Lessons from small failures – Part II: Designing for durability

If you get to design and build generation after generation of essentially the same structure and you get to study the old problems, then you end up with a mature well-designed product. Over the decades automobiles have clearly become both more reliable and longer-lasting. Not necessarily so with structures, especially when they are not built repetitively, as might be a standardized overpass/bridge over an interstate; but put it in an urban area, jazz it up with architectural features and you may be adding problems. DOT’s have test tracks to learn more about particular pavements. No such luxury applies to anything with even a little bit of novelty. Lest you think the immense challenges of designing for durability apply only to the minor stuff Cary Concrete Products is normally involved with, here are snippets from a recent story.

———————————————————-

After Only Seven Years, New I-35W Bridge in Minneapolis Showing Signs of Age

07/27/2014

Star Tribune (Minneapolis, MN)

July 27 –Cracks emerged in concrete girders. A drainage hole on the bridge deck plugged up. Rust showed above piers. Seven years after the collapse of its predecessor, the new Interstate 35W bridge has been showing its age…Repairs to the system were eventually made under warranty, but not before years of finger-pointing between the state and the lead contractor. Taxpayers also covered some costs.

“Contractors are always very reluctant to give anything,” said Tom Styrbicki, bridge construction and maintenance engineer for the Minnesota Department of Transportation (MnDOT). “They always try to drag MnDOT into it: ‘We did it per your specifications. If it leaks, it leaks.’ ”

[My comments: 1) of course there is finger-pointing. Often problems arise from issues in both design and execution; 2) most contractors eventually fatigue from the battles over specifications that are poorly written (my own experience and informal polls peg error-free specs at 1%) and occasionally enforced to our surprise and detriment. The result of that fatigue is unfortunate but understandable to any subcontractor: ‘We did it per your specifications. If it leaks, it leaks.”]

Full story is at http://www.startribune.com/politics/statelocal/268746561.html or http://enr.construction.com/yb/enr/article.aspx?story_id=id:ZDAtwZvbDEgPAf81w4aA1jHIfPjJuLGNDcMwvg2Pfds8xYrzQnvP5LykSce0mHPo

——————————————————————————————————

Our microcosmic story concerns a minor failure and then thinking in detail what it takes to avoid creating designs with inherent problems. The previous post in this series went into detail about the specific problem we recently analyzed and on how surprised we were by the cause of the failure: https://planetcommercialconstruction.wordpress.com/2014/07/21/lessons-from-small-failures-part-i-searching-for-truth/

To summarize: enough moisture entered an unusually large mortar-filled void. When  frozen it was able to generate enough force to crack concrete.  We say that the piece “failed” but in fact the piece performed exactly as it was designed to perform.  The curbs at the posts in this railing system were not durable because the design did not anticipate what occurred. With hindsight we can easily see what happened.

• The rail assembly sits on the sidewalk (and the sidewalk on the deck)
• Water on the back of the curb flows to a small overhang that is part of the sidewalk
• The sidewalk, including the overhang, slopes to the roadway for drainage.
• Water flows into the mortar that is the bed joint of the curb pieces.
• Moisture in the bed joint moves into the mortar-filled cavity (blockout)
• The mortar gets saturated
• The mortar  gets frozen and expands
• Frozen mortar in the bed joint is not a problem, too small to generate large forces
• Frozen mortar in the block out starts to crack the precast curb
• The crack takes in water, ice forms and enlarges the crack
• Visible large cracks form and the curb piece falls apart.
• In 6 years, about 15% of the pieces at this condition have already  failed.
• All other curb pieces, piers, etc., are in excellent condition.

This is the railing system. It’s on a roadway bridge.

Structural detail from contract drawings.

Architectural detail from contract drawing defines a blockout sized for real-world tolerances.

Early stage of failure.

Visible cracking has occurred in this piece.

Inspecting after removal of cracked pieces and cleaning the blockout. Shows crumbled mortar in foreground.

The foresight to anticipate this is highly improbably. For 25+ years I have been regularly involved with mortar-filled openings, and never before saw a failure of this type. Couldyou reasonably expect someone to anticipate this? Failure required both a large volume of mortar and a large volume of water (and a freezing climate). Rules simply about the size of mortar-filled blockouts would not help. (I am sure such rules exist in <0.05% of architectural firms and <0.5% of engineering firms.)  Avoiding this failure would require you to have imagined this scenario. It only is an issue at the pieces at the posts. If you were using this exact railing on many bridges you would be in a different business and in 5 years or less you would have learned that this particular detail does not work. And you would never use it again.

