Publication
Applying metamodeling techniques in the design and optimization of train wheel detector
Publisher:
Emerald
Date:
07-09-2012
DOI:
10.1108/02602281211257560
Abstract: In writing this paper, the authors investigated the use of electromagnetic sensors in axle counter applications by means of train wheel detection. The purpose of this paper is to improve the detection capability of train wheel detectors, by installing them in the optimal orientation and position, using finite element modeling (FEM) in combination with metamodeling techniques. The authors compare three common metamodeling techniques for the special case of wheel detector orientation: response surface methodology multivariate adaptive regression splines and kriging. After analyzing the effective parameters of a train wheel detector, an appropriate method for decreasing the system susceptibility to electromagnetic noises is presented. The results were validated using a laboratory‐based system and also the results of field tests carried out on the Iranian railway network. The results of the study suggest that the FEM method and a metamodeling technique can reduce the computational efforts and processing time. In this paper, combination of FEM and metamodeling approaches are used to optimize the railway axle counter coils orientation, which is more insusceptible to electromagnetic noise than initial arrangement used by some signallers.