Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
This work provides a procedure and methodology for the design of an electro-magnetic sensor based on induction due to ferromagnetic and pure electrically conduc-tive hidden targets. The electromagnetic sensor is made of one or several excitation coils, which induce eddy currents in the target(s) and one or several sensing coils, which measure the target response. The latter signals are then processed to obtain the wanted properties of the target. Detection by induction at radio frequencies (medium frequen-cies) is famous for its relative simplicity, cost effective, and precise size discrimination but has disadvantages mainly in signal strength (especially for far hidden targets) and it is less effective in material discrimination in the case where multi-material targets are present. The presented methodology links three main design fields: excitation, sensor design, and post-processing. Its objective is to improve detection by induction in allow-ing for selectivity in detection or the possibility to distinctively detect targets made of different materials. Implemented simulation models together with experimental verifi-cation led to a highly improved procedure in target feature discrimination. Features are defined here as: size, depth, location, and even material of the target(s). A forward mod-el helped in linking the physical aspects of the defined targets to mathematical para-metrized equations that contributed easily in a flexible discrete inverse model for both transient (pulse induction) and time harmonic (harmonic induction). Feature discrimi-nation is a necessary step to reach the final aim defined here as distinctive detection. It means being able to distinguish between single material targets and mixed material targets at the same time and still discriminate their features. The proposal is flexible even for relative target orientation, robust in changing environmental conditions, relia-ble even for relatively far targets, balanced and robust to overcome manufacturing toler-ances, precise enough to meet industrial sensor specifications of such technology, and can combine different excitations with their benefits and give the required outcome se-lectively.
Ian Smith, Yves Sylvain Gilles Reuland, Slah Drira
Timothy Goodman, René Chavan, Anastasia Xydou, Matteo Vagnoni