Impact-resistance of Reinforced Concrete Structures

Abstract

Reinforced concrete (RC) protective barriers such as rockfall barriers and vehicular barriers need to be designed to resist impact actions. However, there are uncertainties over the required stability of the barrier to withstand the impact forces predicted for the projected impact scenarios. A critical review of the literature covers a range of methods for estimating the impact action imposed by a hard impactor on RC members. Those methods providing estimates of the peak impact force occurring in a transient manner at the point of contact between the boulder and the surface of the RC member, or cushion material placed in front of the member, are classified as force-based (FB) methods, whereas methods providing estimates of the displacement of the target, or a quasi-static force corresponding to the estimated displacement demand, are displacement-based (DB) methods. The FB methods have gained useful insights into protection of the barrier from localised damage. However, the destabilising effects and other global effects (of bending and shear) of the impact action are preferably estimated by DB methods. Fundamental distinctions between the two classes of methods, and different types of forces generated by an impact have been explained in the literature review. Overly conservative estimates of the destabilising action of an impact can be resulted if the peak impact force is applied in a static manner to the model of the barrier (the target). The DB model that has been developed at the University of Melbourne is less conservative than existing codified models and has been verified by comparison with results from laboratory experimentation. The model has also been shown to give estimates of the bending moment of a simply supported beam that are consistent with recommendations by the CEB (Euro-International Concrete Committee) model. The primary objective of this project is to develop simple analytical models for assessing the global response behaviours of a RC rigid barrier when subjected to hard impact, including overturning, sliding and bending. These models were developed based on the same underlying DB methodology, but take different forms depending on the type of response.

In a conventional FB design procedure wherein the impact action is represented by an equivalent static force, the demand on the stability of the barrier increases with its height because of the higher overturning moment that has to be resisted. There are significant costs implications associated with this design methodology and more so in cases where a deep (piled) foundation is required to resist the overturning moment transmitted from a tall barrier. An alternative design approach based on equal energy and momentum principles as proposed in this thesis predicts a higher factor of safety against overturning with a taller, free-standing, barrier when the base dimensions are kept the same. The free-standing approach to design saves costs as the need of a deep foundation is eliminated and stresses within the barrier is always lower when the base is free to rotate. This alternative design approach has been verified by results from systematic physical and simulated impact experimentation.

When space is limited, sliding action of the barriers become a key design consideration. By employing the DB methodology, an analytical model in the form of a closed-form expression has been developed for estimating the amount of sliding displacement of a barrier when struck by an impactor at a lower height. Similar to the overturning action, ratio of barrier mass to impactor mass has been found to have significant effects on the overall stability of the barrier.

The proposed DB model for flexural design of RC wall is based on designing the stem wall with sufficient longitudinal reinforcement (resulting in sufficient stiffness of the wall) in order that neither the steel nor the concrete would surpass the limit state of yield thereby ensuring linear elastic behaviour. This conservative design criterion serves to prevent the stem wall from accumulating flexural deformation following multiple impacts (e.g. by fallen boulders). It has been confirmed by numerical simulations using program LS-DYNA that a stem wall designed based on the proposed model was indeed responding within the limit of yield which is consistent with the design criterion. The LS-DYNA simulation of the example wall did not show any formation of cracks of a size which was visible (i.e. 0.1 mm) nor any permanent deformation to the wall other than in the vicinity of the point of contact at the top which is a localised damage phenomenon. Furthermore, the accuracy of the proposed method in predicting deflection of the stem wall forming part of the flexural stiffness method has also been validated by comparison with results recorded from physical impact experimentation as reported in the literature. The database employed in the validation comprises results from physical and numerically simulated testings covering a total of 18 impact scenarios. In summary, there is sufficient evidence in support of the use of the proposed DB model in practice for designing the stem wall to perform satisfactorily in bending.

Given the robustness of the proposed design methodology, it was employed to design a RC wall specimen for a large-scale impact experiment. The experiment was fully instrumented to measure the bending response behaviour of the wall specimen and the corresponding material strains. As expected, the specimen did not exceed its yield limit within the scope of the experiment. The estimated results have been shown to be in good agreement with the experimental results. Importantly, it was found from both the experimental and predicted results that the inertial resistance developed in the target stem wall played a significant role in terms of the bending response behaviour of the wall. Such an effect is normally neglected in a conventional FB design procedure.

The ultimate goal of this research project is that the analytical models presented will be useful for designers of impact-resistant structures aiming for undertaking a more rational and optimised design. The practical applications of the proposed models in designing a rigid rockfall barrier are illustrated at the end of this thesis. Quasi-static lateral load from debris flow has been incorporated to co-exist with the impact action of the boulder in the calculation procedure. Design checks to ensure stability from overturning and satisfactory performance of the barrier in sliding and bending of the stem wall are also presented at every stage of debris surge. The design example addresses all the requirements in engineering practices to well illustrate the application of the new design methodology.