going to be stretched and folded you would have a situation where
the distance over which the mixing is suppose to take place is very small.
And then it would happen if the diffusion would permit because length scales are now
small. Now would they still mix or not is depended upon actually a metallurgical reason
so this is subject to the kind of phase diagram between A and B, so are the elements A and
B or the alloys a and b miscible or not is something that we can deduce from the phase
diagram, and that is what is going to tell us whether after even this much of process
would they still mix or not. And we can come to the categories of those
alloys as follows. We have got the categories as follows A and B fully miscible, you may
have them partially miscible. And usually when they are partially miscible it also implies
that you have compound formation. And these are the situations where the dissimilar welding
is struck off as a not feasible; fully miscible is a situation where you can say that dissimilar
welding is possible and there is no problem in joining them. And the other extreme is
fully immiscible and you then would have two more possibilities it is immiscible, but without
compounds and with compounds formation. And you could see that the situations which
are very straight forward for us to have joining process taking place, you can say that in
this situation it is very possible and again in this case is very much possible. And you
always have a problem when these are the situations where whether it mixes little bit or does
not mix at all and then you have a compound formation. And how these compounds would form
between the layers of those mixed stretched and mixed layers, where they are going to
form whether they are actually going to withstand the kind of stresses thermal stresses during
the cooling of the weldment. And if they do not with stand that is when we are basically
have the dissimilar joining failing. So, basically you can see that metallurgical
compatibility must be looked at. Now some examples are suitable at this point. We can
say that situations like copper nickel or let us say gold nickel etcetera, where you
have fully miscible phase diagrams, and therefore, there is no problem in joining them. And you
also can think of fully miscible regions as two different alloys of slightly different
compositions, so they are also can be thought of as fully miscible.
Let us take for example, two steels of slightly varying carbon content or two aluminum alloys
of slightly varying copper content or silicon content, so such situations can be thought
of as fully miscible situations. And you also have fully immiscible situations without compounds
for example, iron and copper or for example, steel and brass, so you have those situations
coming here and you have the remaining cases for example, between important technical alloy
categories such as super alloys and aluminum alloys, titanium alloys and aluminum alloys
or super alloys and titanium alloys. So, you those common agents, where you have got later
mental compounds that are forming, and you normally have difficulty in dissimilar joining.
So you can see that after the mixing process is over you can see whether the diffusion
will take place or not. And you can see that the diffusion will takes place in these two
regions, and it will not take place here. So it is not obvious that mixing is actually
going to happen or not. You may a have a situation where mixing is not at all happening in spite
of stretching and folding; or mixing is happening, but fully.
Now looking at the extent of mixing possibility, we can choose what kind of modeling technique
is possible. And we normally will have several types of modeling techniques that are possible
whenever you have two different liquids A and B, what a kind of modeling technique is
suitable can be decided based upon this kind of a analysis looking at the phase diagram
of that relevant elements. So, let us look at what are the modeling techniques
available, whenever you have this kind of a situation. So the techniques available are
as follows. The first technique is basically the assumption that these two liquids are
going to be fully miscible in each other, and therefore, there is no problem in considering
the both the fluids A and B – the molten A and molten B as fully miscible. So, you may
have a situation of mixture models. In other words, the properties of any control
volume are taken basically as an average mixture of the properties of A and B knowing how much
of A and how much of B are mixed in that particular location. So, in other words, you can actually
go ahead and extend the control volume formulation, which we have discussed in the last several
lessons; and we can basically replace the properties with mixture properties and then
go ahead and use. The other methods that are available are as
follows. You have for example, volume of fluid approach, this approach is where in
the control volume you keep track of how much of volume of each of the two phases are present,
and then try to conserve the total volume and advecting each of those phases across
the control volumes. So, you basically you have conservation taking care quite well,
and you can actually handle situations of partially mixed or immiscible regimes using
this kind of an approach. So, level set method is again one more approach, where you basically
think of a parameter phi which is basically the level set parameter. It has a similarity
with respect to what is called as an order parameter.
So, it is like an order parameter, which is telling you where the interface, and interface
between the two phases alpha A and B is, which are molten regions. And you can set saying
that when phi is equal to 0.5 or phi is equal to 0 is the interface between these two liquids.
And this is actually very good method to track the regions whenever they are not being mixed
very well. And the last method of course, is what is called is the front tracking or
two fluid approach. So, in this approach, we basically populate the domain with particles
or marker locations and then we want to track where the interface between the two regions
is. And this is very suitable in situations, where
you have completely no possibility of mixing. And it is also quite complicated, because
you normally will have to handle how the interface is going to split the domain or join two domains.
