Prediction of compressive strength of concrete from volume ratio and Bingham parameters using adaptive neuro-fuzzy inference system (ANFIS) and data mining

摘要

Concrete is the most widely used and one of the oldest material in the construction industry. Compressive strength is one of the most important mechanical properties of hardened concrete because it is related to other properties or performance of concrete. In this study, a soft computing technique called adaptive neuro-fuzzy inference system (ANFIS) was carried out for predicting the compressive strength of concretes from their mix design and flow properties. For this purpose, values of concretes in 80 different mix designs were utilized in the ANFIS modeling. Although there is lowest coefficient of determination (R2) between Bingham parameters and compressive strength (R2= 0.262), the model results with volume ratio (R2= 0.787) is higher than Bingham parameters. Nevertheless, the best results were obtained from models using both variables (R2= 0.944). The results showed that the implemented models are good at predicting compressive strength. A comparison of results indicated that the ANFIS model is more feasible in predicting compressive strength than the data mining models previously developed by the author. In both models, volume ratio is more effective than the flow properties on the compressive strength of concrete but it is not sufficient alone in predicting compressive strength. These results suggested that ANFIS can be used as an alternative approach to predict compressive strength when it is used together, volume ratio with Bingham parameters as input parameters.