In addition to specific comments I have added below, I have two general comments:
A) For automation to bring the biggest bang for the buck, things have to be redesigned. Cars aren't constructed the same way they were 60 years ago, which in turn were constructed quite differently from how they were manufactured 60 years before then - part of that is for safety improvements, but part of it is also with an eye toward automation. When construction firms are so fragmented, it's hard to develop demand for new techniques for the same investment reasons given in the article. When competition is fragmented and profit margins razor thin, it's hard to be the first to step up and pay the design and testing costs of entirely new ways to construct things. And in addition to the engineering, you'd often also have to address local code and whether the new methods meet all the criteria codes carry. It's almost like you'd need a big public/private partnership to design new techniques where governments initiated the process to help fit new codes to new techniques and then gave away that design IP to anyone who wanted to use it.
B) Reduction of unions. Trades and construction have probably resisted the decline in unions more than most industries, but how well have they? How much has unionization fallen in construction and the trades? Specific to the trades, unions often bore a large responsibility to provide well-trained labor. I remember reading an interview with the Sears Tower engineer, Fazlur Khan, where he credited unions with making sure his buildings got built correctly - he said he didn't think the Sears Tower could have been built in, for example, Miami, where unions weren't as strong as Chicago.
Quote:
Originally Posted by aaron38
1) American industries have increased productivity by outsourcing sub assemblies, and often the whole product, to cheaper countries. Can't be done for buildings. And while low skill labor can shingle a roof, it can't do highrise construction.
Highrise components are large enough that pre-fab isn't practical. Could caisson cages be produced in a factory and trucked to building sites? As large as they are, they'd either produce huge traffic jams or not be able to make the turns. So they're built on site.
|
The article isn't only about highrises, though, it's covering construction overall - highrises, SFH, bridges, tunnels, the whole nine yards. Skyscrapers may always be more custom, but certain kinds of highrises could probably be standardized and have additional automation designed for/with them.
Quote:
Originally Posted by aaron38
2) Architecture is unique, and unique resists automation. The self-climbing forms are standardized, but each building is custom, often with floorplates changing floor by floor. Maybe if every building was an identical Miesian box then robots could build it. I've seen in Hong Kong and China where a dozen identical towers are thrown up together. Is that what people want to be more efficient?
|
Skyscrapers are unique. Many highrises are fairly unique. Smaller structures are less unique - certainly a lot could be gained by designing modular SFH and small multi-unit buildings. There is some work being done to that end, but it's still fairly small gains so far.
Barcelona Housing Systems is doing some work to that end.
Quote:
Originally Posted by aaron38
3) Call me old fashioned, but the only thing that scares me more than self-driving cars is self-driving cement trucks. Urban highrises don't have room for concrete to be produced on site, it has to be trucked in. Even without the drivers those trucks have to be washed out, maintained.
|
This is, of course, assuming that concrete will continue to be the primary building material. Part of why it's selected is that it can be cast in nearly any shape and is inherently customizable. If, as part of a push to drive efficiency, customization is made of secondary importance to efficiency, concrete may lose some of it's luster, or reserved for showcase buildings instead of also using it for work-a-day buildings. It may still be preferable, but materials science definitely needs to be part of any push to enhance construction efficiency.
In this regard, the
Forth Bridge,
Forth Road Bridge, and the upcoming Forth Replacement Crossing aka
Queensferry Crossing, near Edinburgh, Scotland, all come to mind as visual examples of how infrastructure projects' design changes over time based on available materials and advances in how to apply them. Each subsequent bridge has been constructed with fewer materials, fewer workers, fewer deaths (73 deaths with the first Forth Bridge, 7 with the Forth Road Bridge, and it should have been zero, but it appears 1 death with the Queenferry Crossing bridge), and less time (8 years, 6 years, and 6 years - would have been 5 but there were weather delays).
Quote:
Originally Posted by aaron38
4) Automation works best in clean, controlled indoor environments where all the variables are removed and the machine does the exact same task over and over and over again. With no dirt, dust, moisture to speak of. You can eat off the floors in electronics factories, everyone walks around in booties.
|
Current automation of extremely fine items works best in cleanrooms. But comparing microchips or the assembly of microelectronics to infrastructure construction is not really appropriate in my opinion. It's a bit like saying you can't
automate animal butchery because automated brain surgery requires extreme precision. The tolerances are nowhere near the same, and how you would go about automating which parts of the process(es) instructs how you address such issues and engineer the overall process.
Quote:
Originally Posted by aaron38
I've seen enough construction photos here, seen men caked in dirt climbing around inside forms, running rebar and conduit and pipes, with every section different. No way can that be automated, not with the types of machines that run assembly lines.
I could go on, but there are dozens of factors why producing highrises and producing iPhones is not comparable.
|
Exactly, they're not comparable, and you would design completely different processes. Engineers can and do completely rethink assembly when automating tasks. Just as a most basic example, how you automate mass-production baking bears very little resemblance to how a baker operates a small bakery shop and yet the end products can often be made to be nearly indistinguishable - it's that's a desired goal.
Quote:
Originally Posted by aaron38
^^^ One thought, does the decline in productivity track an increase in safety practices? Also, is the final product constant between the 1960s and now? For example, units in the 60s weren't all wired for high speed internet and cable, didn't have the luxury touches.
Are we comparing different things?
|
I think the biggest jump is safety happened between about 1890 and 1960 - you saw about a 90% reduction in deaths in that timeframe. It's possible, I suppose, that getting down that last 10% to what is, for many projects, effectively zero deaths could have caused some stagnation in efficiency growth, but I would find it hard to believe that that actually caused a decline in efficiency.
Prev
Linear Mode
