Prediction of Concrete Compressive Strength from Early Age Test Result Using an Advanced Metaheuristic-Based Machine Learning Technique

Doddy Prayogo, Min-Yuan Cheng, Janice Widjaja, Hansel Ongkowijoyo and Handy Prayogo
Pages 856-863 (2017 Proceedings of the 34rd ISARC, Taipei, Taiwan)

Determining accurate concrete strength is a major civil engineering problem. Test results of 28-day concrete cylinder represent the characteristic strength of the concrete that has been prepared and cast to form the concrete work. Waiting 28 days is quite time consuming, but it is important to ensure the quality control process. Machine learning techniques are increasingly used to simulate the behavior of concrete materials and have become an important research area. This study proposed a comprehensive study using an advanced machine learning technique to predict the compressive strength of concrete from early age test results. Early age test data are being used in this case to get reliable values of the two constants which are required for the prediction. A total of 28 historical cases were used to establish the intelligence prediction model. Obtained results show the performance of the advanced hybrid machine learning technique in predicting the concrete strength with a relatively high accuracy measured by four error indicators. Therefore, the proposed study can offer a high benefit for construction project managers in decision-making processes based on early strength test results.

Keywords: Concrete, Compressive Strength, Early Strength, Machine Learning, Prediction