Deep learning and machine learning-based prediction of capillary water absorption of hybrid fiber reinforced self-compacting concrete

Abstract

Deep auto-encoders and long short-term memory methodology (LSTM) basedon deep learning as well as support vector regression (SVR) and k-nearestneighbors (kNN) based on machine learning models for the capillary waterabsorption prediction of self-compacting concrete (SCC) with single andbinary, ternary, and quaternary fiber hybridization were developed. A macroand two types of micro steel fibers having different aspect ratios, and PVA fiberwere used. One hundred and sixty-eight specimens produced from 24 mixtureswere used in the prediction models. The input was the content of cement, flyash, silica fume, fine and coarse aggregate, water, superplasticizer (SP), macroand micro steel fibers, PVA, time that the specimen was immersed in water,and splitting tensile strength. Water absorption was used as output. As per theANOVA analysis of the experiment results, the most effective parameters weremacro steel fiber and time for tensile strength and water absorption, respec-tively. Finally, binary hybridization of 1% macro steel fiber and PVA improvedthe splitting tensile strength while the use of PVA as binary, ternary, and qua-ternary fiber hybridization increased the water absorption of SCC specimens.The auto-encoder, LSTM, SVR, and kNN models predicted the water absorp-tion of fiber reinforced SCC with 99.99%, 99.80%, 94.57%, and 95.50% accuracy,respectively. The performance of deep autoencoder in the estimation of waterabsorption of fiber-reinforced SCC was superior to the other predictionmodels.