Pulsed Eddy Current Inspection of Broach Support Plates in Steam Generators
Abstract
Steam Generators (SGs) are a critical component of nuclear reactors, employing thousands of SG tubes to convert heat generated in the reactor core into useable energy. SG tubes are supported at numerous locations by Broach Support Plates (BSPs) that have trefoil shaped holes, which prevent excessive tube vibrations, while allowing water to easily flow through the support structures. A number of degradation modes occur in SGs, such as SG tube fretting, cracking or denting, requiring periodic inspection. Currently, conventional Eddy Current Testing (ECT) is used to non-destructively assess the condition of SG tubes and components. However, as reactors age, new modes of degradation will likely appear that may be difficult to detect and characterize using conventional ECT, such as wall loss in BSPs and build-up of corrosion products, which typically form as a hard sludge called magnetite. Pulsed Eddy Current (PEC) technologies are an emerging technique that is presented in this work as a method to further advance inspection techniques used in CANDU® nuclear reactors. A PEC probe was designed to inspect the unique shape of the trefoil shaped hole to detect and characterize wall loss and the presence of magnetite in A516 carbon steel BSPs with trefoil shaped holes from within 15.9 mm (5/8”) Alloy-800 SG tubes. PEC was also used to observe how measurements of wall loss were affected by the presence of magnetite. This work presents Finite Element Method (FEM) simulations and experimental results collected to observe these degradation modes. The probe was demonstrated to be capable of detecting far side wall loss as low as 20%, locating and characterizing the relative permeability of magnetite, and of detecting wall loss when magnetite was present. FEM simulations and experimental results were found to be in good agreement, suggesting that additional investigations of the effects of BSP degradation on PEC signal response may also be performed using FEM models.