Concrete slump prediction modeling with a fine-tuned convolutional neural network: hybridizing sea lion and dragonfly algorithms

Abstract

High-strength concrete (HSC) is defined as concrete that meets a special combination of uniformity and performance requirements, which cannot be attained routinely via traditional constituents and normal mixing, placing, and curing procedures. It is a complex material since modeling its behavior is a difficult task. This paper intends to show the feasible applicability of optimized convolutional neural networks (CNN) for predicting the slump in HSC. The following are the parameters that given as the input for the prediction of slump: cement (kg/m3), slag (kg/m3), fly ash (kg/m3), water (kg/m3), super-plasticizer (kg/m3), coarse aggregate (kg/m3), and fine aggregate (kg/m3). In order to make the prediction more accurate, the design of CNN is assisted with optimization logic by making some fine-tuned filter size of the convolutional layer. For this optimization purpose, this work presents a new “hybrid” algorithm that incorporates the concept of sea lion optimization algorithm (SLnO) and dragonfly algorithm (DA) and is named as Levy updated-sea lion optimization algorithm (LU-SLnO). Finally, the performance of the proposed work is compared and proved over the state-of-the-art models with respect to error measure and convergence analysis. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH, DE part of Springer Nature.