Simulación mediante modelos predictivos del comportamiento a tensión del UHPFRC

Ortega Guzmán, Juan José del Carmen | 2020

Ultra High Performance Fiber Reinforced Concretes (UHPFRC) have been the subject of much research due to their superior characteristics compared to conventional concretes. Its applications include the rehabilitation and strengthening of structures. Due to its mechanical characteristics, it is ideal for structures exposed to loads that require a high degree of ductility, such as earthquakes and explosions. In addition, it is used for the accelerated bridge construction system, an application in which the ability of the UHPFRC to resist tensile stress is decisive. Taking into account the above, it is necessary to measure properties such as tensile behavior to be able to analyze the effects of materials such as fibers and buy their performance. However, the complexity of the trials from a technical and financial point of view, as well as the restrictions in time and space, condition these experimental campaigns. In the present investigation, through the use of predictive models of neural networks and multivariable regression, the maximum stress resistance of UHPFRCs, its associated deformation, was predicted and the performance of these models was evaluated. For the training of the models, a total of 934 data were used.