Lessons from shipbuilding productivity | Hacker News Reader

Lessons from shipbuilding productivity

constructionphysics.substack.com

85 pointsworldvoyageur18 comments

IIRC, Liberty ships were of fairly poor quality due to their rushed construction and welding vs riveting, but that they only had to complete a single Atlantic cargo crossing to be worth the construction cost. Sounds a bit like fast fashion in that regard.

Also because there were so many Liberty ships left over, many became privately owned, possibly tramp freighters? Easy to see parallels to the "space trucker" sci-fi trope of a (relatively) small, armed, independently owned and operated cargo vessel.

Good 7min video on Liberty ships: https://www.youtube.com/watch?v=8qDxqBvK3NA

Being an industrial engineer, a suuply chain and logistics guy with a prdound interest in history, the Liberty ships are one of my favorite subjects. So many lessons to be learned from that!

EDIT: I also have a special place in my heart for Star War's tramp feighters and their captains, legit or otherwise.

They were planned to be bad, but after a while they became rather good. The factory approach results in incremental improvements all around, including build quality. Many continued to work long after the war.

> They were planned to be bad, but after a while they became rather good

This is the very definition of the (misunderstood) learning curve.

In this half-fraught context, I'd like to raise the question of a potential new era of "liberty ships" -- i.e. proposal(s) to build modular nuclear plants using shipbuilding technology and scales. Presumably high standards of quality control could be maintained from the start.

On the other end, one wonders how large freighters might become were the Panama and Suez canals not constraints.

At the risk of derailing discussion, this is why Seaborg and Thorcon plan to use shipyards to mass-produce molten salt reactors.

https://thorconpower.com/production/

https://www.seaborg.com/

Seaborg presentation: https://www.youtube.com/watch?v=x-Dz9sfBKEg

One thing about his hours comparisons: it is not clear that British and American yards counted hours the same. This matters for office work like accounting, materials received, etc.- was that counted or not? Which of them was counted, for what contracts? The most that D.K. Brown (in _Nelson to Vanguard_) felt could be said was that the data was suggestive that the British shipyards were more efficient, not that they actually were.

>The Liberty Ships, on the other hand, would be welded. Not only could welding be done faster, but unskilled workers could be trained to do it far quicker than they could be trained to rivet - it took just 10 days to train a welder

I'm very skeptical that it took more than 10 days to train someone to rivet. Riveting is very simple and I seem to recall I taught myself how to rivet with a cheap Harbor Freight blind rivet gun in a matter of minutes. I've shown several non-technical people rivet guns as well and they've been able to use them easily. Drill hole, put rivet in hole, squeeze. That said, ships did not use blind rivets, they used a two-person rivet setup where one person would put in the rivet and the other would hammer it on the other side, but that would still be far easier than welding to teach - it certainly would not require 10 full training days like welding apparently did!

I suspect that riveting was just slower per steel joined per person-hour (much due to the use of 2 people instead of 1), in addition to the excess material used by overlap required by riveting. Training time wouldn't even matter much - if one process took 10 days longer but was 10% faster, then it would pay off over about a half a year of work. I have no figures for how much faster either was in use or training time.

That said, I imagine the steel saved (in both rivets as well as overlap) was probably the main component. Apparently the HMS Ark Royal, a British aircraft carrier during WW2, saved 500 tons of its total 22,000 tons by using welding, so about 2.2%. A Liberty ship is about 10,000 tons, so you'd save about 220 tons by welding. The USGS's data puts a 100 lb hot rolled steel bar at $2.15 in 1944, so very very roughly that's $43/ton or $10,000 per ship saved. A Liberty ship took 592,000 man-hours to build, at $0.30/hr (1944 minimum wage) that's $178k; so the steel reduction could pay for an extra 6% more hours, even if welding was slower, which is significant if you're churning these things out.

[0] https://pubs.usgs.gov/sir/2012/5188/sir2012-5188.pdf

Riveting is not simple. There are various types of rivets and riveting procedures for different occasions. It is simple if you just need two things to stay together, but on a ship there is a lot other considerations.

