A small welding job

Norman - the reality at the time, in the early 1990's, was that these German and Japanese Thermo-Mechanically Controlled-Processed plate High-Strength Low-Alloy steels were so superior, leaving "us" having to buy these steels and put them through our pipe-mills while plate capacity here stood idle. One advantage was their weldability. That was what I sought to address. Previous weld cracking susceptibility tests we had - you could set every variable to the maximum and the TMCP steels would not crack ever.

There were other advantages. Tough at arctic temperatures. High strength. Highly resistant to "sour" crude oils for pipelines. Etc.

I "got" weldability and sour oil resistance.

I set out to understand the weldability. That explanation also explained the resistance to "SOHIC" - sour oil resistance.

Models lead the way, and what mine showed appears to explain everything. I would have needed more tests to prove whether the apparent explanation was the actual explanation - but I had had a bruising journey through my PhD, and the thought of more time in academia was so horrible I didn't even have it.

Regarding this work and how you seem to find it remarkable - in "bang for buck"? - I found others. One is "fatigue-resistant welds"

Then there's been well-chosen welding techniques.

I have helped people who know me with analyses which solved major "challenges" but they are confidential.

This was the best work I got

BTW I did a memoir of the Doctoral research I did

Norman - no-one previously knew there was a more than 3-fold difference in the diffusivity ("movement rate") of hydrogen in "plate steels" - thicker C-Mn steels used for general construction, pipelines, ships, etc. I found that out experimentally early on. That was a surprise. It had my thoughts going on a path which lead to "the sixth-jumping solution". In all fairness - those initiating and putting together the project had that intuition that movement could surely be the only aspect concealing the explanation seeing as so much is / was known about other things. The steel samples they had assembled for me was an embodiment of that intuition of theirs. But in all fairness to me - I found the test(s) which revealed the hydrogen-movement behaviour, when everyone else thought it all-but-impossible. My experiments had about the same qualities as my computational model / solution, I suggest...

My "Wedge Weld Hydrogen Penetration" test - that only works because there are things going on we do not understand. But the pattern of results is so exact - movement-distance is strictly proportional to square-root of time always as ever seen. So there was confidence to use this "unexpected bounty". But the thing is, I did a "scattergun" approach early on, with "dead-certs" not working at all and "almost no hope'ers" astonishing with being workhorses - the WWHP test being one.

You are clearly a scientist - there were about four previous scientific papers on hydrogen mmovement in welds - none with any analysis of what physical phenomena are giving the results. You will be knowing - that is double-unusual.

• tiny number of previous investigations reported (and that being more because cooperation in welding science never ceased during the Cold War)(you'd be expecting thousands minimum given the economic, commercial and military relevance).

• a scientific paper published with no mechanistic model trying to explain the results is rare

So that investigation could have been subtitled "To boldly go where no-one considered it a particularly good idea to go before".

To me it's just "diffusion": with that sort of basic process -- particles wandering around randomly -- it's hard for there to be any other governing law.

Well, the differential equation is simply the difference equation (what you used) taken to the limit of infinite resolution: not a big step in and of itself, but it opens the door to taking advantage of some powerful mathematical techniques (as well as of course the risk of getting lost in math and never finishing your project).

Ah, but that presumes you're storing a number for every point in a 3-dimensional grid and doing a calculation at each point for every timestep. There are more sophisticated ways: for instance one way to solve the heat equation is to take the FFT of your starting state (in all three dimensions, one at a time), after which you can get the solution at any subsequent time by just multiplying each FFT coefficient by an easily-calculated value (an exponential decay proportional to spatial frequency) and doing the inverse FFT transforms. (No time-stepping: just go directly to the desired time.)

Now, that only applies to a 3D rectangular block, and your shape was a bit more complicated than that, so you couldn't have done exactly that; but it's an illustration of the sorts of techniques that are out there (and that these days you could access by just using a good heat equation solver and pretending hydrogen concentration was heat).

You answered that one yourself: demand physical correctness. (And beware of math tricks; admit no bad assumptions.)

Hack is not a dirty word here; it just means not taking as much care as would be taken in a really thorough solution. Like, it may seem a waste of time to take your difference equation, take the limit to make it a differential equation, only to convert it back to a difference equation to actually solve it. And for your purposes of your thesis it probably would have been a waste of time. But the differential equation is closer to the actual physics, and in the process of converting it back to a difference equation you learn what sort of errors you'll be making and have opportunities to improve them.

