In CANDU reactors, a contact between the Pressure Tube (PT) and the Calandria Tube (CT) could cause the development of hydride blisters that may lead to delayed hydride cracking of PTs. To maintain PT integrity, contact between these two tubes must be avoided. Under normal operating conditions (NOCs), the PT-CT contact can occur as a result of in-reactor deformation due to irradiation induced creep, irradiation growth and thermal creep of the fuel channel (FC) assembly. Numerical models have been developed to predict the time of PT-CT contact in a channel, which is required to ensure the operability of the channel and the plan for maintenance. Since the prediction of time to contact is influenced by various uncertainties, such as change in, (i) the dimensions of the FC, and (ii) the material properties and boundary conditions of the FC, probabilistic simulation-based methods have been developed to assess the PT-CT contact risk and establish adherence with provisions of the Canadian Standards Association (CSA) Standard N285.8. Numerical simulations based on a 3-dimensional finite element model (3D FEM) are essential in predicting the dimensional changes occurring during in-reactor deformation of single FCs because, unlike 1-dimensional FEM, 3D analyses correctly take into account the triaxial loading conditions in the calculation of the complete strain tensor. However, 3D FEM analyses are not practical for probabilistic assessments of an entire reactor core with 380 or 480 FCs. This paper (i) outlines the steps that are required to develop a 3D model of a FC using the ABAQUS commercial code along with user subroutines that incorporate the constitutive equations describing the in-reactor deformation of FC materials, and (ii) proposes a new surrogate model of a much simpler analytical form, which can replace the 3D FEM for assessing PT-CT contact in carrying out whole core probabilistic assessments. The proposed surrogate model not only simplifies the FE analysis and provides a more rational basis for probabilistic assessments, but most importantly allows the inclusion of diametral expansion and wall thinning in the prediction of PT-CT gap. This simple function representation gives a cost-effective and computationally efficient approach to probabilistic analysis of FCs.
Read full abstract