Fatigue performance of welded steel structures

Where, when and why this work was done

This work was done in Spring and Summer 2011 as a project while on Welding Engineering course at Cranfield University.

A consortium of interests - military ship-builders wanting fast ships slicing through ocean waves; crane manufacturers; earth-moving equipment manufacturers - revisited a familiar question... Has any means become available to raise fatigue-resistance of welds, so the the performance potential of steel plates can be utilised???

Motivation - "free" welded structures to achieve higher performances

The ability to resist metal fatigue, a phenomenon caused by ever-changing load, is the limiting factor in a lot of design of metal structures. That is; of all the factors which must be designed for, fatigue often most limits what the service performance of the structure can be.

The welds are usually much more vulnerable to fatigue than the lengths of steel section between them. So achieving fatigue resistant welds is good news for steel design...

First fatigue tests

These are about "hitting the learning curve early". On a representative steel (some "S355" from stock), a couple of representative GMAW welds - in dip-transfer and spray-transfer (see "waveforms" previously) - were transverse-sliced to produce these "transverse non-load-bearing weld" fatigue-testing samples.

The purpose was at least two-fold

So here is my initial fatigue testing programme.

Practical skills and fatigue-testing

Science provides the direction: practical skills make intentions possible. Here are the fatigue-testing machine's hydraulic hoses lashed with rope to prevent them becoming a victim of fatigue.

First try for fatigue performance of hybrid Laser/MIG welds TIG-dressed

Unfortunately it was not the sample which fatigued! The fatigue testing machine fatigued the bolts holding the lower grips mounting flange to the hydraulic piston.

Starting that aborted fatigue test:
hybrid Laser/MIG welds TIG-dressed.

DPI - not promising

Dye-penetrant inspection (DPI) did not seem promising looking for growing fatigue cracks in fatigued but unbroken samples


MPI = Magnetic Particle Inspection. Detects surface and near-surface flaws. Is very promising for this application! What is more, see my simple MPI method. Produces excellent results without complicated equipment.

Yet to prove will work for the sample in the ~150mm space between the grips of the fatigue-testing machine. Would pause test. Purpose would be to try to detect growing fatigue cracks during the fatigue test.

Fatigue performance of hybrid Laser/MIG welds TIG-dressed - late July

Managed it - successfully fatigue-tested hybrid laser/MIG TIG-dressed welds . With interesting outcome!

Fatigued fracture surface of hybLM as-welded

Multiple nucleation-sites for cracks on this hybrid laser/MIG weld fatigue-tested as-welded.

Metallography of a fatigue crack in an unbroken sample

Here is one of the early samples fatigue-tested but unbroken which has been sectioned and inspected by optical microscopy to reveal an early-stage fatigue crack.


Other miscellaneous stuff.

(R. Smith, mid2011, 21Jun2017)