Cost Comparison of Traditional Glass Fiber Test Methods and Computer Aided Neural Prediction Supported Systems

Abstract

Within the complexity of the industrial production strategies, computer aided technologies have been becoming a survival key for company administrators for reducing expenses. Furthermore, new production methods and adaptation of dynamic market requirements force owners to apply computer aided solutions to reduce to production time of goods to the market. Nowadays, prefabricated concrete producers are facing the same problem and trying to apply new solutions to overcome these high costs. In this research, artificial neural networks and traditional glass fiber testing methods were compared to reduce the quality control and assurance processes of prefabricated glass fiber reinforced concrete (GRC) production. 143 different four-point flexural test results of glass fiber reinforced concrete mixes with the varied parameters as temperature, fiber content and slump values were introduced the artificial neural networks models. The proportional limit properties (LOP) of glass fiber reinforced concrete and trained neural network analysis are taken into consideration for comparison. The outcomes of the analysis reflected that there is a strong correlation between the proportional limit of glass fiber reinforced concrete on-site test and the artificial swarm-based optimization algorithm results. Depending on this secure data, on-site test quantities are reduced and checked for cost deduction of traditional test results.