Turbine blade tip clearance determination using microwave measurement and k-nearest neighbour classifier

Turbine blade tip clearance and deformation can be monitored by measurement of scattering parameters. In this research, a k-band microwave sensor has been simulated, optimized and implemented in computer simulation technology software which can detect tip clearance variations by measurement of the scattering parameters using a network analyzer. The open-ended circular waveguide with the tip of the turbine blades forms a cavity resonator known as a microwave sensor. Scattering parameters can be used as fingerprints of the blade that are altered due to changes in tip clearance and deformation in accordance with the theory of small perturbations in cavity resonators. In this research, a new method for interpretation of the measured scattering parameters is presented that is based on measuring indices and the k-nearest neighbor classifier. It has been shown that the k-NN classifier provides acceptable accuracy for detection and determination of the blade tip clearance and deformation. (C) 2019 Elsevier Ltd. All rights reserved.