Assessment of high-temperature damaged concrete using non-destructive tests and artificial neural network modelling. (June 2022)

• Record Type: Journal Article
• Authors: Almasaeid, Hatem H.; Suleiman, Akram; Alawneh, Rami
• Is Part Of: Case studies in construction materials. Volume 16(2022)
• Publication Date: 2022-06
• Subjects: CS Compressive Strength -- NDT Non-Destructive Test -- UPV Ultrasonic Pulse Velocity -- D-UPV Direct Ultrasonic Pulse Velocity -- S-UPV Semi-direct Ultrasonic Pulse Velocity -- I-UPV Indirect Ultrasonic Pulse Velocity -- SRH Schmidt Rebound Hammer -- RN Rebound Number -- ML Machine Learning -- ANN Artificial Neural Network -- DL Damage Level -- T Exposure Temperature -- ASE Average Squared Errors -- MARE Mean Absolute Relative Error -- R2 coefficient of determination (R2); Artificial neural network -- Fire -- High temperature -- Damage -- Concrete -- Non-destructive test -- Compressive strength; Building materials -- Case studies -- Periodicals; 691.05
• DOI: 10.1016/j.cscm.2022.e01080
• Languages: English
• ISSNs: 2214-5095

Abstract

Abstract: Evaluation of high-temperature damaged concrete is crucial to ensure the safety of any structure after a fire event. However, using destructive tests, such as taking cores from the concrete, can be costly and dangerous; specifically for damaged structures. Therefore, it is preferred to use in-situ non-destructive testing (NDT) in the assessment of damaged concrete. The objective of this study is to develop an artificial neural network model, based on destructive and non-destructive testing results, to assess the concrete strength after being subjected to high-temperature levels; without the need for further in-situ destructive testing. The effect of high-temperature levels (200–800 °C) on concrete compressive strength was investigated in this study using destructive compression and non-destructive tests on concrete cubes; including ultrasonic pulse velocity and Schmidt rebound hammer testing methods. The results of destructive and non-destructive tests of damaged and undamaged concrete were found to be highly correlated. Therefore, the data of this study and data obtained from the cited literature were augmented together and used to optimise and train the artificial neural network model. The artificial neural network analysis indicated that concrete compressive strength (CS), the magnitude of high-temperature damage, and the level of exposure temperature can be predicted with reasonable accuracy using only a combination of non-destructive tests results. The model had a coefficient of determination equals to 0.944.