This can also be a reminder that a BIM with intelligent objects needs both software that can store and apply the knowledge and  an organization that can capture and organize little tidbits like what we learned from this minor failure.  We are not close to doing that in non-repetitive structures.

Lessons from small failures – Part I: Searching for truth

The “failure” here is simply a piece of architectural precast concrete that cracked open much earlier in its life than expected. The piece is part of a system of curbs and railings on a bridge.  Photos were sent to me and I was able to take a good guess as to what was going on. It seemed obvious;  I’d have given odds of at least 8:1 that my guess was right (spoiler: I would have lost the bet).  Here’s what the piece looked like on the plans, in real life, and as a failed piece.

 

The bridge is  7 years old; failures started showing up a few years ago. The problem is only at the intermediate piers; there are no significant problems anywhere else on the railing.

I knew that the steel post had an anchor plate and was bolted to hot-dipped galvanized anchor bolts cast into the bridge deck and that the tube steel post had some sort of serious coating.  The location and pattern of the cracking indicating a lot of force coming from the cavity where the curb piece surrounded the steel post.  My guess was that something was rusting – maybe some non-galvanized bolts were used or the paint on the post had failed or been abraded and was now rusting. In my experience I could imagine nothing else inside the piece that could generate enough force to crack a healthy piece of precast. The precast had epoxy coated bar and the crack did not follow the bar locations, ruling out rusting reinforcement. The Quick-Look report (www.quick-look.us is our service for fast, first-look inspections based on photos sent to us) identified some rusting object, probably the plate,  as the likely cause of failure.

After getting the report based on the photos the owner engaged me to make an in-person inspection. They wanted both a more careful look and some advice on remediation. We went, we looked, and we watched as they opened up one of the cracked curbs. To my surprise, all the metal was clean (the tiny bit of rust on the edge of the bolted-down plate was insignificant).  The mortar that was used to fill the void formed into the precast to go around the plate and bolts was so crumbled that it might have been just coarse damp sand. It was clear that rust was irrelevant and that freeze-thaw action had turned the mortar into a sponge that once saturated could become a large ice cube capable of generating enough force to crack the precast. Once the cracking started, freeze-thaw and chlorides (the piece is on the edge of a bridge that is in the middle of the snow-and-salt-belt) did the precast in.

After cleaning out the mortar fill

After opening and clearing out most of the crumbled mortar

This piece has a barely visible early stage crack.

Lesson One: Sometimes you have to open up a wall to be sure of what’s going on. Sometimes you will get surprises rather than confirmation.  We understand the limits of looking only at what is immediately visible, at what shows up in a photo. True in property inspection and in many other parts of life.  More lessons from this inspection in posts to follow.

Lovely spring day makes everything look better

Today was a lovely spring day and I even got to take a little bike ride. That puts me in a very good mood. So one of those constant screw-ups that occur here on planetcommercialconstruction struck me as funny rather than annoying. Viewing commercial construction as a comedy of errors is not as unreasonable as it might seem.

In other industries people build prototypes to discover problems before they build the real thing. Since we build large-scale prototypes (that is what a new building or road or park is) we are constantly finding new possibilities for errors. Because so little is “standard” in this business, even processes tend to be non-standard. Last month I was surprised when a plant submitted samples of two shades of red, got an order for some mockup pieces with the right mix number attached, and somehow, despite apparent good business process, managed to make and crate up 4 full size gray pieces. I did not laugh when that happened, but it wasn’t yet spring.

Today’s laugh came in the context of a job that is getting overripe. I sent samples to the GC and they took two weeks to get to the designer.  Those were rejected because they had the color and the finish that was specified  (rather than  what was “intended” –  that’s life on planetcommercialconstruction). We have run out of time. So this time I gave the plant a transmittal with the name and address of the designer so samples could go direct and we could get approval and get this into production. This AM I saw an email from Fed Ex saying they had delivered the package. Great!  I sent the designer an email so he would know it was in his mail room. So when I showed up at the office and saw that a carton about the size of two 12″ square samples had been delivered to me I was surprised.

I looked at the label and saw that the plant  (not the same one as switched red and gray) had put the designer’s name on the label, but not his address. I had failed to read and verify everything on the FedEx email; had I done what every sub is contractually obligated to do, which is to read every one of the gazillion pages of paper, email, drawings, etc. that get sent to us in a project and scour them for conflicts, then I would have detected this earlier and avoided the embarrassment of telling a designer the sample had arrived at his office when it was actually at mine.  The good news is that I took a short bike ride today and so it’s just part of the comedy of errors of commercial construction.

