Overview : 3 months progression by testing


It has previously been outlined that three November 2020 projects

then in later January 2021 the

This article discusses the path ventured, responding the the sequential findings of each of the four tests.

Abbreviations used

Progression through the four mini-projects

The sentiments and consequences for the mind associated with the first two projects on butt-welds and with the latter two projects on fillet-welds could hardly have been more different.

Technically, the underlying reason is that

First two - the butt-welded RHS beam projects

The latter two projects, fillet welds tensile tested in beam test and tensile-rig for beam-test fillet-welds "ignited" a "raging" desire to go forth and solve many challenges now made accessible.

That was not the case with the first two projects

which satisfied in presented a completion with objectives fulfilled. While what they achieved was great, the distinction was that you only have to see those once, with the insight having enduring generality.

The welds were "invisible" in those tests - which come to think of it is exactly what one would want. Observing that reality was a confidence-building "fulcrum" about which to visualise welding outcomes, and validated regarding welding skills and technique.

The reward taken from this part of the test programme was at least two-fold:

The theory brought-to-bear was Euler-Bernoulli beam equations and Finite Element Analysis computational modelling. Both seen to exactly predict what happened as the hydraulics were pumped and the test samples took the loadings.
Note that in all four tests forming this "high-level overview", the hydraulics were hand-pumped, where there was the visceral experience of feeling the responses offered into your hand as you pump the handle, driving up the test loading. Thoroughly to be recommended where possible in tests.

On "enduring generality" - Euler-Bernoulli beam theory was derived around 1750. Then had its application blossom in the second industrial revolution from the 1870's with names like Eiffel and Ferris. So seeing it working 270 years after its derivation and 150 years after its mass application "you can take a lot on-trust".

So those butt-weld samples - OK that central confirmation is good - but what were those "profusion of nuances"?

Delights like

Well, that was the closed-ended contented feeling of the first two tests on butt-welded structural hollow section samples loaded in beam.

I promised this article would discuss the progression through the tests. Which did not happen yet, as the discussion followed other very rewarding topics. Benefiting from all that has been reviewed so far, here is the sequential progression through the first and second tests.
The U-weldment in R.H.S. - fabricate, analyse, test first project clearly did not severely test the welds. With it being seen that the greatest stresses and plastic deformations (bending) were elsewhere in lengths of RHS nowhere near welds.
The response to this was to design the second project Weld test - weld at centre of simply-supported beam . By putting the test weld joint at the centre of a straight beam - apparently very simple - the objective was achieved that the weld was the most severely tested feature. The good stable "distributed resistance" behaviour of the sample under test made very severe loading possible, with deformation manifested by "witness" indications extending over 2/3rds of test sample-beam's depth.
It seemed that no test with the general configuration of an RHS with butt welds could exceed the severity achieved in the second test.

Note that in the second test, the centre-welded RHS beam, "the test of the weld could not have been more severe" and "the weld was invisible to the test outcome" - which is about as good as it gets.

Latter two - the fillet-weld beam-configuration projects

As I best explained earlier succinctly, fillet welds tensile tested in beam test and tensile-rig for beam-test fillet-welds "ignited" a "raging" desire to go forth and solve many challenges now made accessible.

The corollary of which is an untranquil frame-of-mind.

That surprise "exploded" on me. Fairly much because the test(s) worked and seemingly give accurate results - both unexpected.
On fillet welds - which present a lot of accessible variables to optimise.
Furthermore - tensile tests on fillet welds have not been associated with the descriptions "easy" and "cheap". Which the beam test is.
With no upper limit or restrictions on size, material strength, etc apparent on visualising the route ahead.
Then the ramifications - that I had opened-up a way to work on some big issues stalled on means to test then seemingly being unavailable.

These two tests

were two months apart - and those two months were filled with a lot of tension trying to get the next tests to happen.

On the promise of describing the progression:
the strongest driving force by far to get from the third test ("fillet weld 1") to the fourth test ("fillet weld 2") was that one desperate need, to confirm (or discard) that amazing initial weld break strength result.

Continuing on the topic of progression:
there is a fundamental reason for the paucity of interactivity between the investigator and the "Laws of the Universe" in the progression of the test.
The beam-configuration fillet-weld test does not give any indications, qualitative or quantitative, of any form, complex or otherwise, from the sample or through the test rig equipment.
The butt-weld tests were outputting a richness of qualitative and quantitative information along the way to a termination with the sample still intact.
The only thing the beam-configuration fillet-weld tensile test "outputs" is a sense of increasing force needed on the pump handle to advance the test. The test terminating without any warning indications (as have yet been noted) on abrupt complete failure by local fracture, in the weld, of the sample.
It follows that seemingly nothing would be lost by having a mechanical pump drive the test loading, so long as it displayed the hydraulic fluid pressure. Given no useful "feedback" comes through the manual pump handle. That said, the manual pump with cylinder is totally practical, while benefiting from being simple, robust and convenient, so is likely to remain the choice.

Returning to the story which can be told...

So the test(s) had opened-up a huge number of questions, and had not yet provided trustable answers.

The reasons for those two months of tension were

This time was excruciating, but progress, as in any new domain of venture, gains pace, with a timely confluence of events enabling the second beam-configuration fillet-welds test

and brought together the restored cylinder and rig equipment to trial the new test rig , getting more data in that trial.

It is amusing to reflect that all this "chaos" came from being presented with a fortuitous 1/2hour to spare. In which I designed the beam-configuration fillet-welds test, "knocked together" the weld sample, put it in the press and got the press-force break result. All in that 1/2hour.
It all seemed good - but when I did the derivation of weld force from piston force then got the weld stress, "it really kicked-off".

That trial on the test rig confirmed that the beam-configuration fillet-weld test is repeatable.

And precipitated more analysis of the test method.
Which finally brought some satisfaction.
On identifying the established knowledge of fillet weld performance and expressions for fillet weld strength. Then being able to evaluate how my findings compare alongside the established knowledge.

This is covered in-detail in Fillet-weld test evaluation , section "Evaluation and interpretation of tests" and the more-detail following section.

The point being made here that, finally, my findings for fillet welds do engage in an interesting way with established theory.

(R. Smith, 27Jan2021 to 28Jan2021, 29Jan2021 (edits, add progression), 31Jan2021 (break tens, edits))