Optimization of X-ray parameters in radiography nondestructive evaluation

Abstract

A new tool is being developed which automates the optimization process of X-ray NDE. The tool incorporates an adaptive simulated annealing model for exploring the large dimensional parameter space and determining the globally optimum state. The problem is cast in the form of minimizing a cost function, defined in terms of film density contrast values corresponding to a specified set of flaws in the object. The XRSIM model is invoked to calculate film contrast values. The object flaw set can be determined using finite element analysis techniques, if the object loading is known. The centers of the flaw set are chosen to coincide with the centers of elements with highest stress values. A probability of detection (POD) model is used to test the obtained optimization values. If POD values are found to be unsatisfactory, the optimization process proceeds to obtain, two, or more, parameter sets, which can be used for an optimal inspection of the object.