WIT Press


Characteristic Identification Of Oil Dampers For Railway Vehicles Using Neural Networks

Price

Free (open access)

Volume

103

Pages

9

Page Range

725 - 733

Published

2008

Size

3,348 kb

Paper DOI

10.2495/CR080701

Copyright

WIT Press

Author(s)

R. Koganei, K. Sasaki & N. Watanabe

Abstract

We developed a characteristic identification system for devices of railway vehicles which is an essential component in a system of ‘Virtual Running Test Environment’ based on Hardware In the Loop Simulation (HILS) technology, as developed by Railway Technical Research Institute (RTRI) to replace running tests with bench tests. This report describes the outline and the effect of the characteristic identification system for oil dampers using a Neural Network (NN), which estimates input-output relation of the target in 6 degree-of-freedom using multi-axis damper test equipment. Keywords: railway vehicle, oil damper, Neural Network, hardware in the loop simulation, characteristic identification. 1 Introduction Running tests are indispensable to develop the railway vehicle. In Japan, however, a great amount of cost and time are required for the tests. In addition, most of the running tests take place on service lines since there is no test truck available for exclusive use in Japan. Accordingly, a number of test and the test condition are restricted. It is likely to lead to shortening and the quality improvement of the development process if it comes to be able to reproduce a real railway vehicle motion in detail by some bench examinations. Then, we work the construction of ‘Virtual Running Test Environment’ based on Hardware In the Loop Simulation (HILS) technology (Fig. 1), which replaces running tests with bench tests. It is necessary to mount the greater part of the railway vehicle components on the simulation model as software, because it is impractical to mount all of the railway vehicle components as hardware in

Keywords

railway vehicle, oil damper, Neural Network, hardware in the loop simulation, characteristic identification.