Application of Neural Networks for Proportioning of Concrete Mixes

Abstract

In determining Proportioning of concrete mixes, code information, specifications, and the experience of experts are needed. However, all factors regarding mix proportioning cannot be considered Therefore, the final acceptance depends on concrete quality control Zest results. In this process, the uncertainties of materials, temperature, site environmental situations, personal skillfulness, and errors in calculations and testing process come into view. Then the adjustments must be made for proper proportioning. This kind of concrete mix proportioning and adjustments are somewhat complicated: time-consuming and are uncertain tasks. In this paper, as a tool to minimize the uncertainties and errors of the proportioning of concrete mixes, an artificial neural network is used. Not only are the required compressive strengths used to train and test the network, but so are the actual compressive strengths with variations obtainable from the final compressive strength test. The results show that neural networks have strong potential as a tool for concrete mix proportioning.