ARTIFICIAL NEURAL NETWORK MODEL FOR FORECASTING CONCRETE COMPRESSIVE STRENGTH AND SLUMP IN EGYPT

Slump and compressive strength of concrete are commonly used criteria in evaluating fresh and hardened concrete. accordingly, prediction of such criteria is important for the quality assurance of the produced concrete. in this paper, a neural network (nn) model is developed to predict concrete compressive strength and slump in egypt. the artificial neural network (ann) model is developed, trained and tested using 1000 different concrete mixes gathered from different batch plants distributed all over the arab republic of egypt. important parameters that have noticeable effect on the compressive strength and slump are identified and used as the inputs for the anns model. the developed model can be used either to predict the compressive strength and slump for a given mix or to estimate the different ingredients to achieve a targeted compressive strength after seven and twenty eight days. to verify the results of the anns model, seventeen concrete samples are prepared and tested at laboratory and the same ingredients of the mixes are used to predict the strength and slump using the developed ann model. the root mean square error (rmse) of the results for the slump and the compressive strength after 7 and 28 days equal 3.74, 1.79 and 3.05, respectively. these results showed the ability of the developed anns as an effective tool to predict and estimate the compressive strength and slump of concrete in egypt. the analysis of the test results leads to the conclusion that this idea can be used for the development of valid systems for specifications and standards.