Application of the ‘lightweight refinement’ relation to establishing confidence in safety assessment models

This paper presents a ´lightweight refinement´ relation that enables rational comparison of safety assessment models of a system. The comparison process contributes to establishing confidence in adequacy of the models by identification of apparent inconsistencies that require explicit justification. The paper further demonstrates how the ´machinery´ of refinement can be applied to individual models to identify key ´simulation cases´ and aid the task of model review.