A Sea Fight in a Fog: Revisiting the ASCE Controversy about Dynamic Formulae

With another North American academic year getting under way, it’s time for geotechnical professors and practitioners alike to think about where our industry has been and where it’s headed.  This past summer three of our most eminent people (Garland Likins, Bengt Fellenius and Robert Holtz) came together and wrote an excellent piece for the 2Q 2012 issue of the Pile Driver (the official publication of the Pile Driving Contractors Association) on “Pile Driving Formulas”.  The article is centred on the 1941-2 discussion (that’s putting it politely) in the Proceedings of the American Society of Civil Engineers on a committee report on the subject.

Last summer I added to my site on Vulcan Iron Works a history of pile dynamics from the Sanders (not Stanton, as the PDCA article states) Formula to the present.  I considered adding a page on this particular topic, but with a website like this where the compensation is minimal, time failed me.  In a sense, the PDCA article fills an important gap in the narrative.  Having said that, I can’t help but think that the authors took some inspiration from what I had posted.  For example, the inventor of the Engineering News formula is not generally referred to as “Arthur Mellen Wellington” and the quote from his work is the same as the one I used, the more “spiritual” part excised.

In any case, the number of pile driving formulae increased in the first three decades of the last century to the point where, in 1930, the American Society of Civil Engineers appointed a committee to study the issue and recommend a formula.  The timing was interesting.  The following year, the Australian civil engineer David Victor Isaacs published his historic paper which first identified (and developed a method to analyse) wave propagation in precast concrete piles.  Later in the decade the British Building Research Board did their extensive research on wave propagation in piles.  The civil engineering world was taking its first steps to get beyond simple Newtonian impact mechanics in the dynamic analysis of driven piles.

The Committee finally released its report in 1941.  One recommendation was that static load tests be used in place of dynamic formulae.  This was definitely one way to solve the problem, but static load tests are long and expensive, and neglect the use of the pile hammer as a measuring tool.  Another proposed a refinement of existing dynamic formulae.  At this point the controversy erupted.  From September 1941 to February 1942 the discussion raged in the Society’s proceedings.  It involved many of the “greats” of geotechnical engineering: Karl Terzaghi, Ralph Peck, Arthur Casagrande, Gregory Tschebotarioff, Lazarus White and many others.  As is often the case in the earth sciences, from global warming to earthquake engineering, it sometimes got heated and emotional, with some defending the status quo while others pointing out the inadequacies of dynamic formulae.  The PDCA article does an excellent job in distilling this discussion to its basics.

While the end result—a “new” dynamic formula was not imposed on the industry—was a satisfactory one for the moment, the discussion revealed a great deal about geotechnical engineering, some of which has changed and some of which has not.

The first problem was that, for all of their erudition and well-deserved reputation for expertise in the field, many of the commenters were not, for want of a better term, well versed in the ins and outs of things moving, especially as rapidly as takes place during wave propagation in piles.  It is to their credit that the pioneers of this profession were able to transform a profession from a strongly empirical one to one subject to engineering analysis and quantification, and doing so in an environment of complexity and unpredictability as the earth itself.  But the skill set required to do that didn’t always lend itself to the understanding of the phenomena seen in driven piles during driving.  In that respect, the controversy resembled one description of the Trinitarian controversies of the fourth century: a “sea fight in a fog”.  While the shortcomings of the dynamic formulae were clear to those who spent time the jobsite time that these men did, the solution for the problem would have to come from somewhere else.

The second problem was that the computational power needed to analyse the problem was lacking at the time, especially to the practitioners in the field.  Isaacs solved this problem by using a graphical method, a solution seen elsewhere in the profession, but making his method general practice would have involved some kind of instrumentation to verify the results.  On the other hand the BBRB came up with the instrumentation, but their analytical method—a type of d’Alembert solution of the wave equation—was far too complex for practical implementation at the time.  Neither of these methods, even if they had been combined, adequately addressed the soil response to impact, especially along the shaft of the pile.  But in any case the Committee’s inclusion of these methods was not a significant part of their work product, and World War II put a stop to the research.  It’s tempting to think that, without that great and destructive conflict, a workable solution could have been proposed a decade earlier than it was.

The third problem was the frequently unhelpful role of building codes and standard specifications.  Codes enable owners to insure that their work is done properly.  One way they do this is by specifying methods of verification that are both easy and repeatable to evaluate.  What’s “easy” depends upon the tools of the time, but one of the reasons it has been so difficult to displace the dynamic formulae from geotechnical practice is because they—and especially the Engineering News formula—became deeply embedded in the codes and specifications by which many structures were built.  To take these away required their replacement, and risk averse owners of all kinds were reluctant to do this.

As the PDCA article rightly notes, the most prescient commenter was Raymond Concrete Pile’s A.E. Cummings, who noted the existence of Isaacs’ and the BBRB’s work on wave propagation in piles.  This is no accident.  Raymond was involved in every aspect of the installation of driven piles, from the design and manufacture of the driving equipment to the load testing of the piles.  They had a more comprehensive view of the issues involved and, being a large organization, had the means to tackle the problem.  Combined with the advent of digital computers, Cummings’ Raymond colleague E.A.L. Smith was able to write the first code suitable for the analysis of piles during impact driving, and the rest, as they say, is history.

Today of course the analysis of wave propagation in piles, both predictively and inversely, is at the core of pile dynamics.  It’s worth noting, however, that, although there have been many refinements in the methodology and advances in the software used, the basic theory in use is ostensibly the same as it was in the 1970’s.  It’s also worth noting that the use of pile dynamics is still a very specialized business, not only because they involve deep foundations, but also because, as was the case seventy years ago, most geotechnical engineers (except those in research) are not specialists in dynamics or numerical methods, both of which are at the heart of the analysis of piles (and other deep foundations) during impact driving.  Finally, although it’s been a long process to displace the dynamic formulae with wave related methods of analysis in building codes and specifications, it’s unreasonable to say that newer methods will not come along to displace or upgrade them, even in this conservative industry.

One of these days, significant breaks with current practice will appear to be considered.  Hopefully we won’t go through another “sea fight in a fog” as we did in the 1940’s and make the transition to newer, vetted methods smooth and efficient, for the benefit of both our profession and for those who use the structures we design and build.