Salesforce Transit Center: San Francisco's $2.2 Billion Cracks (popularmechanics.com) 81
Slashdot readers jimminy_cricket and Thelasko share a report from Popular Mechanics about how San Francisco's Salesforce Transit Center went from the Grand Central of the West to a $2.2 billion construction debacle. Here's an excerpt from the report: Built at a cost of $2.2 billion, the Salesforce Transit Center and Park formed the cornerstone of the Bay Area's ambitious regional transportation plan: a vast, clean, efficient web of trains, buses, and streetcars, running through a hub acclaimed as the Grand Central Station of the West. Naming this structure -- the embodiment of a transformative idea -- could yield marketing gold for Salesforce. It also could make [Marc Benioff, founder and co-CEO of Salesforce] a household name on the level of Bezos, Gates, or Zuckerberg. Benioff took the gamble in 2017, pledging $110 million over 25 years, with $9.1 million up front and the rest committed to supporting operations when the trains started running. For now, the train box sat vacant on the bottom level, awaiting a 1.3-mile tunnel connection. [...] As he took the stage on his birthday at the Moscone Center, Marc Benioff must have been confident his gamble on naming rights had paid off. He couldn't imagine that at that moment, less than a mile away, the ambassadors trained to welcome the public to the STC were now frantically waving commuters away. Rather than Grand Central Station or the High Line, the Salesforce Transit Center and Park suddenly resembled the Titanic.
Earlier that day, workers installing panels in the STC's ceiling beneath the rooftop park uncovered a jagged crack in a steel beam supporting the park and bus deck. "Out of an abundance of caution," officials said, they closed the transit center, rerouting buses to a temporary terminal. Inspectors were summoned. They found a similar fracture in a second beam. Structural steel is exceptionally strong, but given certain conditions -- low temperatures, defects incurred during fabrication, heavy-load stress -- it remains vulnerable to cracking. Two types of cracks occur in steel: ductile fractures, which occur after the steel has yielded and deformed, and brittle fractures, which generally happen before the steel yields. Ductile fractures develop over time, as the steel stretches during use, explains Michael Engelhardt, Ph.D., a professor of civil engineering at the University of Texas at Austin and chair of the peer-review committee overseeing the STC's response to the cracked-beam crisis. The cracks discovered beneath the rooftop park were classic brittle fractures. The tapered 4-inch-thick steel beams -- 2.5 feet wide and 60 feet long, with a horizontal flange on the bottom -- undergirded the 5.4-acre park on the building's fourth level, and buttressed the roof of the bus deck on the second level. By themselves, the cracks formed a point of weakness with potentially hazardous consequences. But they also suggested the possibility of a larger crisis. If two brittle fractures had appeared in the building's 23,000 tons of structural steel, couldn't there be others?
Earlier that day, workers installing panels in the STC's ceiling beneath the rooftop park uncovered a jagged crack in a steel beam supporting the park and bus deck. "Out of an abundance of caution," officials said, they closed the transit center, rerouting buses to a temporary terminal. Inspectors were summoned. They found a similar fracture in a second beam. Structural steel is exceptionally strong, but given certain conditions -- low temperatures, defects incurred during fabrication, heavy-load stress -- it remains vulnerable to cracking. Two types of cracks occur in steel: ductile fractures, which occur after the steel has yielded and deformed, and brittle fractures, which generally happen before the steel yields. Ductile fractures develop over time, as the steel stretches during use, explains Michael Engelhardt, Ph.D., a professor of civil engineering at the University of Texas at Austin and chair of the peer-review committee overseeing the STC's response to the cracked-beam crisis. The cracks discovered beneath the rooftop park were classic brittle fractures. The tapered 4-inch-thick steel beams -- 2.5 feet wide and 60 feet long, with a horizontal flange on the bottom -- undergirded the 5.4-acre park on the building's fourth level, and buttressed the roof of the bus deck on the second level. By themselves, the cracks formed a point of weakness with potentially hazardous consequences. But they also suggested the possibility of a larger crisis. If two brittle fractures had appeared in the building's 23,000 tons of structural steel, couldn't there be others?
