In order to evaluate the influence of certain experimental or material-related conditions on results of the rapid chloride migration test (RCM test), statistical tests within and between samples are necessary. Thus, it needs to be clear which scatter a sample or a population of results is already subjected to without external influences due to the test method itself. So far, however, literature values for the appropriate statistical variable (coefficient of variation, CoV) explicitly valid for mortar or fine-grained concrete, e.g., for concrete repair, are missing. Therefore, we suggest a specific mortar CoV based on our own results of RCM tests performed on a cement-rich, sprayable mortar based on ordinary Portland cement. For the evaluation of external influences on a sample in comparison to a reference sample, we developed a significance criterion based on a statistical hypothesis test. The sensitivity and the reliability of this criterion is demonstrated on various results from RCM tests on mortar specimens, according to the test specifications and with deliberately chosen deviations from it. In addition, we point out the parameters included in the calculation of the rapid chloride migration coefficient that are the most sensitive to unintentional errors.
Keywords: coefficient of variation; rapid chloride migration test; repair mortar; t-test.