Rare pleasant surprise

This posting was inspired by a request for a final waiver. Last summer I took on a tiny job – a single  beam cover (turned out to be 3 pieces when we were done). The job was a poster child for small aggravating jobs that take way too much time in relation to the revenue. It was one of the inspirations for the posting about designers being squeezed. https://planetcommercialconstruction.wordpress.com/2013/10/08/fads-fashions-and-trends-on-planet-commercial-construction/   There was no hostility in the process, just a lot of back and forth and coordination, all way out of proportion to covering a steel beam at a garage door. We were furnish-only and the erector was quite bureaucratic and would not use their normal lifting hardware because of an inconsequential hole size mismatch on a plate. I was so frustrated that I jumped in and ordered eye bolts without waiting for a change order, just hoping that the GC would pay me for the bolts and the fiddling time. (The problem was not mine but I wanted to wash my hands of it and while I normally hate taking risks on extras, it was very <$100). Late in December I was astonished and surprised to find in my mail box a check for the whole contract, including every change order and NO RETENTION. What a pleasant surprise. And this week’s email asking for a final waiver? It only mentioned the base amount.  The contrast between the smooth and rapid payment (at least on my end) and the slow and bumpy process of approvals and coordination was amusing. Ah, the planet of surprises, small and large!

Fads, fashions, and trends on planet commercial construction

 Some planets are static for long periods of time. Some not. One nice thing about Planet Commercial Construction is that it demands attention, so little is routine. Change is rampant.  When the a new or oddball situation repeats, I quickly wonder: is this randomness or a trend?

Three times this year I have been told to ignore the contract drawings.  On a small plaza renovation, my instructions came from the field.  On a decent sized downtown commercial addition, the GC was unable to ever give us a coordinated set of architectural and structural drawings; we were promoted so our shop drawings became the coordination focus. Today on a tiny job (but jam packed with confusion and complexity) where we flagged an A-S elevation conflict for the top of a beam we sat on, the GC said “ignore the contract set, just follow the steel shop drawings.”  Those drawings had followed the unconstructible contract details for supporting our piece and so they showed all sorts of extra plates and angles we did not need. Good bet that it’s a trend. 

What’s going on here? I will speculate.  All were private, CRE-driven work. Overall we (contractors and designers) remain in a depressed industry with low margins. Developers are much better at squeezing and wringing cost out of a project than are public owners and so the architects are working cheap. In turn the fees of they pay engineers are rock bottom. So what were once 80% plans are now “for construction”.  Subs who have to submit shop drawings encounter all sorts of issues but the EOR and maybe even the AOR is not eager to put more effort in.  Sort of implicit design build but under a design-bid-build contract. We keep hoping and expecting that the GC will play a serious coordinating role, but in my experience at best they dabble at it.

This process increases the chance or errors and problems. And despite direction to subs about ignoring the contract drawings, in the event of a serious dispute the GC has a huge ammo chest, since the sub is not likely to be in compliance with the contract drawings.  I find it nerve-wracking.  Am I the only one on the planet witnessing this? 

Benches, Sidewalks, and Construction Tolerances

I have a project in which one of the elements we are furnishing and installing is a bunch of large precast concrete benches. They appear on a landscape sheet that also calls for aligning bench ends with sidewalk scores.  Simple in theory; prone to problems in real life.

Bench layout and detail  (click to enlarge)

To avoid those problems I pushed on the GC and we met: me, the GC, his flatwork sub, the architect, and the landscape architect. Together we worked out an approach that had a good chance of avoiding problems with alignment. On my native planet (Earth) all measurements, dimensions, and manufactured items have tolerances. Tolerances (or accuracy if you prefer)  for car engine parts are a few thousands of an inch, in precast it’s a few 16ths, when tossing  hand grenade  it can exceed several yards.  Here on planetcommercialconstruction there is not uniform awareness of such tolerances.    As part of managing expectations and avoiding problems with our installed material, I asked the two architects about their assumed tolerances. Without hesitating, both said “none” – it should all fit perfectly. If you build, rather than design, you will of course be flabbergasted by this blatant disregard for the nature of material reality (or, disregard for the reality of materials). If you have years of experience, this reality will not be surprising  For those with interest in tolerances you can buy the book, or just go there and read the excellent introduction: http://www.amazon.com/Handbook-Construction-Tolerances-David-Ballast/dp/0471931519 )