Bad steel (Score:4, Funny)
Re:Bad steel (Score:5, Interesting)
I believe from a longer article I read the steel was sound but when they cut holes through for bolts, rivets and pipework the cutters and welders never sealed the cutouts properly and that lead to crystalisaiton around the cuts which caused micro-fractures that developed into larger fractures. I'm not a materials specialist but that seemed to be the gist of the problem.
Once again, you predict the past - incorrectly (Score:3)
GP told you what happened, two years ago.
You then proceeded to tell them (and by extension, the engineers who tested the darn thing) that they're wrong because you have a feeling.
But please do stick to making bad guesses about past engineering mistakes and stay out of public policy, where your faith in whatever comes out of your ass could actually effect other people.
Re: (Score:3)
Essentially these are known and well understood technologies that perform reliably when used correctly. Somebody in planning, manufacturing or installation screwed up massively. With the severity of the problem, it is likely that there is more than one thing that went wrong.
Re: (Score:2)
Somebody in planning, manufacturing or installation screwed up massively.
Most likely, a supplier cheated, making a few extra bucks for himself and dooming the vastly expensive structure.
Re:Bad steel (Score:4, Interesting)
Reading the article, it seems to be a combination of lack of proper treatment of holes cut into the beam and steel that just barely meets the requirements. There was also a question of the time the holes were cut, but apparently it was unknown to be critical at the time of construction. But it seems if the steel had been superior (as it was supposed to be, being made in the US) or the holes would have been done right, the issue would not have happened.
Re: (Score:3)
If it met the requirements it met the requirements. If the steel was known to perform better by a huge margin then the design would have been changed to use less of it.
My reading of the article was that the holes were speced to be polished to bright metal, whoever specified that clearly knew. Besides, I doubt that the metalurgist who analysed the failure was applying principles that have only just been learned about, chances are most steel failure modes he explained have been known about for some time. Mayb
Re: (Score:2)
If the steel was known to perform better by a huge margin then the design would have been changed to use less of it.
Steel has several properties including strength, stiffness, and ductility. Steel strengths are highly variable, perhaps 25,000 - 150,000 psi for low alloy stuff with relatively small $/ton. However designs are often controlled by stiffness, an inherent property of the material, only slightly changed by large amounts of alloy so improved strength doesn't buy much then.
Ductility is imp
Re: (Score:2)
If it met the requirements it met the requirements. If the steel was known to perform better by a huge margin then the design would have been changed to use less of it.
The requirements are somewhat fluid. And nobody (except you) is talking about a "huge" margin.
Re: (Score:2)
If it met the requirements it met the requirements. If the steel was known to perform better by a huge margin then the design would have been changed to use less of it.
The requirements are somewhat fluid. And nobody (except you) is talking about a "huge" margin.
Margin in engineering is a fun topic. I deal with it in silicon. Margin has three ends to chew away on. (1) The variation in the stuff you build (2) The bad stuff that can happen along the way that you may or may not know about and (3) The uncertainty in the failure limit. The remainder has two parts - the remaining margin that is used to calculate worst case yield and the extra remaining margin which is the delta between what you claim for yield and your knowledge of the true extra yield you have, so you c
Re: (Score:2)
The important thing here is that the errors were discovered before something catastrophic happened. It's possible to repair this, and it will be repaired, the rest of the structure will be inspected and any other problems will be fixed.
Even with the mistake things worked correctly, the damage was spotted, the structure closed, inspections made and eventually repairs will be undertaken.
The structure didn't collapse, and no one was hurt. The inspectors and engineers deserve praise for the way this was handled
Re: (Score:1)
I disagree. Sure, the secondary safety mechanisms worked after this was was discovered by accident, but that is no reason to cheer. These secondary safety mechanisms are not supposed to be needed. If they are needed, something got messed up really badly and that should decidedly not happen.
