DBPapers
DOI: 10.5593/sgem2017/22/S09.026

DEFORMATION ANALYSIS OF THE CZERSKO POLSKIE ROLLER DAM STEEL DRUM USING CONVENTIONAL GEODETIC METHODS AND NEURAL NETWORKS

J. Sztubecki, M. Mrowczynska, M. Sztubecka
Wednesday 13 September 2017 by Libadmin2017

References: 17th International Multidisciplinary Scientific GeoConference SGEM 2017, www.sgem.org, SGEM2017 Conference Proceedings, ISBN 978-619-7408-02-7 / ISSN 1314-2704, 29 June - 5 July, 2017, Vol. 17, Issue 22, 211-218 pp, DOI: 10.5593/sgem2017/22/S09.026

ABSTRACT

Defining the stability of a structure depends on several factors. It is difficult to predict them and take into account in the design phase. Over time, many of them will cause changes in the location of its components. Measurements are required to be taken during the operation of a structure in order to find out about any changes in its geometry. This paper contains an elaboration of the results of measurements of the geometry of a steel drum which is a major component of the historical roller gate dam - Czersko Polskie - in Bydgoszcz. The availability of the drum surface for the measurements was limited to the visible part, downstream of the structure. Nevertheless, using the adopted calculation procedure it was possible to determine the position of the drum axis as well as the twisting and deformation of the drum surface. In order to estimate the extent of deformation of the steel drum, conventional methods of surveying were used alongside new methods based on artificial neural networks (Hecht-Nielsen neural network).

Keywords: deformation, geometric evaluation, steel structure, roller dam, neural networks