And that would require lot of book keeping with respect to the location of the interface
re-meshing etcetera. So, we have basically a variety of methods that can handle a variety
of mixing possibilities. And what is very suitable for dissimilar welding
for metallic materials particularly the kind of alloys that we normally encounter in structural
engineering scenarios, basically we will have requirement only to look at these approaches.
The reason being, that most of the metallic materials that we are looking at are either
fully mixing with each other, or partially mixing with each other. So you normally do
not have situations where they are totally not at all mixing with each other. So you
may not require to go to the situation of a front tracking methods at all.
You will have a situation when you have a inter layers that will not be mixed to a large
extent, so in the case of inter layers, you may want to use level set method to know the
location of the inter layer, if it was to melt and move a little bit away from its original
location. So, you have situations where the mixture model is going to be the used maximum
and front tracking or two fluid approach is used minimum in the context of a dissimilar
metal joining the modeling aspect as for as the metallurgical scenario is concerned. Let us go to some additional details of the
mixture model, because that is what we have seen as a very suitable for dissimilar joining
for most of the metallic materials. So, we will go to some more details of that what
we mean by that etcetera. So what are the things that you are going to be mixed? So,
enthalpy for example, you could always take it as a fraction, so you could think of the enthalpy as a mixture
quantity, if you have A and B that are being joined then the enthalpy of A and B can be
thought of as mixed within the controlled volume. And the fraction of A and fraction
B that is present in that control volume can be taken and then you can average it out,
and you can say that f A and f B sum of it is 1. So, therefore, instead of f B, I can
write 1 minus f A. Now this kind of a method would actually also
see whether the enthalpy change at the melting can be taking it as the same way. Can we write
the melting point also like this or rather 1 minus f A, can we write like this? The answer
is no, the reason is as follows. If you take the phase diagram of elements A and B, and
let us take the melting point of A is here, and the melting point of B is here, then this
kind of a mixture model assumes that the melting point of any region in between is going to
just an averaged value between the two. So, it means that the melting point of any
alloy in between is given by this law. Now this is basically Vegard’s law, which need
not be applicable for engineering applications. And you can see that even a very well mixed
system like copper-nickel would show that in between you would have situation like that,
so you will have a freezing range that is coming up and the melting points are not given
by the straight line approximations. So, you can see that already we can see that
we have a deviation from the mixture model that we need to account for, so one should
not blindly put this for everywhere. It is also not true that we need to apply the mixture
model in the entire domain upfront, because during the melting you have a situation that
what is melting is only a pure material. So, you can say that during the heating and melting,
you can say that you do not have to apply the mixture model. So, you could use the pure
metals properties on either side and really the mixing is happening only after their molten,
so you can actually see that only for the liquid domain you start seeing the mixture
model is required, so up to the point of melting you do not have to apply.
So, you can actually delay the melting point averaging until the point that both sides
have some amount of molten region; and after that, you can actually substitute the averaging
of melting point with the phase diagram information, if you have it readily and then use that information
instead of averaging the rest of it. So, similarly the other parameters also should be averaged
carefully such as for example, the thermal conductivity, or heat capacity etcetera. So,
we always need to look up the data bases for information about the property averaging whether
it is correct or not, but otherwise by default, if you use a mixture model like this then
you would get a numbers that are not very far away. So, once we have used this mixture model to
start looking at the mixing of the two liquids, then you would encounter a situation as follows.
You have basically the regions that are molten A and B to different extents. Now we already
said that in the case study we have taken B has a lower thermal diffusivity, and it
meant that B melts first. And you already saw that we took the example where the melting
point of B is higher, so which means that as you are cooling B also starts to solidify
last. So what this means is basically this region
is spending less time in molten state, and this region is spending more time in the molten
state. So even when we assume that the stretching and folding, and the diffusion process is
taking place, you saw that that process is actually given more time on one side, and
less time on the other side. And what is the conclusion from the amount
of time that is available for stretching and folding and diffusion, even assuming that
they are both fully miscible, if it has less time then basically the gradients are sharp.
So you can say that this would imply that the solute gradients are steep, and this would
imply that the solute gradients are shallow. So, clearly we can see that already based
upon our discussion that when you have two different materials that are being joined,
and then we have thermal properties that are different, we already see that they both may
not melt same time. And if they were to melt in such a way that one of them is melting
earlier, then you may have a situation where that region may spend more time and then that
is also leading to different gradients. Now, once the heat source is moved on, and
this region is now going to solidify. This region when it is going to solidify, how does
it happen, does it happen the same way as we have looked at in the welding of similar
metals, does it solidify the same way, namely the grains of A or B side can they grow straight
into the melt. Now that is where actually interesting thing is possible. Essentially
the problem when you have dissimilar welding is now converted to a problem where you have
got the solidification of the melt pool in the presence of strong composition gradients.
