Development of an image analysis code for hydrided Zircaloy using Dijkstra's algorithm and sensitivity analysis of radial hydride continuous path

Abstract

PROPHET, a software that conducts an image analysis of hydrided Zircaloy for Radial Hydride Fraction (RHF) and Radial Hydride Continuous Path (RHCP) evaluation using Dijkstra's algorithm was developed. Enabling autonomous image quality adjustment capability, and user-defined noise and hydride detection sensitivity control options, PROPHET instrumentally ensures limited sensitivity to potential variation in quality of the Optical Image (OM) and post-characterization of hydrided Zircaloy. Sensitivity analyses were conducted for magnification and resolution of OM, input fracture toughness, and employed algorithms. x200 and 0.8M (i.e., 1024×774) were recommended for magnification and image resolution, respectively. Fracture toughness ratio of 1:50 for hydride to Zr-matrix was confirmed to be a sound choice and RHCP exhibits limited sensitivity to fracture toughness around fracture toughness ratio input of 1:50. Algorithm has a large effect on RHCP. Dijkstra's algorithm exhibits unmatched fidelity to identifying the least-resistant path (hence maximum RHCP). RHCP and RHF exhibit notable varations along the angular positions of a homogenously distributed hydrides on the surface of a tubular Zircaloy. The uncertainties of RHCP and RHF due to spatial hydride distributions call for (1) multiple image analysis for a statistically reliable RHCP evaluation or (2) low magnification but ultrahigh resolution OM to capture the significant fraction, if not all, of the entire cladding surface of interest. RHCP successfully correlates strain energy density of hydrided Zircaloy with limited accuracy. Introduction of a kernel function which instrumentally adjusts the importance of one or more factors that affect the mechanical behavior of hydrided Zirclaoy (i.e., hydrogen amount or RHF) may complement the accuracy of prediction.