Hacks that work are great; it's just that not all of them work. You mostly validated your solver, by checking against an analytical (textbook) solution, so that part worked, but I don't believe you validated the aspect of the solver that gave you sharp jumps across boundaries between different diffusion coefficients.

Has that improved? (Or has it gotten worse, with no longer even a British attempt to compete?)

Norman - no-one previously knew there was a more than 3-fold difference in the diffusivity ("movement rate") of hydrogen in "plate steels" - thicker C-Mn steels used for general construction, pipelines, ships, etc. I found that out experimentally early on. That was a surprise. It had my thoughts going on a path which lead to "the sixth-jumping solution". In all fairness - those initiating and putting together the project had that intuition that movement could surely be the only aspect concealing the explanation seeing as so much is / was known about other things. The steel samples they had assembled for me was an embodiment of that intuition of theirs. But in all fairness to me - I found the test(s) which revealed the hydrogen-movement behaviour, when everyone else thought it all-but-impossible. My experiments had about the same qualities as my computational model / solution, I suggest...

My "Wedge Weld Hydrogen Penetration" test - that only works because there are things going on we do not understand. But the pattern of results is so exact - movement-distance is strictly proportional to square-root of time always as ever seen. So there was confidence to use this "unexpected bounty". But the thing is, I did a "scattergun" approach early on, with "dead-certs" not working at all and "almost no hope'ers" astonishing with being workhorses - the WWHP test being one.

You are clearly a scientist - there were about four previous scientific papers on hydrogen mmovement in welds - none with any analysis of what physical phenomena are giving the results. You will be knowing - that is double-unusual.

• tiny number of previous investigations reported (and that being more because cooperation in welding science never ceased during the Cold War)(you'd be expecting thousands minimum given the economic, commercial and military relevance).

• a scientific paper published with no mechanistic model trying to explain the results is rare

So that investigation could have been subtitled "To boldly go where no-one considered it a particularly good idea to go before".

-----------------------

Diffusion of atoms is extremely important, thoroughly studied and fairly easy to measure over distance and time in semiconductor fabrication. I was on a team that designed and built the necessary instruments for automated production testing. The electrical properties of Silicon are very sensitive to the concentration of trace amounts of other atoms.

My PhD research should have been subtitled "To boldly go where no-one thought it a particularly good idea to go" (parody on "Star Trek" upbeat theme) and when I completed it, after the scientifically posed "Conclusions" the common-language ultra-brief summary would have been "You're screwed".

The properties of the TMCP steels seems only attainable using the TMCP production route. The TMCP steels have a very "clean" microstructure of very fine ferrite grains. Nothing else. Tiny precipitates which with a scanning electron microscope (SEM) you can resolve are just that - some very fine solid precipitates. Presumably reaction products of the "fine" additions of the likes of Titanium to an already very clean pure highly deoxidised melt. (I had not the budget and it was of no importance to my work to use electron spectographic methods to analyse what they were - what SEM I got was a "gift" anyway).

Other ways to get the same properties would cost a lot lot lot in alloying elements and would have none of the weldability and sour-crude-oil resistance. Total non-starter.

I have never seen one of these, and no photos are released, but I do know a few who have seen the rolling mill stand(s) in a Thermo-Mechanically Controlled-Processed plate steel plant and they are apparently jaw-droppingly awesome. They "work" the steel with heavy reductions at blood-red heat (not light reductions at yellow-heat). There was not possibility to make that investment here.

So the answer is "No".

Yes. But with hydrogen in steel the hydrogen is moving millimetres in minutes. What you'd do for every other element - take a slice and measure the concentration-at-position in the hours days and weeks following - very specifically cannot work for hydrogen in steel. Restating: if you try to slice a sample with intent to measure the hydrogen in it, the hydrogen would be long-gone by the time you had your slice.

There is then an additional problem. Suppose you had your slice of steel with the hydrogen it had previously still all there in the conentration profile it had... How are you going to measure that hydrogen concentration in-situ??? Element #1 - you can name a spectroscopic method which would tell you what concentration of hydrogen is there? (hypothetically - "neutron spectroscopy" - but there is no such instrument with fine beam to spot-probe and plot concentration

That is why no-one had managed to touch this topic before - despite all the steel welded in the world, no-one had information where the weld hydrogen went and in what time-scale. Yes, cracks in the weld metal and heat affected zone said "'ydrogen woz 'ere" but that's about it...

