Theoretical model-based quantitative optimisation of numerical modelling for eddy current NDT

Abstract Eddy current (EC) nondestructive testing (NDT) is one of the most widely used NDT methods. Numerical modelling of NDT methods has been used as an important investigative approach alongside experimental and theoretical studies. This paper investigates the set-up of numerical modelling using finite-element method in terms of the optimal selection of element mesh size in different regions within the model based on theoretical analysis of EC NDT. The modelling set-up is refined and evaluated through numerical simulation, balancing both computation time and simulation accuracy. A case study in the optimisation of the modelling set-up of the EC NDT system with a cylindrical probe coil is carried out to verify the proposed optimisation approach. Here, the mesh size of the simulation model is set based on the geometries of the coil and the magnetic sensor, as well as on the skin depth in the sample; so the optimised modelling set-up can be useful even when the geometry of EC system, the excitation frequency or the pulsed width is changed in multi-frequency EC, sweep-frequency EC or system and pulsed EC. Furthermore, this optimisation approach can be used to improve the trade-off between accuracy and the computation time in other more complex EC NDT simulations. Keywords: eddy current nondestructive testing (EC NDT)numerical modellingquantitative optimisationmesh sizeregions of interest (ROI) Acknowledgements The authors would like to thank the NSFC (grant no. 51005039), EPSRC UK (grant no. EP/E010458/1) and the special plan for the enterprise correspondent of commissioner's action issued by Guangdong Province, Ministry of Education and department of Science and Technology of China for funding the research. The authors also thank Associate Professor Yong Li of Xi'an Jiaotong University for providing the ETREE model and for valuable discussions. Notes 1. Email: xinhua.li@ncl.ac.uk 2. Email: anthony.simm@ncl.ac.uk 3. Email: g.y.tian@ncl.ac.uk Additional informationNotes on contributorsXinhua Li 1 Anthony Simm 2 Guiyun Tian 3