Concrete Durability Properties Using Multi-Head Attention Mechanism-Recurrent Neural Network Based on Sugarcane Bagasse Ash

The main waste that the agriculture sectors create is sugarcane bagasse ash (SCBA). These days, there is a huge demand for fine aggregate in construction industry, and it is extremely challenging to dispose of industrial and agricultural waste. A combination of its ability to contain both silica and aluminum ions, SCBA can efficiently replace some of the fine aggregate in concrete. It is employed to partially replace fine aggregate in M50 grade concrete after being calcined at 600 C for two days to improve its silica content. Five concrete mixes, one of which is a control mix, are cast with replacement percentages that range from 0% to 20% in steps of 5%, using SCBA in place of fine aggregate. For 28 days, the concrete specimens are allowed to cure. According to the experimental study, 15% substitution of fine aggregate with SCBA results than other replacement levels. Therefore, it is determined that 15% is the ideal amount to replace the fine aggregate with SCBA. Durability tests, including those for water absorption, porosity, sorptivity, bulk density, acid resistance, and corrosion resistance are performed to determine the best percentage. It appears that replacing 15% of fine aggregate with SCBA performs better. These data are determined using Multi-Head Attention Mechanism-Deep Neural Network (MHAM-DNN) that attains less Root Mean Square Error (RMSE) of 0.5648 compared to existing methods like Self Attention-DNN (SA-DNN).