Model verification, refinement, and testing on independent 10 - year carbonation field data

ABSTRACT: An extensive range of data generated from a 10-year long-term experimental and field carbonation study was used in this work to test, validate and refine an independent existing model for carbonation prediction in reinforced concrete structures. These data are detailed in the CSIR, 1999 report. In the report, data was generated for twenty one (21) concretes mixtures of nominal strengths 25, 35, 50 MPa made using ordinary Portland cement (OPC) or rapid hardening Portland cement (RHC) with or without 15, 30, 50% fly ash (FA), each mix being subjected to five (5) different curing regimes prior to exposure of samples to the natural Pretoria weather. Strength and carbonation measurements of the field exposed specimens were determined at the ages of 3.5, 6 and 10 years, along with their compressive strengths. Also determined were compressive strengths at 1, 3, 7, 28, 90, 100 days, 3.5, 6, 10 years for specimens stored under laboratory conditions. It should be underscored that the model origin was not in anyway related to these data. But when tested using laboratory data from the CSIR investigation, the model prediction of carbonation rates was remarkably accurate. The model also accounted for the cement or binder type quite well. However, when field strength data was used in the model, the prediction results diverged from measured carbonation rates. In analysing the strength data, it was found that long-term field data gave higher compressive strengths than the corresponding laboratory strengths. These observations can be attributed to the effect of carbonation on exposed samples, leading to relative strength increase. The model has been modified by the author in order to account for the carbonation effect on strength. The good performance by the model on independent data demonstrates its underlying robustness for use in service life prediction