Advancing a comprehensive understanding of concrete durability.

摘要

The durability of concrete is vital to the health of our infrastructure and economy. Unfortunately, traditional, linear approaches to experimentation and specification have not provided a complete understanding and control of this complex material. This work seeks to further a comprehensive understanding of concrete durability through approaches including multi-variable mathematical modeling of premature deterioration, analysis of the microstructural changes of cement paste and mortar during freezing, and the development of quantitative computer models in parallel with actual experiments.; The first section includes a summary of the collection and analysis of data describing concrete pavements in a number of Midwestern states. This involves the development of a thorough survey, completion of that survey by local organizations, and the creation of a database from pooled data. Statistical analysis of that data reveals a number of statistically significant trends in a pavement's tendency to deteriorate. Statistically important variables include total alkali and sulfate content of the cementitious material, the presence of type C fly ash, ambient paving temperature, age, and permeability of the base course material.; In the second section, deformation mapping of cement paste during freezing illustrates the power and accuracy of this new technique. It also illustrates the effects of water/cement ratio, age, presence of aggregate, and repeated cycling on deformation during freezing. Finally, the increased resolution of this new technique allows for the identification of a number of new microstructural features of the freezing process.; Computer modeling closely mirrors the deformation analysis described above. Effects of water/cement ratio, age, and the presence of aggregate are all predicted by the models. Also, the trends and magnitude of bulk deformation predicted are very similar to measured results. Any differences can be attributed to the elastic nature of the models, as opposed to the inelastic nature of the experiments. This broad agreement of prediction and results serves to illustrate the power and accuracy of both the deformation mapping technique and computer models.