Civil Engineering Association

Study on Reinforced Concrete Frame with Solid Infill Masonry using Artificial Neural Network (ANN) RC Infill Wall Model in ANSYS (Ajay Gupta)

Despite being the most common construction practice throughout the ages, infills have not found the space it deserves, in the structural design. There is lack of proper and easy method to consider the effect of the in-filled. So, this research is a small effort in the search of the alternative approach for analyzing the infill frames. The FEM Models are normally incapable of considering all the effecting factors such as
non-linear behavior of the infill materials, lack of fit, non-homogeneity of the materials, etc. This research gives some idea to the structural engineer how to guess initially the parameters of interest during the design of infills. Structural design process is an iterative process and an approximate initial guess can reduce the time and cost involved in the analysis. The tentative design parameters can be predicted using the Artificial Intelligence and this computing power of the modern day computers has been used to fulfill the intended purpose. The data sets, which are generated by computer from the simulation of the infill-frame structure done in sophisticated software (ANSYS v10.0) capable of non-linear analysis, are used for the training of Neural Network. Few other unique data sets are taken for the validation of the Network trained. The comparison of the results from the ANN and that of software were in reasonable agreement with each other except in few rare cases.