IDENTIFICATION OF CORROSION DUE TO CHLORIDE PENETRATION ON REINFORCED CONCRETE ELEMENTS USING INFRARED THERMAL IMAGE PROCESSING

• Main Author: Suyadi
• Format: Dissertations
• Language: Indonesia
• Subjects: Teknik sipil
• Online Access: https://digilib.itb.ac.id/gdl/view/81583
• Institution: Institut Teknologi Bandung

This research aimed to identification the corrosion level of concrete reinforcing steel by the Principle Component Thermography (PCT) analysis of image data from active thermography testing. For predicting the corrosion rate of reinforcing steel by thermographic method, 20 MPa concrete was heated at a distance (dobJ) of 30 cm - 50 cm using a 500watt halogen lamp for 30 minutes. Thermogram data acquisition using a speed of 8 frames per second (fps) with a variation of the camera distance to the test object (dcanJ is 60 cm, 70 cm and 80 cm. This research does not consider the rate of image change over time. The correction of thermogram errors is done during image preprocessing by entering the reflected temperature (Tref/) value in the object temperature (Tab) equation. The difficulty of identification in the thermal contrast method and the subjectivity factor in the signal noise ratio (SNR) or region of interest (ROI) method analysis as well as the advantages of the PCT method, are the basis for the proposed update in this research. In this research, image processing uses the PCT method with corrected thermogram sequence data input. The PCT is an Empirical Orthogonal Function (EOF) method based on the Eigen Value Problem (EVP). The different values of thermal properties between the two materials result in spatial thermal distribution anomalies in the concrete. The EOF is a spatial pattern of the principal component that shows the distribution of concrete surface temperature variations. The eigenvectors of each EOF mode are the temporal variability of the EOF which shows the spatial variability of the data in the space dimension. In the variation of spatial eigenvectors correlating with the second largest eigenvalue of the PCT analysis results, there is a thermal pattern anomaly showing differences in the level of effusivity in the area. These spatial variations are composed into a data set known as EOF2. The sharp contrast of the anomaly depends on the degree of corrosion of the reinforcing steel in the concrete. EOF2 is sensitive to these deviations because it emphasizes areas where the thermal response deviates from the pattern defined by EOFI. Use ofSNR analysis in EOF2 mode to determine the eigenvector threshold (EOF2th) contour value. The Corrosion Index (IK) value, which is a comparison between the area of the EOF2th contour with the maximum spatial variation of the EOF2 mode (EOF2max) as a result of PCT analysis, can be used to detect this thermal anomaly. The results of the research show that there is a relationship between the Corrosion Index value and the level of corrosion of reinforcing steel (Thcor). Increased reinforcement corrosion tends to increase the Corrosion Index value. The experimental test results show a fairly good prediction of the reinforcement corrosion rate when the dcam is 60cm.