Lessons Learned from Blind Simulations of Three Reinforced Concrete Structural Components

Abstract

There is a high degree of uncertainty and complexity in the behaviour of reinforced concrete (RC), making it difficult for most analysis methods to accurately characterize the response of RC structures. Blind prediction contests provide a unique opportunity to truly evaluate the accuracy of analysis methods without having any prior knowledge of test results. Recently, the Concrete Structures research group at Carleton University participated in three blind response prediction contests. The first contest involved a quasi-static cyclic test of an RC gravity column. The second one was about monotonic testing of a deep RC beam intended to represent a slice through a thick foundation mat. The last contest involved predicting the torsional response of an RC U-shaped wall. To predict the response of these specimens a set of nonlinear finite element analysis software and a newly developed simple macro-modelling approach both formulated based on the Modified Compression Field Theory were used. Prior to modelling the blind tests, a number of similar specimens from the literature were analyzed to verify the accuracy of the modelling methods. This study discusses the lessons learned by the research team from participating in these competitions. The importance of having modelling experience prior to using finite element analysis tools is highlighted by discussing the consequences of making poor modelling assumptions. The reliability of the proposed macro-modelling analysis method in predicting the response of shear-critical RC structures is also demonstrated. The findings will help engineers to avoid some of the pitfalls that they may encounter in blind simulations or modelling real-world structures.