Paper Title
Apply ANFIS to improve on site Concrete Compressive Strength Estimation using Ultrasonic Pulse Velocity Test

Abstract
Among many non-destructive test methods, the Ultrasonic Pulse Velocity Test is one of the most used tests for measuring compressive strength of concrete. However, according to the relevant literatures, it showed that the Ultrasonic Pulse Velocity Test has some ±15% errors in predicting the compressive strength of concrete. In addition, in recent years, artificial intelligence has gained a high degree of trust in the study of prediction results. It has developed rapidly, and different methods has been applied to different fields, such as Artificial Neural Networks (ANN), Support Vector Machine (SVM), Adaptive neuro-fuzzy Inference System (ANFIS).In this study, a total of 99 test data from the Ultrasonic Pulse Velocity Test in a large residential building complex was collected. The data are used as the training and testing group data of artificial intelligence model. Then prediction models such as the linear and nonlinear regression prediction models and Multiple Regression prediction model, and ANFIS Prediction model, are established. The results of each model are compared with the actual compressive strength of concrete, and the Mean Absolute Percentage Error (MAPE) is calculated. The error is expected to reduce effectively. The results show that the MAPE of the predictive models based on linear and nonlinear regression methods and Multiple Regression prediction model are 12.92%, 17.48% and 11.76%, respectively. The MAPE of ANFIS model is 7.51%. It can be concluded that using artificial intelligence to establish the concrete compressive strength Prediction model is able to improve the accuracy of the Ultrasonic Pulse Velocity Test for the test in construction sites. Keywords - Non-Destructive Test, Concrete Compressive Strength, Ultrasonic Pulse Velocity Test, Artificial Intelligence, Adaptive Neural Fuzzy Inference Systems, ANFIS