Re: (Score:2)
You really don't know what your talking about. This wasn't discovered 'by accident'. It was discovered during a routine re inspection. All structures are routinely inspected for failures like this. For example the roadway bridges you drive your car across are reinspected at a minimum yearly and potentially much more frequently if the bridge has had issues in the past.
It's absolutely routine to have issues with materials/methods in large scale construction like this. It's so common in fact that at the end of
Re: (Score:2)
Reading the article,
OK, who bought gweihir's account? Because you must be new here.
But it seems if the steel had been superior (as it was supposed to be, being made in the US)/quote
So you're saying "most likely, a supplier cheated"? I dunno ...
Re: (Score:2)
You seem to be unaware that I am an engineer and a scientist. These things intrigue me.
quick fixes (Score:1)
Replacing the two outer brittle members that frac
Re: (Score:2)
Replacing the two outer brittle members that fractured clean though sounds like whistling by the graveyard to me, past the tip of the QC iceberg
I'm sure they're doing the best they can with the power out.
Re: (Score:2)
Indeed. They are either going to find that the steel is actually sub-standard or that the whole thing is badly engineered, or both. Doing this right, you will have ample redundancy in the design. But the breaks clearly show (whatever the PR people say) that the whole thing has exceeded these redundancies. Not good at all.
Re: (Score:2)
Somebody in planning, manufacturing or installation screwed up massively.
Installation. The holes were cut in the beams but not cleaned properly before being welded.
Re: (Score:2)
That is not enough for this effect. Proper cleaning of the holes is just one thing and its absence alone should decidedly not be enough to cause this massive an issue.
Re: (Score:2)
Vecchio explained to the board that the cracks were due to a “perfect storm” of the three factors that Engelhardt says characterize brittle fractures: weakness in the metal, damage during fabrication, and the stress of load during use.
I'm baffled by the inclusion of the 'stress of load during use' statement. There is of COURSE going to be stress during use. Even the 'weakness of metal' is a little misleading, as all test done before and after show the steel passing the required tests.
“The toughness level at the surface of the sample was good but as you went to mid-thickness, the toughness dropped down quite a bit. Toughness in the centerline was very low, so the defects were sitting in material that had very low toughness. The plate itself did meet the requirements for this type of construction.”
Re: (Score:2)
Actually, the main culprit sits somewhere else. The holes, as done on these two beams, were the _trigger_, but the reaction they triggered has been set up by other mistakes.
Re: (Score:2)
Re: (Score:2)
If I read TFA correctly, it was a combination of defects in the steel compounded by not grinding out the holes that were torch cut for access.
Re: (Score:3)
The really interesting thing to me (I also read the longer article) was that in another section, they had done the hole cutting in a slightly different order and there they had no problems at all! It was only when the crystallization happened and then that affected the next step in the process, that cracks developed.
The steel was fine, proper procedure when making the holes was to grind down the outside to bright metal which was not done.
Re:Bad steel (Score:4, Insightful)
As a shipbuilder, I can tell you that, all things being equal, some steel is more "suspicious" than other steel, notably imported steel from China.
Re: (Score:2)
Re:Bad steel (Score:4, Interesting)
Setting up to conduct hardness tests on steel beams that are 60 feet long by 4 inches thick with customized flaring at the ends is a non-trivial problem. I doubt that it could be done at the point of use. Designers, engineers, and construction crews must rely on written specs and warranties that cannot be independently verified.
This looks like the counterfeit bolt problem, writ large. But much worse really: with bolts and other small parts, one can take a sample of each lot and test to destruction. Not really possible with cast or fabricated beams of this size.
The solution is to avoid the problem at the design stage. Design around construction techniques that use industry standard pieces that can be effectively quality tested on site. Use trusses of trustworthy, standard components rather than massive,specially designed, beams, etc.
