Reverse English on Skagit Collapse


It took about two minutes for the Columbia River Crossing spin machine to suggest that the Skagit Bridge collapse proves we need to build the CRC.

Two opinion pieces this weekend volley the ball back over the net, with the real policy point – we need to prioritize maintenance:


4 responses to “Reverse English on Skagit Collapse”

  1. An over-size Canadian truck impacts an arched transverse girder on a Skagit bridge; instantly the bed of the Columbia liquifies and pilings sink.

    Time for some theory of structures? Yes!

    This is only the latest collapse of a major bridge in Washington State, and it never will rise to the fame of the original Tacoma Narrows event. Tacoma Narrows was plenty STRONG–just look at the abuse it took before failing! But its deck was not STIFF enough to retain form under load. Even worse, it had not been designed to be STABLE under normal aerodynamic loads. The three criteria for SAFE STRUCTURES are, in descending order of importance, STABILITY, STIFFNESS, STRENGTH. All three are necessary, but Tacoma Narrows demonstrated that the second and third are irrelevant if the first does not obtain.

    No one was killed at Tacoma Narrows either. Sometimes Providence protects us.

    Was the Skagit bridge poorly designed? Yes, and no.

    The first major structure in steel was the Eads Bridge in my hometown of Saint Louis. Designed and built in the era of horse-drawn wagons and steam trains, it still is doing very well after nearly 140 years, thank you. Structural engineering then came of age, with riveted steel members designed according to the method of “allowable stress.” The idea is pretty simple: if steel has an ultimate strength of 20,000 pounds per square inch (20 ksi in shorthand) we shall design our bridge so that no member is stressed to more than, say, 14 ksi under anticipated loading, and all will be well.

    This worked extremely well in practice. Our old Willamette bridges, Broadway, Steel, Burnside, Hawthorne, Ross Island, Sellwood comprise an encyclopedia of the versatility and durability it enables. Conde McCullough and David Steinman epitomize this era bridge building, locally, nationally, internationally.

    After WWII things changed. Steels became stronger and engineers sought to utilize that strength through more sophisticated methods of design, thus reducing mass and cost of bridges. “Ultimate strength” and “plastic modulus” design criteria became popular, and they were not so bad.

    The second Interstate Bridge and the Morrison Bridge are from that era, and they readily exhibit the principal effect of the new criteria: stand on the middle of the Burnside and it is solid as a rock when heavy vehicles pass; stand on the middle of the Morrison when a truck passes and the vibration can induce vertigo. Same deal with the old and new Interstates. Why?

    Steel is a marvelous material. It can be alloyed and treated to assume tremendous strength, but its stiffness (elastic modulus) is fixed by the atomic lattice bonds of its principal element, iron. So a bridge built with “stronger” steels inherently tends to be less stiff, simply because there is less material and fewer lattice bonds to resist deformation.

    STIFFNESS and STRENGTH are not covariant.

    And the real kicker remains STABILITY under expected loads.

    Loads on a bridge are the trickiest things to figure, vide Tacoma Narrows. Modern design puts this first with the LFRD method: Load Factor Reduction Design. Every conceivable combination of various loads is estimated, each kind of load (traffic, temperature, wind, seismic) being weighted according to its probability, and the bridge is designed to meet the worst case realistically expected over its life.

    The Skagit bridge was not designed to this standard, but during its history it was, apparently, impacted a number of times before the fatal blow was struck last week. My point is that according to LFRD philosophy that sort of load could and should have been considered retroactively, and measures taken preclude its occurrence. We cannot control wind or seismic loads, but we can control traffic; it is the only thing we can control with certainty.

    Not understanding that fact of engineering life is an enormous failure by WASHDOT, especially its Executive Director, and by extension, Jay Inslee, the Governor of the state.

    It has been stated that the Sakgit bridge was “fracture-critical.” This is misleading. Photographs show that movement of the transverse arched girder displaced the upper constraint of a vertical strut in compression, which then buckled, transferring load to other members. Those radically deformed in response, some perhaps fracturing, causing catastrophic failure.

    The initial event was displacement of the terminus of the compression member: a disturbance of STABILITY. The second event was buckling of the compression strut: a failure of STIFFNESS. The final event was the mass of twisted and fractured metal: a failure of STRENGTH.

  2. This is a very informative post marred by the ridiculous claim that the engineering failure of a bridge built 59 years ago is the fault of a governor who took office four months ago.

    By “control traffic” do you mean equip the bridge with over-height radar?

    Before I knew of the difference in clearance between the left and right lanes I thought that was a proper thing to do for this bridge.

    But because the right hand lane has considerably lower clearance than does the left, radar wouldn’t do the job. Even if the beam were angled to match the profile of the bridge clearance, a vehicle clearing the radar in the left lane might then move to the right lane and strike the structure.

    It seems to me that the only way to protect the bridge is to mandate that all vehicles higher than the clearance of the right lane must follow the current detour path across the Riverside Drive bridge. And I will agree that it seems pretty stupid that Washington DOT hasn’t done that already, since the bridge apparently has been hit repeatedly in the past.

    But, again, “the buck stops here” is about policy not technical execution. Is it the governor’s fault if hackers break into Access Washington and steal drivers license records?

    Maybe in some indirect “you should have set and enforced a policy against that” way, but the State Legislature has to share that same responsibility, since they constantly are tying the hands of the executive branch.

    Rightly in many instances, but with power comes responsibility.

    It’s really the concrete lobby in Olympia that bears the blame, because they fight tooth and nail to ensure that every penny collected at the pump goes to building new roads and bridges made from their favorite construction material.

  3. Didn’t the Hood Canal and original 1-90 floating bridges sink? What gives up there in WA when it comes to bridges?

Leave a Reply

Your email address will not be published. Required fields are marked *