But I am surprised that welding was deemed simpler. Welding seems to me even more complex than riveting.

I am going to assume that it is more the amount of time to gain enough proficiency to be productive. The amount of physical dexterity and coordination required can be quite high for tasks like this, especially when having to work quickly in non-ideal conditions.

It is one thing to be able to swing a hammer, but to do it with force and accuracy when what you are hitting it is not right in front of you and above your head, is quite another task entirely.

I always find infuriating how people under estimate the complexity of doing things professionally. Software developers seem terrible at this even though they really should know better. Everything looks easy under ideal conditions at an abstract level.

Swinging a hammer might be difficult but you have to understand that in those times most young men of working class would have a lot of experience swinging a hammer.

I have been raised on a farm when I would be doing a lot of various tasks like that. I observe how people who have not had that kind of experience are absolutely helpless when faced with tasks that require that require to use your hands with strength, dexterity and accuracy. I assume a hundred years ago people would have way more experience than I have.

So when we are talking about people having to be trained, I assume that most of them would already have extensive experience handling a hammer or an ax from early childhood.

Talking about the hammer swinging was more trying to illustrate how something simple can actually be a lot more complicated to carry out in actual practice. Not what I think was the likely limiting factor in learning this riveting process, not sure they would have been using an actual hammer, hopefully they had hydraulic hammers to do the physical part!

It was a unclear example for illustrating learning processes that require physical dexterity and manipulation of materials can actually be more complicated than initially appear.

I was more thinking I can see how you can arrange it so welding would require less of these skills before you can be 'productive'. Alignment tolerance, fitment etc. Getting good welds on the other hand that would pass modern inspection...

Without understanding the actual processes and the expectations of what was required of riveters vs. welders don't feel I can say anything. I think I was more commenting on the grandparents dismissive stance on riveting not being that hard to learn!

I have a similar experience to you also growing up on a farm, and also from watching others learn physical skills when I was working as a carpenter and a little bit of blacksmithing.

Can also be that riveters were in high demand for other tasks at the time, and welders cheaper to train and retain.

I went and watched some people installing hot rivets on youtube, and while I'm not sure it's much more demanding in terms of skill, it does look like an absolute pain to do, especially in tight spaces.

Working in the guts of a ship next to an open furnace, while people pass red hot metal around, then trying to squash that metal with a hammer (probably a large one) sounds like an absolute nightmare, and I bet it involves all sorts of fiddly clamping.

I don't think a blind rivet gun is a good comparison: the difficulty of hot riveting (if anybody has experience, please correct me) look to be around keeping the rivet at the right temperature, getting the other side properly clamped, walloping it just right under extremely adverse conditions, and making sure you don't set anything or anybody on fire in the process.

Yeah, I worded my comment poorly - I agree that the actual process of installing a hot rivet is very difficult manual labor. What I was getting at was that the concept of a rivet is fairly simple, and would not take anywhere near 10 full working days to train. One of the issues is that a rivet is pretty intuitive: the metal rivet holds together two other pieces of metal. Is it in the hole and does it have a head on both sides? It'll hold. It'd be intuitive to a child. But a weld is not as intuitive: you can make a nice looking weld that doesn't penetrate deep enough and fails easily. This can obviously be catastrophic when welding ships, so I can understand training taking 10 days.

I guess is it's probably tricky to maintain rivet temperature - as soon as it contacts the hull, all the heat is going to flow out and the rivet is going to cool very rapidly. That would presumably have a negative effect on clamping pressure, because the rivet wouldn't shrink so much after it's been hammered in place.

I think the guy upthread had it right, however: probably the main difference was the manhours involved per rivet, just because you seem to need like four people working on each rivet at once. And, honestly, if a task is just horrible to do, that's got to have an effect on productivity.

> And while the new assembly methods resulted in improved productivity, ... levels remained below what the British could achieve ... using their more traditional methods ... were able to produce ... at lower cost and fewer labor hours than the US was able to

Tell your boss this whole story when management next decides to introduce some new methodology :-)

Reply on news.ycombinator.com