The latter one refers to the uncertainties due to lack of knowledge, which can be decreased by the
application of additional data or information, better modeling, and parameter estimation methods. It should be
emphasized that in the reliability modeling, it is possible to divide the second kind of uncertainty into statistical
uncertainty and model uncertainty, whereas the first type of uncertainty is called random variation (or physical
uncertainty, noise factor). Statistical uncertainty refers to estimation of model parameters based on the available data
where the observations of the variable may not represent the real situation perfectly, and thus, the recorded data may
be biased. Additionally, different sample data sets usually provide diverse statistical estimates. Model uncertainty
results from the use of one (or more) simplified relationship which is supposed to represent the “real” relationship or
phenomenon of interest. Such an approach results from lack of knowledge or increased availability of data. Another
important kind of uncertainty is related to the uncertainties due to human factors. Such uncertainties result from
human errors and interventions undertaken in the design, manufacturing and operation. For example, they can be
caused by misuse, gross errors and human mistakes [22,23,24]. They can be considered by creating robustness
through product changes or using an extra safety, however, in practice they are primarily subjects to quality
management [10].