Influence of real material properties on the response of a numerical model: case study concrete pole

Abstract

The modeling process during the design of structures is a fundamental task of the engineering practice. The model predictions are dependent on the underlying assumptions and simplifications. The influence of the abstraction level on the prediction quality of the numerical model should be known. This study focuses on the question how much the approximation of the model response can be improved if the theoretical input parameters prescribed in codes and norms are replaced by parameter with a high accuracy. In this study, a 10 m long spun-cast pre-stressed concrete pole is investigated intensively. Due to the production process, irregularities concerning the cross section and the material properties inside the concrete pole and along its length can occur. With several laboratory experiments, the local geometry and material properties of the pole are determined. The randomness of the properties is taken into account in the model and its influence on the model response is investigated. The model predictions are compared to experimentally determined values describing the dynamic behavior of the pole. This approach enables studies on the sensitivities of the input parameters on the model response. Furthermore, the required accuracy of the model input parameters for sufficient model predictions is discussed.