A Dynamic Identification Method of Metamorphic Relation Based on Separation of Input Pattern and Output Pattern

Abstract

Due to the high cost and extended period of the benchmark development, there are few verification examples for verifying nuclear power independence software. This case is named the test oracle problem. As a result, it is a challenge to adequate verification. Because metamorphic testing is no need to construct verification examples, it is critical to alleviating the above problem. Some domestic independent nuclear design software has successfully applied metamorphic testing in the verification activities. Since Oracle problem has become a common challenge in software verification in nuclear power field, metamorphic testing has broad application prospects. Aiming at great arbitrariness and low efficiency of existing methods, we introduce a dynamic identification method based on the separation of input and output patterns. In this paper, this method is applied to the burnup program (namely NUIT) to determine a set of metamorphic relations between burnup depth and nuclide density. Other burnup programs can directly apply them and learn from this example. And this method can be extended and used in the different professional nuclear fields, such as reactor physics, thermal hydraulics, source term, and radiation shielding.