The growing complexity of engineering problems necessitates the development of advanced modelling approaches that address not only the correctness of numerical treatment of models but also the process of model creation. In particular, mathematical models are based on modelling assumptions, which can be violated even before computing with the model. Thus, it is necessary to distinguish between conceptual modelling errors, related to breaking basic modelling assumptions, and modelling errors related to practical realisation of models, such as the quality of computer implementation and approximation quality. While the errors related to practical realisation of models are well-understood and studied, the conceptual modelling errors are much more difficult to address, because a deep understanding of the model creation process is required.