The problem should have been seen and avoided by the architects, or by the engineers who reviewed the architects' work before signing off on the design. But it is only recently ----less than 50 years--- that CAD has reached a point where this kind of beam could be designed, so the industry lacks the wisdom gained from bad experiences. Being able to design a custom piece of steel within the known tolerances of the material is now possible, but it can only be done safely if the end product can be tested to assure that it is in fact within those tolerances.
Re:Bad steel (Score:4, Informative)
I used to work in QC at a green sand cast iron foundry, and face similar procedures that a steel mill has. We'd take samples before and after an important casting and include the mass spectrometry reports in a shipment (much of our work were small batches according to a customer specification). For less important things we'd take a sample twice a shift (every 3-4 hours) or at the request of the DISA operator if he felt the alloy was a bit off. Green sand was tested every 2 hours for moisture and clay, and logged. Additional tests were made on request by those on the line as well. Tensile strength tests are performed on samples at customer request, with multiple samples throughout the casting run. Those were less rare for gray iron because it's not really preferred for tensile strength. It would take me quite a while to machine and measure all the sample rods and pull them to failure on a hydraulic machine designed specifically for the task.
(not sure how I remember so much about a job I did decades ago)
Re: Bad steel (Score:2)
> Design around construction techniques that use industry standard piece
That's nice, if you're building something new on an empty suburban greenfield site. When you're attempting to shoehorn a huge underground transit center into an old built-out urban neighborhood with existing structures, custom parts are often the only thing that even makes it possible in the first place.
Re: (Score:2)
Recently there was a huge incident in the aerospace industry. A Chinese steel manufacture fucked up somehow and the quality and content of the steel was mislabeled. the company I was with had to go back and look at ALL the documentation for their products to ensure they didn't have anything built of the fucked up steel.
Re: (Score:3)
China's definition of steel is whatever scrap was melted down that day. They keep all the good stuff for their military and sell the garbage to everyone else, including the locals.
Re:Bad steel (Score:5, Funny)
There is no such thing as Chinese steel.
Its called Chinesium: Unknown Metal of unknown quality, used for all sorts of applications where cost is the only object of concern.
Re: (Score:2, Informative)
You can spin it anyway you want trying to place the blame on China, like Americans do now for every problems they have, but it doesn't change the fact that the steel used in this particular incident is Made in USA [ktvu.com], like this other Made in USA product [cnbc.com] you are so proud of.
Steel Manufacturers: Herrick
- The Stockton-based company fabricated the steel that is currently cracked. The transit center’s senior construction manager said Herrick manufactured 30 percent of the project’s steel. A total of seven steel companies (all US-based) [ktvu.com] helped in providing steel for the project.
grinding (Score:2, Funny)
its all bad steel now...
ever since the atomic blasts all new steel is tainted
basically someone didnt grind the holes that served as access and they got microfractures... combined with steel that was poorly formed i.e. hard on the outside and soft inner...
but really the story should be that america managed to build a new bus and train station... I mean how often does that happen ?
Re: (Score:1)
Conventional wisdom warned against Benioff buying naming rights to the transit center.
Don't think "Salesforce Transit Center", think "Salesforce Mugger, Prostitute and Druggie Hangout" to see why this is such a bad idea. There are some things that you really, really don't want to have naming rights for.
Re: (Score:3)
I went round this earlier this year, before full operations had resumed. It was very clean, with no sign of druggies, homeless, etc..
The gardens on the top deck are very pleasant.
Re: (Score:1)
Yeah, that's the Libertardians' new chant, they say it about San Francisco, Seattle, and Portland. It originated in an "article" in the Daily Caller (yeah, some people actually read that still) complaining about how "socialists" have taken over the city governments of the "Left Coast" and are destroying them. Never mind that the governments of all three cities are far more business-friendly than they are liberal (Seattle has a grand total of one Socialist on the City Council). Supposedly you can't walk d
Re:Bad steel (Score:4, Interesting)
It was the weld access holes for the hangars of the floor below were cut in a way that it led to added thermal stresses. The joint detail was complicated and difficult to build. The solution on one of the beams worked better than the other.