That's why I work as a welder - I frightened everyone doing what I did

- and to be honest was very broken and damaged myself after getting through that. Then furthermore, what I discovered didn't fit with text-book writings

- and no-one questions what is written even if "God" seems to be saying otherwise.

Etc.

Best wishes,

No way!!!

"Random" is the problem - it is hardly likely to be so.

For two element types to co-exist in one space with no interaction is hardly probable. Fick recognises that. He gave us a baseline: "What would happen if there were no interaction". Everyone then has to be realistic - that is a "baseline" reference case but is almost never going to happen.

My computing solution is a "caveperson with a wooden club" method and should be recognised as such.

You are obviously very bright.

Then there are other things going on we absolutely do not know about. "Asymmetric diffusion", where the rate the solute enters the solid solvent does not match the rate the solute leaves the solid solvent, was previously known and observed. Broadly the "in" rate is Fickian, but the "out" rate is slower and "something else" for hydrogen in steel.

Even what we do know - that for treatments like cold-working steel the product (multiplication) of solubility and diffusivity stays the same with increasing cold-work (cold work increases - S increases; D decreases) (the "permeability" - seen abundantly elsewhere in-support)

- which means that solubility and diffusivity must be dependent variables on the same one underlying independent physical state - got massive explosive vitriolic response when I counselled that at the time about 20 years ago to someone dealing with hydrogen. Academics who were absolutely livid at the suggestion which is to be seen right there "exact" in experimental data.

Back to movement and "something else" happening...

I found that this is happening in steel welds. That is why I needed my computational solution.

"Fickian" assumptions are used as the best guess in explaining clauses in Standards for welding and hydrogen - but I showed for certain the reality is something else.

I could visualise experiments to try to "break into that" / "get a window into what is going on" - but that would have needed a new project, and I'd gone through two Universities and three supervisors just getting this project done...

Best wishes,

Yes. But with hydrogen in steel the hydrogen is moving millimetres in minutes. What you'd do for every other element - take a slice and measure the concentration-at-position in the hours days and weeks following - very specifically cannot work for hydrogen in steel. Restating: if you try to slice a sample with intent to measure the hydrogen in it, the hydrogen would be long-gone by the time you had your slice.

There is then an additional problem. Suppose you had your slice of steel with the hydrogen it had previously still all there in the conentration profile it had... How are you going to measure that hydrogen concentration in-situ??? Element #1 - you can name a spectroscopic method which would tell you what concentration of hydrogen is there?

------------------- Raman lidar.

------------------- (hypothetically - "neutron spectroscopy" - but there is no such instrument with fine beam to spot-probe and plot concentration

That is why no-one had managed to touch this topic before - despite all the steel welded in the world, no-one had information where the weld hydrogen went and in what time-scale. Yes, cracks in the weld metal and heat affected zone said "'ydrogen woz 'ere" but that's about it...

That's why I work as a welder - I frightened everyone doing what I did

- and to be honest was very broken and damaged myself after getting through that. Then furthermore, what I discovered didn't fit with text-book writings

- and no-one questions what is written even if "God" seems to be saying otherwise.

Etc.

Best wishes,

------------------------------

Is the rate of diffusion through a thin sheet relatable to the rate of diffusion the same distance into a solid?

For real-time measurement:

'The effective diffusion coefficient of hydrogen in the presence of defect sites is always smaller than that of ideally dissolved hydrogen in the defect-free perfect crystalline lattice."

----------------------

Inversely, the rate of diffusion might be a measure of the quality of the steel.

Correct. Exactly so. "Trapping". By 300degC and upwards all steels are showing the same diffusivity to hydrogen accordign to all experimental data - they all interlock to say that.

Thanks "Raman Lidar"

Too complex to say that, really, I believe

Thanks "Raman Lidar"

--------------------

If the H2 concentration is high enough there are commercial sensors.

--------------------------- Carbon monoxide detectors also sense hydrogen, this claims they display 10% of the actual concentration.

Oh, it's definitely a random process: thermal energy causing all the atoms to vibrate and the looser ones (like hydrogen) to wander around. But there are a lot of different random processes.

Your objection would have been better addressed to the words "wandering around" -- because thinking it over again, I see you have a point. Suppose that on average the hydrogen atoms spent almost no time wandering around and almost all of their time stuck in some trap (some local energy well). Then, at equilibrium, a part of the material with twice as many traps would have almost twice as many hydrogen atoms, since at equilibrium the concentration of hydrogen atoms *that are wandering around* is equal everywhere. (That last is the part of my argument that survives.)

