ESTIMATION OF THE CONCRETE STRENGTH SETTING OFF FROM THE PHYSICAL PROPERTIES OF THE AGGREGATES BY ARTIFICIAL NEURAL NETWORK AND LINEAR REGRESSION METHOD AND COMPARISON OF THE RESULTS

Determination of the effect of the physical and mechanic properties that the aggregates have on concrete strength is only possible with the tests conducted. These studies take a long time and mostly are not economic. Therefore, different methods formed by utilizing the experimental studies done before are used to determine the strength characteristics. Results obtained from the model formed with the strength properties of concrete made by setting off from the physical characteristics that the largest component aggregate forming the concrete have and the results obtained with Linear Regression are compared in this study. Concretes complying with TS 706 standards have been prepared by keeping all components other than the aggregate forming the concrete constant and the 7 and 28 days compressive strengths of these concretes have been measured. The values determined experimentally have been estimated by developing models in Artificial Neural Networks and Linear Regression methods. 20 variables including crushed sand, Specific Weight of fine aggregate and coarse aggregate, Water Absorption, Dry Unit Weight, Bulk Density, Tight Unit Weight, Los Angeles Abrasion Strength, Fineness Modulus, and Flatness Modules have been used at the input stratum and 7 and 28 days compressive strengths have been used at the output stratum. It has been observed at the comparisons that the training and test results in the models can be estimated very close to test results.