(4) Virtual Cement
This monograph describes, in a unified fashion, the results of our modelling research since 1989. There are two types of models: microstructure models and property calculation models. Both will be covered in this monograph. As of 2000, experimental results from the Inorganic Building Materials group at NIST have been added, on concrete rheology and x-ray and scanning electron optical microscopy of cement and concrete.
Over the last 20+ years, computational materials science has developed as a scientific discipline, powered by the revolutionary advances made in computer processing speed and memory capacity. The main application of this discipline has been to random materials, where analytical approaches are inadequate. This applies especially to concrete, which is a random material over length scales ranging from nanometers to meters.
This monograph is an attempt to summarize what has been developed thus far in the computational materials science of concrete, or often called the computer modelling of concrete. The monograph focuses on the work done at NIST, as this has been the main center for the computer modelling of concrete in the 1990's. Developments elsewhere are also covered in this monograph. Experimental results have been added because the computational materials science of concrete does not stand alone as a discipline. Experiment and theory must work together, and progress in one results in progress in the other. The idea of the synergy of theory and experiment, which has been actively pursued in physics for centuries, must now be followed also in the materials science of concrete, in order for real progress to be made in understanding, controlling, and optimizing this complex material.
The main idea that permeates this monograph is that of concrete as a multi- scale material, from nanometers to millimeters and beyond, with a different microstructure on every scale. The approach at every length scale is to develop a model of the microstucture, so that a physically realistic microstructure can be built in the computer. This construct is then operated on by various algorithms (finite difference, finite element, random walk,etc.) to predict physical properties. Mathematical techniques are then used to link these predictions together between different length scales so as to provide a true multi-scale picture of a microstructure-property relationship. To date, this final scale-linking step has only been carried out for ionic diffusivity. Also, because of the way the work has gone, there is a lot more material in the monograph on the micrometer (cement paste) and millimeter (mortar and concrete) length scales than on the nanometer (C-S-H) length scale.
This monograph is intended to be a document "in-process", with more material added as more research is carried out. We hope that this monograph, in this form, will be useful to readers. E-mail messages detailing problems with the monograph, or telling of the monograph's usefulness, are welcome.
(1) E.J. Garboczi, D.P. Bentz and G.J. Frohnsdorff, Proceedings of J.F. Young Symposium, Materials Science of Concrete Workshop, Lake Shelbyville, IL, April 27-29, 2000.
(2) E.J. Garboczi, D.P. Bentz and G.J. Frohnsdorff, Concrete International, 22 (12), pp. 24-27, December 2000.
(3) D.P. Bentz, J.R. Clifton, K.A. Snyder, Concrete International, 18 (12), 42-47, Dec. 1996.
(4) J.W. Bullard, C.F. Ferraris, E.J. Garboczi, N.S. Martys and P. Stutzman, Chapter 10.3 in Innovations in Portland Cement Manufacturing, J.I. Bhatty, F.M. Miller and S.H. Kosmatka, eds., Portland Cement Association, pp. 1311-1331 (2004).
(5) E.J. Garboczi, J.W. Bullard, D.P. Bentz, Concrete International 26 (12), 33-37 (2004).