• Dixit Soliya
• Hardik Solanki
• Anil Belani

In recent years, the increase in the number of terrorist attacks, blasts in oil rigs, explosion accidents in petrochemical industries have shown that the impact of blast loads on building and its components are a serious matter which should be kept in mind during the design of a building. Preliminary design phase is most important in blast resistant design process. It includes whether the building component will be safe or not with initial guess values of parameters such as steel diameter, size of element, allowable response, etc. This phase requires structural designer’s intuition to predict the initial guess. Artificial Neural Network (ANN) has shown great potential in prediction of values even with nonlinear relationships between parameters. In this research, ANN is used for the preliminary design of a RCC front wall of a building. Parameters such as peak over pressure, the natural time period, stand-off distance of blast, dimensions of the Wall, Steel Diameter, boundary conditions of the wall, allowable response were considered as input. Ductility ratio, maximum deflection and rotations were considered as output parameters. Spreadsheet was used for simpler calculations to generate required data based on values calculated using formulas and methods provided by Codal provisions such as special report “Design of Blast Resistant Buildings in Petrochemical Facilities” published by ASCE, UFC-3-340-02, IS:4991-1968 and TM 5-1300. To reduce the tedious process of predicting different values of output parameters through statistical methods, ANN can be used to train and simulate the prediction of the data. MATLAB software was used for training and simulation of ANN. Based on trained values, output parameters can be generated within allowable error values as ANN can create relations between most of the parameters. Simulated outputs within the input range of trained ANN provides satisfactory results while values outside the input range, doesn’t provide satisfactory results.

Artificial Neural Network, RCC front wall, Preliminary Blast Design, Ductility Ratio, Non-linear Prediction

"Preliminary Design Of A Blast Resistant RCC Element Using Artificial Neural Network", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 4, page no.518-524, April-2019, Available :http://www.jetir.org/papers/JETIR1904A77.pdf

"Preliminary Design Of A Blast Resistant RCC Element Using Artificial Neural Network", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 4, page no. pp518-524, April-2019, Available at : http://www.jetir.org/papers/JETIR1904A77.pdf