That or the “4 inch thick steel beams” (TFS) spanning 60 feet... (4” thick flange, IIRC the beam was about 5’ deep.)
Re: (Score:2)
Conclusion is poor steel quality control then?
Uh, no? Where in the article did you draw THAT conclusion? They specifically stated that some of the holes/welds made during construction were not cleaned properly. In fact, they checked other identical beams and found no issues because they WERE done properly.
China (Score:1)
Re: (Score:3)
Re:Free idea (Score:4, Interesting)
All aluminum has an oxide layer, that's the nature of the element. All raw aluminum in contact with oxygen immediately forms an oxide layer. Grind it off and a new layer is formed as soon as you're done. Anodizing is just a chemically enhanced version of this process.
You'd be hard pressed to find a glue that is strong as a quality steel weld.
As for embrittlement, it depends of the type of steel and how it was processed. Heat treating and how it's performed (annealing, quenching) has a profound impact on steel's performance. A bad weld or cut on otherwise good quality steel can compromise it's performance.
Re: (Score:3)
You'd be hard pressed to find a glue that is strong as a quality steel weld.
Or as inexpensive. Exotic glues have a pretty significant material cost. Tools, labor, and consumables for welds are not zero, but you get a lot of bang for your buck.
A bad weld or cut on otherwise good quality steel can compromise it's performance.
What I don't understand is why they didn't build the access holes into the original parts. Then weld plates over them. It's less work at the site and you're far less likely to screw something up when it's included in the design. Cutting is way hotter than welding, and the steel's failure point should be no big surprise. Bad process is a better
Re: (Score:2)
IMNSHO, it comes from having designers and architects who have never actually spent any time on a work site. Wandering around during a one hour inspection will not give you any idea what problems the guys who have to build your stuff actually encounter, you have to spend time out there and get dirty.
Re: (Score:2)
Aluminum suffers from fatigue stresses that steel does not. It's why aluminum isn't used as a structural material in buildings.
Aluminum is used in planes and they have to look for fatigue cracking every certain number of flights. Over the life of an airliner significant parts of the aluminum will be replaced due to fatigue cracking.
If you used aluminum as structural elements in a building you'd have to tear the beams and columns out every few years due to fatigue cracking. Your typical building moves a LOT,
Re: (Score:1)
Hmmm... very interesting, actually.
So, like their product? (Score:1, Offtopic)
So, the same quality as their product?
Their product is pure network effect; there is nothing about that software that should mean that almost every freaking company on earth uses it.
They are the 1970s IBM or 1990s Microsoft of CRMs - nobody ever got fired for choosing them, simply because everyone else chooses them.
Re: (Score:3)
I've never used SalesForce, but the availability of people trained and certified has got to be part of it. My sister studied animation, and got certified with SalesForce just spending a couple of hours a day for a few months on their website, and now has one of the best-paying jobs of her life. She was prompted to look into it because a friend of hers, who studie
Re: (Score:3)
I can attest to this network effect. Like switching from English units to metric units in the USA, the cost of the switch and of potential failures is something companies do not want to take on.
Re: (Score:2)
>Like switching from English units to metric units in the USA,
Them be imperial units. England went metric for engineering purposes a long time ago.
Expect this thread to run. It usually does.
re: Salesforce and support (Score:3)
Yep, precisely! I do work for a company that uses Salesforce. We started out only using a very limited set of its features, and only purchased licenses for a select few people in the company to use it. But over time, they kept finding more reasons to expand on that, so we hire a full-time I.T. staffer now who works from home and does nothing but our Salesforce support.
