Development and implementation of a methodology for hybrid fire testing applied to concrete structures with elastic boundary conditions

Fire tests remain a precious tool to comprehend the behavior of structures under accidental fire conditions. The common practice in fire testing is to isolate the tested element in a furnace in which the mechanical support conditions are maintained constant throughout the test. However, such tests fail to capture the effect of the structure surrounding the element of interest when this effect cannot be realistically modeled by a free or fixed support condition. It has been observed in large-scale tests that the behavior of entire structures under fire is different compared with the behavior observed in traditional tests on isolated elements. This indicates the importance of capturing accurately the boundary conditions between the element and the remainder of the structure when characterizing the behavior of this element in fire.
The literature describes a few attempts at performing fire tests under realistic boundary conditions. In the latter, the tests were still performed on isolated elements but the boundary conditions were updated during the test taking into account the characteristics of the remainder structure. This technique, called hybrid testing, represents an appealing solution to test structural elements under realistic boundary conditions.
Hybrid testing is a methodology which offers the advantage of testing singular structural elements (or a group of structural elements) named physical substructure PS while at the same time considering the characteristics of the remainder substructure named numerical NS, thus allowing to model realistic boundary conditions. Pioneering work has been done in the seismic field where this technique is now well described, but the implementation of this methodology for structural fire testing raises important challenges due to the specificities of the field. A few hybrid fire tests have been performed in the past on columns and slabs. Their analysis shows that they all use a similar methodology, which is referred to as the first generation method in this work.
The objective of the thesis was to develop and implement a hybrid fire testing methodology on a reinforced concrete beam extracted from a moment resisting frame. Initially, it was intended to build on the first generation method, but after its detailed analysis in the development stage it has been observed that the process can be unstable. The value of the stiffness ratio between the numerical substructure and the physical substructures has been identified as critical in governing the stability of the test, dictating whether the hybrid test needs to be applied in displacement control or force control. This is a severe drawback of the first generation method, as the stiffness ratio is unknown and changing during the test; besides different degrees-of-freedom can require different procedures during the test. Therefore, it has been shown that the first generation method should not be applied as it can lead to instability prematurely during the tests.
To overcome the drawbacks of the first generation method, the objective was to develop a new technique that leads to interface equilibrium and compatibility while at the same time is unconditionally stable (i.e. independently of the stiffness ratio). Thus a novel methodology was developed and applied to the case of a concrete beam (PS) being part of a concrete moment resisting frame (NS). The novel method makes use of the PS’s stiffness in addition to the NS’s stiffness as it was the case in the first generation method. The stiffness matrix of the PS is unknown during the test therefore the initial tangent stiffness matrix is considered during the calculations. The latter choice influences the value of the time step to be adopted during the test. Every time step the boundary conditions are updated and it will be discussed how the chosen value can influence the results.
A predetermined matrix is used to describe the behavior of the NS during the hybrid fire tests. This approach does not capture the nonlinearity of the remainder but at the same time the implementation is relative simple and the negative effect of the time calculation is eliminated. The procedure to compute the predetermined matrix of the NS is presented in this thesis. One possible direction in the future development of hybrid fire testing is to model the NS in the finite element model.
The algorithm of the proposed method is developed and implemented in nonlinear finite element software SAFIR in order to perform virtual hybrid fire tests. The same algorithm is translated in order to be implemented by the company in charge of the control system at the CERIB furnace facility.
The thesis also presents a traditional fire test that has been performed on the beam, in order to highlight the differences when testing structural element without and with the real boundary conditions. For the hybrid test, three degrees-of-freedom are controlled at the interface. The furnace facility has an important role to perform successful test where the equilibrium and compatibility are ensured and no instability occurs during the test. The impediments encountered during the tests will be discussed along with the recommendation for a successful hybrid fire test.