Neural Network Prediction of the Effect of Nanoparticle on Properties of Concrete

Abstract

Artificial intelligence techniques are used for optimization in engineering studies and it is observed that it gives better results than other classical methods. Artificial intelligence techniques, which are widely used in the solution of civil engineering systems, are examined with their technical aspects and basic principles. To improve the mechanical and physical properties of the concrete, nano and micro scale materials can be added and their effects on concrete properties are investigated intensively. Nanoparticles added to the cement-based materials, it is possible to accelerate the improvement of the early age characteristics of the cement. In this study, 9 different mixtures were designed with nanoparticles and ultra-fine materials such as cement, water, standard sand, super plasticizer, fumed silica, precipitated silica, F-class fly ash and silica fume. Mechanical and physical tests were performed on these samples. Mechanical properties such as pressure and bending strength and physical properties such as water absorption were found. The effect of mechanical and physical properties on artificial neural networks has been tried to be estimated. Concrete mixes; The accuracy of 7-day estimation data was found as R² = 1 and the error rate was RMSE = 0.00108, the accuracy of the 28-day forecast data was R² = 0.9888 and the error rate was RMSE = 0.0887. An experimental study with the results obtained using the neural network of the results obtained have been found to show very close values. So, it will be possible to obtain the desired mixture values with an effortless, economical and shorter time.