Because it's been one of only a couple things we use that I.T. doesn't get too involved with, I'm not all that familiar with it myself. Bu
Old news (Score:5, Informative)
Re:Old news (Score:4, Interesting)
And alternatively, US steel may just not be any good anymore. The intense efforts to find somebody else to blame is telling in itself.
Re: (Score:1)
I get that you hate the US, but China is the wrong place to defend.
Re: (Score:2)
Not if they're innocent in this particular case. That would be like having a serial killer in a city and attributing ALL murders to him regardless of evidence.
Re: (Score:1)
The person bringing up "hate" here is you. Seems you have a rather simplistic world-view. Maybe read the referenced article where quality problems with the steel are discussed? As in a far too soft core?
Re: (Score:2)
Oh, and where exactly did I defend "China"? Do you have some sort of reading dysfunctionality?
Re: (Score:2)
You replied to "China sucks" with "nuh-uh, US sucks", as you're wont to do. Seemingly without realizing your habits. Reflexive love of totalitarianism, I'd guess.
Re: (Score:2)
I did not. But somebody with really low reading comprehension and pretty low intelligence may get that impression. In actual reality I did not comment about China at all and your "analysis" only shows that your mastery of the written word is extremely low.
Re: (Score:2)
because there's never been a streetcar in the Bay Area.
Re: (Score:2)
Fransisco or Francisco? (Score:2)
Spellchecking must be pretty expensive nowadays...
Re: (Score:3)
The interesting thing is that no native English speaker will notice because they read "San Fr" and skip to the next word.
Re: (Score:2)
Re: (Score:3)
You'd think with this being Slashdot and everyone having heard of Cisco, nobody would forget how to spell it. Why, yes, that is the Golden Gate Bridge in their logo and they were founded in San Francisco.
Boston's Big Dig (Score:2)
The cracks didn't cost $2.2 billion (Score:5, Informative)
Second: the cracks didn't cost $2.2 billion, the building cost $2.2 billion. After an extensive review, it turned out that the cracks were limited to the two places mentioned in the summary, and have been fixed. The article is just the story of what happened, and it's interesting enough for a civil engineering novice like myself, but it isn't exactly a debacle. A mistake was made, the building was closed while a safety inspection was performed, and then the problem was fixed and the building reopened.
Third: Salesforce contributed just $110 million of the $2.2 billion total, and yet get to name the building after themselves. Not permanently, the naming rights expire after twenty-five years, and I find I'm a little conflicted about this. On the one hand, it's outrageous that Salesforce gets to name a public building after themselves which was almost entirely taxpayer funded. On the other hand... names are free. Having corporate names plastered all over our infrastructure is an affront to the notion of public goods, but I'm finding it difficult to be too critical of public officials who accept tangible money like that in exchange for something so ephemeral.
Re: (Score:2)
On the one hand, it's outrageous that Salesforce gets to name a public building after themselves which was almost entirely taxpayer funded
A little late to that game. Have you never seen a sports stadium?
Re: (Score:2)
What if it comes out that some Salesforce exec secretly supported Trump, and Californians start calling it the Nazi Transit Center, demanding for it to be renamed? Will gov't officials cave and give back the money to Salesforce? In Maryland, living state legislators commonly get government buildings named after themselves. What will happen should one of them be found out to be a pedo? Etc. Law of unintended consequences and such -- but it's just a bad idea.
On the flip side, in 2008 San Francisco actual [wikipedia.org]
Re: (Score:1)
This.
In addition, Salesforce has paid only $10M so far which is a pretty small fraction of $2.2B. The naming rights deal is similar to sports arenas and less than what Benioff is contributing for naming rights to UCSF Children's Hospital ($100M over 10 years). Salesforce had nothing to do with the design or construction of the Transbay Center; I don't see why they should have an opinion about a construction error.
It's sad that public works require handouts from the wealthy to be built and operated. It's nic