Now, the hydrogen atoms have to spend some of their time wandering around, or there wouldn't be diffusion at all. But it might be close enough to zero that the difference can be neglected. In any case, if a fraction X of them are trapped and the rest are wandering, then the concentration ratio is 2:1 for the trapped ones and 1:1 for the wandering ones, so an overall ratio of (1+X):1.

Of course that's just c "Ludwig Boltzmann, who spent much of his life studying statistical mechanics, died in 1906, by his own hand. Paul Ehrenfest, carrying on the work, died similarly in 1933. Now it is our turn to study statistical mechanics. Perhaps it will be wise to approach the subject cautiously."

(Opening lines of "States of Matter", by D.L. Goodstein).

You seemed to know the reason when you were writing your thesis: it's the result of a difference in energy levels between the two sides, with one side having a greater chemical affinity to the substance that is diffusing and thus readily sucking it out from the other side. (And with hydrogen in particular, the hydrogen molecule having to dissociate into hydrogen atoms before it can penetrate the steel, with that dissociation requiring a lot of energy.)

Not all academics are like that. Feynman wrote that when he was investigating the Space Shuttle disaster, he got along fine with the people who worked on the solid rocket boosters: he knew stuff from theory and they knew it from practice, but all were on the same page.

That said, your style is a sort that raises the hackles of theorists: you're making theoretical statements (trespassing on their turf) but doing so using crude arguments: it makes them feel like they're under barbarian invasion. It's a theorist's job to explain experiments and to look past any crude justifications an experimenter offers, but not everybody is good at their job.

That much I gathered. But there are a lot of things that can happen in the world: sometimes failing British companies have been bought by their German competitors and revamped.

"Ludwig Boltzmann, who spent much of his life studying statistical mechanics, died in 1906, by his own hand. Paul Ehrenfest, carrying on the work, died similarly in 1933. Now it is our turn to study statistical mechanics. Perhaps it will be wise to approach the subject cautiously."

(Opening lines of "States of Matter", by D.L. Goodstein). ... Norman Yarvin snipped-for-privacy@yarchive.net

---------------------

Read at your own risk:

Einstein accepted that Quantum theory explained observed results but he didn't believe it was the real answer. He died of heart failure.

Ahhh... The problem is that these are detecting hydrogen after it has already left the sample. The fundamental scientific issue is that how hydrogen leaves the "solid solvent" iron/steel is complex in ways we don't know.

Scientists have run round and round in circles with the Devanathan and Stachurski (?) electrochemical permeation cell because it introduces two surfaces not part of the what happens within a solid object - the surface though which the hydrogen is input and the surface through which hydrogen leaves (the electrochemical permeation cell has the sample as a disc / wafer). It has been realised that what happens at those surfaces tends to dominate the results and little of what you measure in this apparently "perfect" method is what's happening

That is why I enraged some "scientists" - at the outset I "threw-out" any theory or method which did not have a ready demonstration that it really works like that.

Against much fury, I started with "hydrogen does move in steels". Then had only hydrogen bubble formation in glycerol as my detection method. Something wise scientists will tell you - it's something like "You cannot browbeat God". That is at the root of finding your way into a subject, seeing what is there and conveying what you have seen ("creating new knowledge").

Your "Raman Lidar" would work for a thin sample extracted from eg. a weld, at liquid nitrogen temperature (hydrogen is "condensed" on "traps" and is immobile) and that sample suddenly brought back to room-temperature. You could have the laser beam scanning the sample - likely a raster-pattern, plotting {rate of hydrogen evolution (inferred by hydrogen concentration in the gas surrounding} vs {position (x-y coordinates)}. There is no guessing how well this would work (?).

Roaring with laughter!!!

Yes, I had that self-perception.

You are "running with the baton" now (analogy to a relay race), and I work as a welder mainly on marine (boats, etc) and marine civils.

That said; in welding I went for an interview at an esteemed offshore engineering Company in the oil&gas sector and the panel of four interviewers behind the desk had to hold onto that desk to avoid falling off their chairs at one response. They had picked a fascination I did not share with the sequence of small rough-as-... steel fabrications companies I had worked for and a somewhat disengaged answer had then wiped-out. My response conjured up a world where more than half the people who tendered their resignation scripted it in the "four knuckles format".

Happy to read your comments. Best wishes