Computational Materials Science of Concrete
To better see how the computational materials science of concrete fits into the field of concrete research, it might be good to first discuss its intellectual sources, at least from my point of view. First was the work on the structure of amorphous semiconductors like silicon and germanium in the 1960’s and 1970’s. Physicists had before developed crystal physics to a high degree, and had even allowed for crystal defects like dislocations. However, the problem of amorphous semiconductors, or of glass, was entirely different - there was no underlying crystal lattice. How was one to do any calculations at all? Analytical approximations were tried, with only a limited degree of success [17]. Then models were built with several hundred atoms, pushing the computing power back then to the limit, which were linked together randomly [18]. Algorithms were applied to these models to compute properties, which were then compared to experiment in an attempt to explain the experimental results.
Second and third came two developments in the materials science of concrete community, which appeared to be unrelated to the previous amorphous semiconductor work, but which were similar to it. These were both highly original, highly innovative developments. In 1984-5, Wittmann, Roelfstra, and Sadouki published two important papers on numerically simulating the structure and properties of concrete in 2-D [19]. In these papers, 2-D models were developed for simulating the shape and arrangement of aggregates in concrete with finite elements applied to compute properties like thermal conductivity and elastic moduli. In the very next year, 1986, Jennings and Johnson published their work from the project led by Geoff Frohnsdorff on a 3-D model of cement paste microstructure development for C3S pastes [5]. This was the equivalent of the amorphous semiconductor models, but at the cement particle scale, not at the atomic length scale. Particles of various sizes that followed a real cement particle size distribution were dispersed randomly in 3-D. Various rules were applied to these continuum spherical particles to simulate the dissolution of cement and the growth of hydration products.
The fourth development was a paper showing how a random walk algorithm could be applied to models of porous materials to compute electrical and diffusive transport [20]. The combination of the ideas of random walks, digital images, and the existing cement paste hydration microstructure development model [5] led us directly to the first NIST digital-image-based cement paste hydration model [9]. Fortunately, at the time of its first development, 1989, we had just enough computer power to barely implement a 3-D model of sufficient size (1003 pixels). Once the model was on a digital lattice, almost any finite element or finite difference algorithm could be readily applied to it, so that almost any physical property could be simulated and compared to experimental results.
The subsequent spread of the computational materials science of concrete has been aided by two developments in information transfer. Information transfer, electronic and otherwise, was also one of Geoff Frohnsdorff’s constant themes, as any of his acquaintances knows well. The first was the Computer Modeling Workshop, started in 1990. In June, 2006, the 17th annual Computer Modeling Workshop took place at NIST. As of this workshop, we have had a total of 473 attendees, of whom one third were from institutions outside the U.S. Out of that total, there were 225 students and post-docs, 87 faculty members, 99 industrial researchers, and 62 government researchers. Many current research leaders have attended this workshop. I note that at this present Quebec conference, the numerical modeling track forms the first of another series of concrete modeling workshops, run by Prof. Klaus van Bruegel of Delft University. A second workshop is planned in 2008 in the Netherlands. Delft is one of the centers in Europe for concrete modeling and so it is good for a series of such workshops to be run in Europe as well as in the US.
The second information transfer development has been the NIST Electronic Monograph: Modeling and Measurement of Cement-Based Materials [21], which is accessed by over 10,000 individual users every month, from over 90 countries. The origin of the Monograph is due, in a large degree, to Geoff Frohnsdorff. In 1997, a technical publisher asked me if I was interested in writing a book on the computer modeling of concrete. I was interested, but asked my group leader, James Clifton, who then asked his boss, Geoff Frohnsdorff. The answer that came back to me, mainly from Geoff, I believe, was that as a government scientist I was not allowed to write a book for a commercial publisher, but that I was encouraged to write a book to go on the web. In 1997, at least to me, a hardback book on my bookshelf was a lot more prestigious than something on the Internet, which I had not used much at that point. In retrospect, I’m glad that Geoff made that decision. The Monograph turned out to be a collection of papers in book form, and I have been able to add to it and update it like I never could have done with a hardbound book on my bookshelf. It has also had a far wider circulation than any hard copy book I could have written.
The current state of computational materials science of concrete at NIST is embodied in the Virtual Cement and Concrete Testing Laboratory (VCCTL) [22], which is being developed at NIST with the active cooperation of eight leading corporations and associations in the concrete field: the US National Stone Sand and Gravel Association via the International Center for Aggregates Research, Degussa Master Builders, Sika Technology AG, W.R. Grace, the US Portland Cement Association, Association Technique de L’Industrie des Liants Hydrauliques, Verein Deutscher Zementwerke eV, and the US National Ready Mixed Concrete Association (until 2006, Holcim US, Inc. and Cemex were an important part of the consortium and founding members). Computer power and model effectiveness have grown together over the years so that this large, integrated software package can be designed to mimic a complete physical testing laboratory, with databases of cement and aggregates instead of bins and hoppers, material combination and concrete curing models instead of mixers and molds, a software interface instead of a cart to take materials and samples around the laboratory, and accurate models for performance prediction instead of instrumented testing machines. The VCCTL exemplifies the idea expressed in this paper that we need to make physical testing “smarter” by measuring fundamental material quantities that are needed by these kinds of models to be able to unleash their full predictive power.
Next: Prospectus Previous: Experimental Materials Science of Concrete