Predicting the Expansion of Concrete Exposed to Sulfate Attack with a Regression Model Based on a Performance Classification

11 Oct 2018  ·  Xiangru Jian, Paulo J. M Monteiro, Kimberly E. Kurtis ·

This paper mainly describes the development of a new type of regression model to predict the long-term expansion of concrete subjected to a sulfate-rich environment. The experimental data originated from a long-term (40+ years), nonaccelerated test program performed by the U.S. Bureau of Reclamation (USBR). Expansion data of specimens composed of 54 different mixtures were measured periodically throughout the entire test program. In this analysis, the mixtures were first classified into three groups using K-means clustering based on their expansion patterns. Within each group, the expansion rate was predicted as an exclusive regression function of the water-cement ratio (W/C), tricalcium aluminate (C_3 A) content of cement, cement content of cement or time of expansion. Then, a support vector machine (SVM) was employed to determine the classification criteria by relying on the characteristics of the mixture proportions rather than the experimental performance, thereby enabling the model to offer predictions for new mixtures without test data. An analysis of the model indicated that concrete specimens with different mixture proportions, especially with different W/C values or C_3 A contents, are unlikely to share identical expansion patterns and should be considered and predicted separately.separately.

PDF Abstract
No code implementations yet. Submit your code now

Categories


Applications

Datasets


  Add Datasets introduced or used in this paper