Statistical Analysis for Autoclave Expansion of Sulphate- Resisting Cement

Abstract

Excessive volumetric change may have great influence in collapse of hardened cement mortar. Autoclave expansion test provides an index of potential delayed expansion and thereby the soundness of cement mortar. This paper presents an effort to model autoclave expansion in relation with factors affecting soundness of cement mortar.A statistical analysis is performed for a data base of 575 chemical and physical tests obtained from archive of quality control at Kerbala plant for SRC cement. These data sets are refined to 444 by excluding outliers to avoid errors in measurement, recording, and observations. Multiple linear regression by SPSS16 software is used to develop a statistical model relating autoclave expansion with some explanatory variables (MgO %, F. CaO %, C3A %, C3S %). Stepwise regression shows that linear model explains 89.4 % of variation in autoclave expansion by percentage of free lime to 0.05 degree of significance, while excludes other proposed variables. If the degree of significance is changed to 0.1, C3A may be included in the statistical model as second explanatory variable. The low percentage of MgO makes its effect insignificant for inclusion in the developed model. Also, the analysis shows that the restricted percentage of Tricalcium Aluminate C3A in the sulphate resisting cement may serve positively in mitigating the expansion of hardened cement mortar. On the other hand, the developed relation may give indication to control a recommended percentage of free lime in the raw materials to assure allowable soundness of cement.