GONZALEZ SALCEDO, Luis Octavio; GUERRERO ZUNIGA, Aydee Patricia; DELVASTO ARJONA, Silvio    ERNESTO WILL, Adrián Luis. Artificial Neural Model based on radial basis function networks used for prediction of compressive strength of fiber-reinforced concrete mixes. []. , 29, 2, pp.37-52.   20--2019. ISSN 0124-8170.  https://doi.org/10.18359/rcin.3737.

A complex nonlinear relationship exists between the factors influence the compressive design strength of steel fiber reinforced concrete. This relation between input variables, the factors, and the output variable as it is the compressive design strength can be obtained by using an artificial neural model, which has characteristics of self-adapting, self-study and nonlinear mapping. An application of a radial basis function artificial neural model is presented in this paper. Compressive design strengths of steel fiber reinforced concrete endured mixes with diverse proportioning was predicted and compared with the experimental measured results. The predicted values were analyzed by R lineal correlation factor. The results showed that the predicted values based on radial basis function networks presented coincidence with the experimental values, and the predictability of the mechanical property of the neural model is better than that of the multi-layer neural models developed previously by the authors. The training of the neural models allowed us to conclude that the use of materials relationships is a better indicator for the comparison between different dosages of concrete mixtures that lead to similar compression strengths. A future agenda is opened in the generation of new methods of studying in metal fiber reinforced concretes compression design strength reinforced in the field of engineering.

: Fiber-reinforced concrete; compressive design strength; properties prediction; artificial neural networks; radial basis function; artificial intelligence.

        · |     · |     · ( pdf )