Here I will note but not address the disconnect between that attitude (“no tolerances, everything fits perfectly”)  and the specifications that are full of stated tolerances and are an integral part of the contract.  I’d like to use this witnessed scenario as a springboard to speak to the intimately interconnected issues of:

•  inefficiencies in how we build structures on planetcommercialconstruction
• why progress with BIM and other software solutions to the challenges and cost of design is much slower and tougher than many people hope will be the case

Efficient movement of a design concept through all the steps that will result in a physical reality requires a lot of knowledge about many materials, products, and processes. Most folks who have not actually tried building complex modern structures cannot imagine how much knowledge is involved among all the specialists. In terms of the first issue above (inefficiency), because we involve people with knowledge post-design, post-bid, there is a lot of backtracking and meetings just like the one today. A lot of revision and re-design occurs.  The injection of relevant knowledge into the process earlier would be much more efficient in terms of all participants’ resources.

BIM relies on intelligent objects (doors, walls, benches, etc.). But if these objects do not have embedded within them full knowledge of their own tolerances they cannot be part of an automated design engine that delivers hoped-for results. Ultimately what they really need is the deeper,  more complex, and difficult to acquire and to  incorporate  knowledge of how tolerances of separate materials/products/assemblies interact. This is tough stuff to round up and harness.

You don’t have to trust the pronouncements here of a quirky precaster.  Check out http://dspace.mit.edu/bitstream/handle/1721.1/43747/263921735.pdf Joshua Lobel’s  master’s thesis. His title is wonderfully playful if you know the vocabulary:
Building Information: Means and methods of communication in design and construction.  He distinguishes between design information and construction information and notes that BIM may lead to loss of knowledge and further disconnect between the two.

A devilish grin amongst the details?

Some of the facts of life on this planet (commercial construction in the USA) :
• You are governed by lots of documents (contracts, plan, specs, ASI’s …)
• Plans often have conflicts, improbabilities, or impossibilities within them.
• Specs are almost always flawed, in serious or minor ways.
• The plans and specs are often in conflict.

When first going through the documents you cannot always know if the architect is serious or pulled an “oops”. Sometimes this is true on second pass. Sometimes resolution of a tiny detail can be prolonged. We have a job to supply pieces of precast concrete that hold plates that somehow help diffuse air through sewage during treatment. Or something like that: truth is I do not fully understand the system of which these pieces are but one part.

Section through precast diffuser plate holder

Full section  (Link is for those who want to see the section so big that they can read it)

One end has a special metal casting for the air inlet. This we embed in the precast concrete;  the metal piece is supplied to us by the GC (for my comments on that casting see https://planetcommercialconstruction.wordpress.com/2013/01/23/minor-mentoring-is-a-pleasure/ ). The other end has a lifting insert that we buy. Normally no big deal, but this one is unlike any ever seen by me or by the 3 other gray-haired buddies I consulted with.

Also unusual is the mention of bolting to forms – makes no obvious sense in the context. And the capacity of a 1” bolt is an order of magnitude more than the piece weighs. Why use a 1” lifter? Maybe there is a reason. Maybe not.

This being a very large (=bureaucratic) owner, I over-communicated to avoid rejected submittals. So both on the shop drawing and in a separate substitution request, I stated my assumptions, questioned the insert, and proposed something more appropriate. All in a manner both crisp and thorough. That was Feb 18 (this year). Since then I received a formal approval on everything but the substitution of the insert – it was stated it had to be for a 1” bolt. Normal, readily available inserts don’t look like this one and vaguely similar ones are deeper (because bigger bolts mean more capacity means more embedment in the concrete) and there is not room for a deeper insert where the insert goes in the piece. So I made a formal request for a vendor and part # for this now-elusive insert. I also again asked what this 1” bolt does. Yesterday (March 14) the GC called; he was still trying to get answers and said he would call an engineer at the agency.

Today, I still don’t know if there is a function for the 1″ bolt (GC may not even use the insert for lifting) , or if this is an insert that exists in another realm or on a different planet (we looked at plumbing catalogs but nothing matched) or if the reviewer who rejected it just is a stickler for enforcing whatever someone else once wrote down . Or maybe someone is “just kidding” about the 1”. Time may tell. I’ll let you know. Those of you who live on other planets may find this story bizarre, but anyone  living on planetcommercialconstruction will find it quite normal.  While this uncertainty about simple things can be annoying, it keeps this business from being predictable, routine, and boring.