1 January 2017 in Procedia Engineering
1 December 2015 in Transactions of the VŠB – Technical University of Ostrava, Civil Engineering Series
The subject of this paper is to describe some of the aspects manifesting in the use of the elastoplastic material model library multiPlas, which was developed to support non-linear computations in the ANSYS system. The text focuses on the analysis of numerical simulations of a virtual tension test in several case studies, thereby the text endeavours to describe the problems connected with modelling non-linear behaviour of concrete in a tensile area.
1 October 2017 in IOP Conference Series: Materials Science and Engineering
Identification of a concrete material model parameters using optimization is based on a calculation of a difference between experimentally measured and numerically obtained data. Measure of the difference can be formulated via root mean squared error that is often used for determination of accuracy of a mathematical model in the field of meteorology or demography. The quality of the identified parameters is, however, determined not only by right choice of an objective function but also by the source experimental data. One of the possible way is to use load-displacement curves from three-point bending tests that were performed on concrete specimens. This option shows the significance of modulus of elasticity, tensile strength and specific fracture energy. Another possible option is to use experimental data from compact tension test. It is clear that the response in the second type of test is also dependent on the above mentioned material parameters. The question is whether the parameters identified within three-point bending test and within compact tension test will reach the same values. The presented article brings the numerical study of inverse identification of material model parameters from experimental data measured during compact tension tests. The article also presents utilization of the modified sensitivity analysis that calculates the sensitivity of the material model parameters for different parts of loading curve. The main goal of the article is to describe the process of inverse identification of parameters for plasticity-based material model of concrete and prepare data for future comparison with identified values of the material model parameters from different type of fracture tests.
1 January 2017 in AIP Conference Proceedings
Today, the inverse identification or optimization of material model parameters is very often used to find input parameter values for use in relevant nonlinear material models. These parameter values should enable the responses of structures obtained from numerical simulations to very closely approximate the real responses of such structures obtained from experiments. Due to the popularity of concrete as a construction material, much attention is paid to nonlinear material models that aim to describe its behavior. This paper is focused on the optimization-based inverse identification of the parameters of two related nonlinear concrete material models which are known as the Karagozian & Case Concrete model and the Karagozian & Case Concrete model - Release III. Within this paper, the identification of the material model parameters is performed on the basis of interaction between nonlinear numerical simulations, optimization algorithms and experimental data, the latter of which take the form of a loading curve measured during a triaxial compression test. A comparison of the responses of both of the used material models when the optimized parameters are employed is, of course, part of this paper.
10 July 2018 in AIP Conference Proceedings
The contribution deals with the description and demonstration of an algorithm for the generation of spatial geometry – the material structure of concrete. The term “material structure of concrete” is taken to mean a mix of randomly distributed aggregate and cement binder, the latter being used to bond the aggregate. In order for the generated material structure to be as realistic as possible, a photograph of the material which is to be created is used as the input (i.e. data source) for the algorithm. The generation process itself involves the use of spatial noise functions, which are able to generate simple patterns as well as very complex fractals. The algorithm is based on the principle that similarities can be found between a suitably placed section cut through a noise space and the abovementioned input for the algorithm – a real photograph of the material. The functionality and individual steps of the algorithm are presented via two examples of concrete generation with differently shaped aggregate grains.