So, it is not the same as in the similar welding, where you have got the melt of the same compositions
as the solid, and then solid is growing into the melt; and then as long as there is a heat
removal then the solidification will proceed. But this is not that case, you have got a
situation, where the solid is having a different composition from the melt and that melt is
not having the uniform composition, it has a very steep gradient. So, under what circumstances
such a situation would actually lead to solidification or is it possible that the grains may not
grow fully in, and it may require nucleation of fresh new grains in the melt, so that you
may not have actually a continuous growth of the solid. So you will have those scenarios
coming up. And let us just take two very different kinds
of possibilities and analyze, what is going to happen here. So let us take the case as
follows. The case 1, first case is taking the phase diagram to behave in this manner.
Let us take that it will have this kind of a possibility. And let us take the temperature
for our analysis as left. And if this was the case then how would the scenario look
like. At this scenario, if you want to look at the
so called Gibbs energy versus composition plots, more information about this can be
obtained from physical metallurgy text books, but for now you can just look at what I am
analyzing. You can see that the temperature T i is where these two compositions are at
equilibrium, so this is basically the solidus composition and this is the liquidus compositions
that are in equilibrium with each other as solid and liquid. And you can see that that
is actually example is given by this situation, the solid composition and the liquid composition
that are in equilibrium, so that there is a common tangent construction that is possible.
And how is the region that is looking like here, if look at this region then you have
a situation where pure A is in contact with alloy melt, this is the situation. And you
see that what is in contact with alloy melt, alloy melt, let us take the composition of
alloy melt of C l and let us take the temperature to be T i. Then you can see that what is in
equilibrium with an alloy of the composition C l is thermal numerically a solid of a composition
of C s, but then what is actually in the welding scenario is actually pure A. Which means that
at this location composition should reach C s from that of pure A so that local equilibrium
is achieved. So, you can see that what is in equilibrium with a liquid melt in the melt
pool is having a composition of C l. So, what is in equilibrium with that kind
of a melt pool is actually a solid of composition that is may be for example, 4 percent or 5
percent of B for example. And that is not in contact with the melt actually, what is
in the contact with melt is pure A, so you can see that if the pure a composition can
change by diffusion, so this thing can through diffusion. If it can change through diffusion
to a value C, s and then locally it could be miss possible after that if you then try
to bring the temperature down, then the solidification can take place as we would normally understand
in the physical metallurgy. So, you can see that the growth of grains
from pure A in to alloy melt is possible, but only after some amount of diffusion is
taking place. So, how the micro structure would be, let us draw like this A; we are
looking at these grains. Can they grow, we see that, if they want to grow then there
will be region here that I would like to expand A, the composition in this region is gradually
enriched from pure A to alloy of A plus B. So, as the grains of were trying to grow into
the melt, you have to expect that the composition locally has to change, and then only the growth
is possible. Now such a change in the composition is possible provided the temperature as come
down, so with the small amount of under cooling you can expect that this is possible.
And therefore, you can expect that on the side of the dissimilar joining where the partition
coefficient K C s by C l. The partition coefficient K is less than 1; that is, when the phase
diagram is going downwards. So on that kind of a side it is possible for us to imagine
that through a small amount of diffusion at the interface the local equilibrium can be
achieved following that the grain growth the growth of these grains into the liquid can
take place to complete the solidification. So, you may expect a micro structure very
similar to that of a similar welding on one side that is one case that we are looking
at. Let us look at the other case where this is
not possible that is case 2. For case 2, we then take the phase diagram. In the other
way, and we would then have a situation like this. And this a melting point, and this is
T i, so you would have a phase diagrams like this. This is composition.
And you have a situation where the partition coefficient is greater than 1 and; that means,
the phase diagram is going upwards, there it is going downwards; both are possible in
various alloy categories. And you have a situation where exactly the same kind where the B is
now facing it the liquid. And we are now seeing whether it is possible for this fellow to
grow or not ok. Now, what is facing B, you can say that pure B is facing alloying melt.
