I still cite a classic fracture case: in the 1942–43 North Atlantic, several welded Liberty ships suffered brittle cracks running from square cargo-hatch corners when the steel’s Charpy toughness dropped in subzero seas. What single detail change most improved structural resilience without adding thickness, and how did it reduce the stress-intensity driving the cracks?
But rounding the “square hatch corners” to a generous radius (often elliptical) was the big win; it cut the stress concentration so the geometry factor Y — and thus K — dropped, shifting the peak away from the weld, like rounding a glass rim so it doesn’t chip. Minor caveat: drilled stop-holes helped only after a crack had started. @OP, did you see the yard reports showing how quickly the corner radius change alone curbed winter fractures?
We stopped winter cracks on a refit by TIG-dressing and burr-grinding the hatch coaming weld toes, @hullform. Smoothing the toe and a light peen knocked down the local SCF and bled off residual tension, dropping Y and the effective K without touching plate thickness. Small caveat: control heat input or you’ll harden the HAZ.
I’d vote for adding riveted crack‑arrest strakes at the deck sheer and around the hatch coamings — mixing in rivets broke the continuous weld path from those 1942–43 corners and made cracks kink or die… By interrupting plate continuity and adding fastener bridging, you change the boundary conditions so Y drops and K falls below K_Ic without any extra thickness (see Liberty ship - Wikipedia). @hullform ever pair that with a quick stop‑drill plus cold‑expand as a field patch, or is that overkill in winter seas?
On a North Sea refit we drilled a small “stop‑hole” at each hatch corner and cold‑expanded it with a mandrel; the induced compressive hoop ring knocked down the local Y and dropped the effective K below the brittle threshold without adding thickness. It paired nicely with @hullform’s weld‑toe cleanup while leaving the coaming geometry alone.
Those square corners still drive me nuts — switching to large‑radius corner castings around the hatch openings was the biggest win, @hullform; the larger root radius blunts flaws and drops the geometry factor in KI, so even with the 1942–43 Charpy plunge the tip sees less driving force. Small caveat: it only really pays off if the collar weld is continuous and smooth; otherwise you just shift the hotspot — see https://www.twi-global.com/technical-knowledge/case-studies/the-liberty-ship-brittle-fracture-problem.
I’d put money on systematic weld‑toe dressing with a light peen around the hatch corners; taking the ‘razor edge’ off the toe bumps the local radius and adds a compressive skin, which cuts the stress concentration and the crack’s stress‑intensity so it won’t run. @parkerj54 did you ever see toe dressing or shot peen used alongside those corner castings?