Optimization of fibre reinforced foam concrete for the mechanical behaviour by artificial neural network: Fid-Bau Portal

Optimization of fibre reinforced foam concrete for the mechanical behaviour by artificial neural network

Thiagu, H. / Madhavi, T. Ch.

Foam concrete has grown popularly in recent decades due to its unique advantages over plain concrete, including low energy consumption and manufacturing cost, favourable thermal response, ease of fabrication and demolition compared to other lightweight materials. It is also a lightweight concrete. It is a concrete mixture in which air bubbles are created into the concrete using a foaming agent. In this research, we provide the results of an experimental evaluation of the strength properties of FRFC (fibre reinforced foam concrete) including varying concentrations of steel fibres (0, 0.2, 0.4, 0.6, 0.8 and 1%) and polypropylene fibres (0.4, 0.5 and 0.6). Compressive stress, flexural strength, split tensile strength and impact strength were measured after casting and testing specimens for 28 days. Based on the tests conducted, the proportion of 0.8% of steel fibre and 0.5% of polypropylene fibre gives better results compared with foam concrete without fibres. Fibres play a main role with combinations of different percentage. The foam concrete with and without fibres optimization is tested experimentally and validated by soft computing techniques employing ANN (artificial neural network) modelling. The ANN results indicate R² values of 0.98617 in foam concrete with fibre and 0.96319 in foam concrete with fibre, indicating that the ANN model is an extremely accurate predictor of the strength parameters of foam concrete with and without fibre. Therefore, the most effective ML (machine learning) algorithm for determining the optimized fibre percentage through the various strength parameters of the foam concrete is the ANN model.