06-16-2012, 10:41 AM
(This post was last modified: 06-16-2012, 12:57 PM by Rickabilly.)
As promised, I am writing this to explain the mechanics of metal warpage during welding and other heat processes.
Obviously this is a really deep topic and I could write a scientific paper on this that even I wouldn't really understand, which would be pointless, so instead I am aiming at welded joints in welders terms leaving out as much math as is possible to transfer the basics of what happens in a weld that causes shrinkage and how you can reduce the effects. I am keeping it limited to three examples of butt welds and nominating three temperatures
TL, the temperature at which the metal flows in the pool,
TS, the temperature at which the pool has just solidified,
and
TR, Room temperature.
And further to this I will simplify the weld moving across the joint by comparing a moving weld pool to separate pools in a line Pool1 (P1), Pool2 (P2) all the way across the joint to say P10.
Example A, Butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, with in the first place a 1/16" gap and a full weld across in one go with no tacks.
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, but, no warpage occurs as the shrinkage is not resisted at all.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking but not at the same rate P1 has cooled much faster than P2 will and so is shrinking faster applying a compressive force to P2. At P4, P3 is solid and cooling/Shrinking more slowly than P2, which is cooling more slowly than P1, so P1 is adding a compressive load to P2 and so P2 is loading P3 and so on across the weld at about P6 an equilibrium is reached where each successive P# is cooling at about the same rate as the next P# but due to the fact that P1,2,3,4and5 have all cooled at different rates the compressive forces applied to successive p#s has warped the bar, as the highest compressive deformation has occurred on the P10 side the ends of the bar move towards the P10 direction, these compressive forces were not significantly resisted until the weld was half way across the bar and so the bend will be significant.
Example B, the same as A butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, with in the first place a 1/16" gap and a full weld across in one go, but with a single strong tack at P10
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, as P1 shrinks the tack at P10 resists the shrinkage causing the bar to bend at the weld slightly the ends of the bar move towards the P1 side.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking but as P1 is cooling quickly it develops compressive strength more quickly than P2 develops compressive forces against P1.
At P4, P3 is solid and cooling/Shrinking more slowly than P2, which is cooling more slowly than P1, so P1 is now supporting the compressive forces applied by the cooling of subsequent P#s.
and so on, at the end of the weld the tack at P10 is remelted and it cools to TS shrinking slightly at this point the temperature across the bar is almost uniform except the P1 and P2 which are still a little cooler, as P7,8,9 and 10 cool the warped bar tends to return slightly, depending on many factors the bar could bend slightly either way, more likely towards P1, but is very unlikely to be perfectly straight or as warped as example A.
Example C the same as Example A; Butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, except with no gap and no tacks,
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, but, warpage occurs at a high rate as the shrinkage is resisted by the solid cool(and so compressively strong) pre weld metal at P3.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking, P3 is now liquid but P4 is cool rigid steel and will not compress so the extra compressive forces from P2 are applied to P1 which is relatively weak compared to P4, The bar will bend heavily towards P1 and continue to do so until P10 is melted as P10 cools to TR the bend will recover slightly but this example will no doubt show the greatest warpage due to weld shrinkage.
So from this we can see that in simple welds simply maintaining a Small gap between the weldments and using a solid tack at the end point of the weld we can dramatically reduce warpage.
Following the examples A to C it can be derived also that the amount of warpage is related to the difference in temperature gradient across the bar from the unwelded parts to the welded parts, for this reason if the weld area is preheated, the difference in cooling and therfore shrinkage rates will be smaller reducing warpage,
A further point is that welding processes that have a stronger energy gradient (more heat in a smaller spot) tend to increase warpage, so TIG with a very high heat in a tiny spot can cause massive warpage where flame welding which in it's own right preheats the metal as a result of it's weak energy gradient causes less warpage, MIG and Stick are about even but MIG has the advantage of blowing a cooling gas onto thinner work pieces which has an affect.
If folks find this useful or interesting I can go through any number of examples that demonstrate the mechanics in different joints, but I do accept that this is very difficult to visualise for some and so rather than waste time I will follow your lead. Please take part in the poll so I know what you want.
Best regards
Rick
Obviously this is a really deep topic and I could write a scientific paper on this that even I wouldn't really understand, which would be pointless, so instead I am aiming at welded joints in welders terms leaving out as much math as is possible to transfer the basics of what happens in a weld that causes shrinkage and how you can reduce the effects. I am keeping it limited to three examples of butt welds and nominating three temperatures
TL, the temperature at which the metal flows in the pool,
TS, the temperature at which the pool has just solidified,
and
TR, Room temperature.
