Modeling the Compressive strength of Engineered Cementitious Composites (ECC) using Machine Learning and Artificial Neural Network (ANN)

Engineered cementitious composites commonly known as bendable concrete is highly known due to its ultimate strain capacity between 3% to 5% as compared with 0.01% of normal concrete [1]. ECC is similar to High performance fiber reinforced cementitious composites with a difference of nonuse of coarse aggregates. ECC is known for its good compressive, flexural and tensile strength. Low carbon foot print due to usage of fly ash makes it ecofriendly and cost effective. Micro crack formation and self-healing behavior of ECC makes it popular in the construction industry. Due to its ductile behavior, this material has a wide use in the area which is prone to natural disaster. Different types of fibers can be used in ECC like Polypropylene fiber (PP), polyvinyl alcohol (PVA), natural fibers etc. Green light weight ECC can be developed with the use of high volume of industrial waste [2]. This paper deals with modeling the compressive strength of ECC using machine learning and ANN.154 Data sets were collected from different experimental works performed to determine the compressive strength of ECC. Data sets were normalized and artificial neural Network and Machine learning techniques like Random Forest, Linear Regression, LASSO Regression, Elastic Net, Ridge Regression, Support vector R are used to analyze the datasets and validation is done using experimental work. The results show a good correlation with the predicted and experimental compressive strength values.