The development of safe, reliable, and economic products requires their optimal design. For this purpose, the parameters in the model equations should be known precisely. In reality, however, this is usually only rarely the case, since the parameters often exhibit a more or less high degree of uncertainty. Such uncertainties are either the result of a random process (aleatoric uncertainties), like variations in the manufacturing process, or are caused by a lack of information (epistemic uncertainties). Examples of the latter are vagueness or lack of knowledge in the parameters or boundary conditions. Consequently, suitable methods are necessary which allow for the inclusion of parameter uncertainties into engineering design computations.

Against this background, an analytical approach to evaluating equations with fuzzy numbers as parameters was developed, which allows for the inclusion of parameter uncertainties into engineering design computations. Furthermore, suitable sensitivity measures are provided that allow a quantification to which extend the uncertainties of the single model parameters contribute to the overall uncertainty of the model answer.

Kontakt: Dipl.-Ing. Arthur Seibel