Using logistic regression in tracing the significance of rubber-textile conveyor belt damage

The main purpose of this work is to investigate different levels of damage that occur on rubber textile conveyor belts. The conveyor belt damage depends on the heterogeneity of the transported material, transporting conditions and on attributes of the conveyor belt itself. The conveyor belts that are used in this study were divided into three categories: new, worn and renovated. In total, 67 test specimens were examined. The logistic regression method was selected in order to analyze individual levels of conveyor belt damage. Using this method, two differently shaped impact drop hammer tips (spherical and pyramidal) were applied and analyzed. Impact height of the drop hammer and the type of conveyor belt were treated as the independent variables. The output variable was selected as the one that best differentiated between significant and insignificant belt damage. The result of the work enables to estimate the odds of formation of the significant conveyor belt damage for the individual variables. A surprising result, which is obtained from the performed tests and model analyses, is a fact that application of a new, unused conveyor belt, leads actually to increased odds of the belt's significant damage. (C) 2014 Elsevier B.V. All rights reserved.