Application of machine learning and fuzzy logic to predict the effect of impact loading on the life of rectangular reinforced panels as offshore reinforced concrete structures

Tafarroj, Mohammad Mahdi; Hatami, Hossein

doi:10.22034/ijcoe.2024.457912.1081

Application of machine learning and fuzzy logic to predict the effect of impact loading on the life of rectangular reinforced panels as offshore reinforced concrete structures

Articles in Press

Document Type : Original Research Article

Authors

Mohammad Mahdi Tafarroj ¹

Hossein Hatami ²

¹ Lorestan University, Khorramabad

² Lorestan University

10.22034/ijcoe.2024.457912.1081

Abstract

Depending on some factors like material, design, and magnitude of the impact, impact loading can have significant effects on various structures. For example, the effects of impact loading on marine structures are significant. It can be caused to structural damage, fatigue, corrosion, material degradation, and many other problems. Wave action, vessel collisions, and underwater explosions are the factors that can all contribute to impact loading. Cyclic impact loading resulting from the waves on offshore structures reduces the life of structures in the elastic area. In this study, as a comprehensive experimental investigation considering four mix designs, 64 rectangular composite panels were made with 100 mm2 area and 30-, 45-, 60-, and 75-mm thickness and tested by impact loading. Tensile, compressive, and flexural loading were done on all specimens. Steel fibers with 0, 0.25, 0.5, and 0.75 percent and 25 m of length were utilized to make the concrete composites. A hammer with 180 kg weight and 7500 J power was dropped on specimens for impact loading with a drop hammer test machine (DH-TM). Specimens were dynamically loaded by drop test from a 60 cm height. The composition of steel fibers and expanded metal sheets with each other significantly increases energy absorption. Moreover, the initial peak force increases, while crushing length and specimen’s deformation reduces. Models based on two well-known techniques artificial neural networks (ANN) and adaptive network-based fuzzy inference systems (ANFIS) have been developed. The performances of both techniques were compared and finally, the most appropriate one for predicting new data is introduced.

Keywords

Machine Learning

Self-compacting Cementations Composite

Impact Loading

Energy Absorption

offshore structures

Subjects