Deformation Forecasting Based on Multi Variable Grey Prediction Models

Abstract

The classic prediction methods take the system behavior as a stochastic process, using probability and statistics, searching the laws of massive historical data. However, since the statistical approaches are efficient with large volumes of data, they cannot work well in case of plenty information unavailable. The main purpose of the grey system theory is to predict uncertain systems behaviors' with limited number of data. It does differ from statistical analysis method as it does not deal directly with the original data and searches the intrinsic regularity of the data. In this study, deformation consisting on the crest of the Keban Dam in Turkey is aimed to determine by using multivariable grey prediction models GM(0,N) and GM(1,N). The outcomes show that GM(1,N) produces much more reliable results than GM(0,N) on prediction of deformation. The outcomes also confirm that there are very high level of relation between water level and deformation in a dam.