Uncertainty modelling of large deformation process using a new aggregation rule of evidence

Abstract

The aggregation rule is critical when a high degree of conflict between different sources of data or expert opinions exists. In this paper, we will present a new evidence aggregation rule that makes use of Ground Probability Assignment (GPA) developed by Yager along with a proposed Credibility Factor of Evidence (CFE). Observed experimental evidences on simulation responses or commonly used values of uncertain parameters are used for the CFE estimation. The new aggregation rule is adopted for uncertainty modelling of a large deformation process represented by the Taylor impact test for which Johnson-Cook and Zerilli-Armstrong plasticity models provide different answers. The uncertainty modelling procedure also accommodates both epistemic and aleatory uncertainty embedded in material constants of the adopted plasticity models. Results of uncertainty representation, propagation and quantification for Taylor impact tests of AISI 4340 Steel have shown that the suggested aggregation rule is very efficient in recognising key information or knowledge from different sources, which helps to reduce epistemic uncertainty.