And further to this I will simplify the weld moving across the joint by comparing a moving weld pool to separate pools in a line Pool1 (P1), Pool2 (P2) all the way across the joint to say P10.
Example A, Butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, with in the first place a 1/16" gap and a full weld across in one go with no tacks.
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, but, no warpage occurs as the shrinkage is not resisted at all.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking but not at the same rate P1 has cooled much faster than P2 will and so is shrinking faster applying a compressive force to P2. At P4, P3 is solid and cooling/Shrinking more slowly than P2, which is cooling more slowly than P1, so P1 is adding a compressive load to P2 and so P2 is loading P3 and so on across the weld at about P6 an equilibrium is reached where each successive P# is cooling at about the same rate as the next P# but due to the fact that P1,2,3,4and5 have all cooled at different rates the compressive forces applied to successive p#s has warped the bar, as the highest compressive deformation has occurred on the P10 side the ends of the bar move towards the P10 direction, these compressive forces were not significantly resisted until the weld was half way across the bar and so the bend will be significant.
Example B, the same as A butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, with in the first place a 1/16" gap and a full weld across in one go, but with a single strong tack at P10
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, as P1 shrinks the tack at P10 resists the shrinkage causing the bar to bend at the weld slightly the ends of the bar move towards the P1 side.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking but as P1 is cooling quickly it develops compressive strength more quickly than P2 develops compressive forces against P1.
At P4, P3 is solid and cooling/Shrinking more slowly than P2, which is cooling more slowly than P1, so P1 is now supporting the compressive forces applied by the cooling of subsequent P#s.
and so on, at the end of the weld the tack at P10 is remelted and it cools to TS shrinking slightly at this point the temperature across the bar is almost uniform except the P1 and P2 which are still a little cooler, as P7,8,9 and 10 cool the warped bar tends to return slightly, depending on many factors the bar could bend slightly either way, more likely towards P1, but is very unlikely to be perfectly straight or as warped as example A.
Example C the same as Example A; Butt weld of two bars 4"(100mm)x 40"(1000mm)x 3/8"(10mm).
These will be welded end to end to make an 80"(2000mm) long piece,
the welded edges will be ground to a 1/4" x 45 degree weld prep, except with no gap and no tacks,
1, strike the arc , creating a pool P1 at TL and proceed to weld,
2, as the pool moves and becomes a second pool P2, P1 cools to TS as this piece is solid and is cooling it reduces in volume according to a linear scale say 1unit of volume per degree until it reaches TR and as the whole bar except for the welded bit is at TR the heat is rapidly removed from this first bit of weld, but, warpage occurs at a high rate as the shrinkage is resisted by the solid cool(and so compressively strong) pre weld metal at P3.
3, as the pool moves again to P3, P1 and P2 are solid and so shrinking, P3 is now liquid but P4 is cool rigid steel and will not compress so the extra compressive forces from P2 are applied to P1 which is relatively weak compared to P4, The bar will bend heavily towards P1 and continue to do so until P10 is melted as P10 cools to TR the bend will recover slightly but this example will no doubt show the greatest warpage due to weld shrinkage.
So from this we can see that in simple welds simply maintaining a Small gap between the weldments and using a solid tack at the end point of the weld we can dramatically reduce warpage.
Following the examples A to C it can be derived also that the amount of warpage is related to the difference in temperature gradient across the bar from the unwelded parts to the welded parts, for this reason if the weld area is preheated, the difference in cooling and therfore shrinkage rates will be smaller reducing warpage,
A further point is that welding processes that have a stronger energy gradient (more heat in a smaller spot) tend to increase warpage, so TIG with a very high heat in a tiny spot can cause massive warpage where flame welding which in it's own right preheats the metal as a result of it's weak energy gradient causes less warpage, MIG and Stick are about even but MIG has the advantage of blowing a cooling gas onto thinner work pieces which has an affect.
If folks find this useful or interesting I can go through any number of examples that demonstrate the mechanics in different joints, but I do accept that this is very difficult to visualise for some and so rather than waste time I will follow your lead. Please take part in the poll so I know what you want.
Best regards
Rick
Whatever it is, do it today, Tomorrow may not be an option and regret outlasts fatigue.
I can do it, with a photo. Now how do I do a slideshow?
It's actually nice to assemble it in text form just to square it all up in my own head every now and then.