bigjimslade wrote:
I have been trying to find sources that address the welding v. riveting question. Dulin has a page on welding but it does not really address the question of why. None of the other standard references address it. If anyone knows of a source(s), let me know. I have asked in other fora to no avail.
We do know that there was some level of distrust of welding. When Liberty Ships and T-2 Tankers were breaking in half, weld quality received the blame initially. It turned out the quality of welds had little to do with it and poor quality steel that became brittle with cold and poor design for welding (e.g., hatches with square corners) were to blame. The latter is similar to the problem the Comets had.
On the Iowas, the bow is welded. All but the upper four strakes are welded end to end. The structural members are welded. The hull is welded where riveting would have been impracticable. And the Superstructure is welded.
So a lack of trust in welded joints does not appear to be the concern.
I have been puzzled that they butt welded the hull strakes but in much of the ship there appears to have been a distrust of butt joints. In the barbette supports, there are butt plates welded over the butt joints; something that is weaker than a simple but joint. Then, oddly, the turret supports are riveted.
Lack of trust in welding was a concern. But welding offer advantages because it is faster, saves steel, and produces a lighter finished structure. So where production rate is of greater concern than durability, like building liberty ships faster than the Germans can sink them, the lack of trust is accepted as a necessary price to pay.
On capital assets like battleships, welding is also accepted in parts where the individual members are not highly stressed, or where failure would not be critical. Butt joints within each strake are not critical, because the butt joints do not line up across different strakes, even if one butt joint fails, the load will be transferee to areas of adjacent strakes where there are no joints, and the rip will stop.
Joints across strakes are critical, because the seams align across the top and bottom edge of all plates within a strake, If a welded seam fail at one location, it can propagate across multiple plates, leading to the entire seam unzipping.
So it is common for battleship hills started in mid to late 1930s to be a mix of welded and riveted parts. The percentage of hull that is welded appears to correlate with experience the navy had with welded ships, and the state of welding technology and availability of metal alloying elements required to make steel weldable in a country.
For example, on the Bismarck, most of the hull was welded. But the Germans had been building largely welded warship hulls since the 1920s, and had a lot of experience. But the torpedo defence system, which are expected to be subjected to the greatest load and experience the largest deformation, were still riveted.
On Yamato, parts of the hull outside the citadel was welded, but the hull around the citadel was riveted. The Japanese were also early adopters of welding, but japan had some bad experience with welding that required completed warships to be rebuilt replacing welded seams with riveted seams. So they became more conservative.