Admixed high-performance concrete property prediction by novel regression-based models

Abstract

High-performance concrete (HPC) is one of the most important elements in constructing bridges, skyscrapers, and dams. This concrete additive plays a very important role in performance and response to inflow loads such as earthquakes and dead loads. Fly ash (Fa) and Micro-silica (Ms) are additives added to concrete by cement to reduce water to cement. Increase the ratio and increase the hardening of the cement. This will improve the compressive strength (Cs) of the concrete. Modeling is required for this type of structure. The radial basis function (RBF) is one of the models that can produce better and more rational results. This model combines two optimizers, the Sine Cosine Algorithm (SCA) and the Artificial hummingbird algorithm (AHA), in the framework of RBF-SCA and RBF-AHA, which are considered to be new and effective initiatives in the field of algorithms. The lowest amount of error parameters contains: (RMSE = 2.58), (NMSE = 6.59), and (U95 = 7.16) for RBF-AHA in the train section and the test section (MBE = – 0.1929). The (Tstate = 0.285) in the train section of the RBF-SCA has the lowest compared to another section. RBF-AHA has the highest R2 value of 97.15% in the training area. Both hybrid models can have the desired error and the correct percentage based on the given output. However, the RBF-AHA model may look more powerful in this modeling.