And let us take that alloy C l like this. Now you can see that what is in equilibrium
with C l is actually a much enriched form of B with lot of A, which is having a composition
of C l, and you can see that B is actually molten you can see that is liquid that is
in equilibrium with that kind of a thing. So, you can see that until you reach this
composition C s star, you do not have a situation where the solid is actually having any stability
with through the liquid. For all the compositions where the pure B
is going to get enriched with A, you see that it is only the liquid which is stable only
after it as crossed this composition. So, you can say that C l or C s only when you
cross the composition, you can see that solid is stable compared to the liquid and solidification
can start taking place. So, in other words, you can see it is not possible for B grains
to grow, because it is it is possible only beyond a very high amount of composition.
So, you can see that this growth that is direct growth, direct growth of grains of B into
melt not possible. Now if B were not able to solidify then how
would the liquid on that side freeze, it would happen basically by nucleation of solid in
the region, and that would actually normally take some amount of under cooling. Which means
that on the B side you will not see a micro structure similar to the similar welding situation,
where the grains are growing into the melt instead you would see that this interface
is only going to be what is molten and then the grains here will not have any relation
with the grains of B. So, you can see that you have very, very different situations.
And you also saw that on the B side, you will have perhaps more amount of gradients and
you also see that more amount of unmixed region is possible on B.
So, you can see that in a dissimilar case, the way the melt pool is going to solidify
can be very different on both sides; on one side, it may be achieving the micro structure
similar to the similar welding; on the other side, it may not able to take place at all.
And that is reason is basically motivated from partition coefficient being less than
one or more than one and looking at the situation of a pure metal in contact with alloy melt
in the presence of a gradient. So, at this juncture then we would then see
what would happen further. As the solidification is then taking place whether directly from
the melt or not, then this is the situation that would happen. And you have got the region;
this is the alloy of A, B and may be even C, because some addition of C is possible.
Now you would see that on one side, it may be possible for the grains to grow in; and
on the other side, it may not happen at all, and you see that some other grains are forming.
And you would see that as this is going on, you would have also situation where the composition
is gradually changing. And you will have a region where the composition may be such that
certain inter metallic and form. And what would be the shape of such inter metallic
region; it would be the same shape as that of the fusion zone. And in a longitudinal
section, it would look like this. You would see that there will be regions that would
be containing different inter metallic; and these inter metallic will be basically coming
from the phase diagram which ever are all possible.
So, if you have a phase diagram like this. So, you have some inter metallic here this
beta for example, is appearing when you have composition that is enriched, so you may have
regions containing B appearing in this kind of a passion located in a region that is basically
following the shape of the fusion zone. And these regions are going to be subjected to
stresses, because of the shrinkage of the material as the temperature is brought down
and whether they are able to withstand the stresses or not will tell whether the cracks
will actually form or not. So, a lot of discussion is possible when we
look at what are all the components that are going to form, we will not do that because
that is specific to a particular alloy category. But at this juncture we would say that there
are so many things that would change from the point where you have got the heating,
and then later the initiation of melt, and then you have got the mixing, and then the
solidification. At each stage there are different things that
will be happening compare to the similar welding situation and therefore, you may expect the
final result to be either successful or not depending upon the properties of the kind
of compounds that can form and also the properties of the rest of the two materials that are
being joined. At this juncture, I would like to close by giving some remarks about how
we proceed to understand further. So what we planned to do is as follows. It
is possible to understand how this entire processes, namely the heating, melting, mixing,
these thermal processes can be understood through simulations. So, what I planned to
do is following the two part lesson on numerical simulations to thermal field, and fluid flow
in welding, I would like to show you some slides on how the evolution of a thermal field
and fluid flow will take place. And then we will also then combine with this
lesson to show, if the material properties were to be different then how those profiles
would look different when you have these three stages. So, we would actually combine them
to make a slide presentation shortly in a later lesson to show how the difference will
be happening. And regarding solidification, we will actually be able to see the difference
only when we look at the micro structures. So, few microstructures can also be shown
to see the difference of solidification on both sides when dissimilar welding is happening.
Now, how close is the analysis that we are applying to the alloyed problems such as brazing
or brazed welding or cladding or there is surface alloying or additive manufacturing,
this is a problem that is to be seen, and we will take that as a separate lesson where,
we can look at only a geometrical variation; and rest of the discussion, wherever you have
got the difference of properties playing a role, then we will refer back to this lesson
where we have discussed them in detail. With that, we will close this topic of dissimilar
welding, and then we will continue further with simulations of dissimilar welding for
illustration purpose. Thank you.

2 thoughts on “Dissimilar Welding

Leave a Reply

Your email address will not be published. Required fields are marked *