Case Study on Inspection of Radiant Heater Ethylene Cracker Tubes

A tube crawler has been developed to quantify the carburization in 25%Cr-35%Ni alloy tubes used in furnaces for ethylene production. Eddy current probes are used to detect the increase in magnetic permeability associated with chromium depletion in the alloy matrix. Diameter measurements detect the swelling associated with carburization, which means the crawler can also be used on higher alloy tubes. The eddy current probes can also detect cracking along the brittle chromium carbide phases. The design, physical set-up and operation with National Instrument’s Labview of the crawler will be described, along with the data analysis procedure in Mathworks Matlab and Quest Integrity’s LifeQuest reformer software, typical results and reports.

Remaining life assessments are made using Abaqus Finite Element Analysis (FEA) of the coil and hanger geometry, and changes in temperature during startup, typical production runs, and decoke cycle. The FEA incorporates plastic and creep material properties. Mechanical testing of ex-service HP 50 alloy has been undertaken to improve a material property data-base for carburised material. Work is underway on mechanical testing of ex-service 35Cr-45Ni alloy. The mechanical test data correlates the crawler magnetic permeability measurements with tube hardness, ultimate tensile strength, 0.2% yield stress, and elongation to failure under tensile testing and creep testing regimes. Optical and electron microscopy has been used to characterise the ex-service material, with Energy Dispersive Analysis of X-rays (EDAX), image analysis and Electron Back Scattered Diffraction used to quantify the carbide phases through the tube wall, which can be correlated with the magnetic response and tube swelling. Finite Element modelling of the carbon diffusion and carbide precipitation using phase equilibria predicted by Thermocalc for HP Nb alloy, is being used to incorporate damage accumulation rates caused by the carburisation into the global FEA models.