This paper describes the methods of performance evaluation and remaining life prediction for an aged reinforced concrete (RC) T-girder bridge via visual inspection data. The Bridge Management System (J-BMS) that was previously developed by the authors, and which is capable of forecasting the deterioration process of existing bridge members, was applied to evaluate the safety indices (soundness score) and re-maining life of the target bridge based on these test results. Using these methods, the remaining life of an aged RC-T girder bridge(SK-bridge) can be quantitatively estimated by applying the bridge rating expert (BREX) system, which is a subsystem of the J-BMS RC version that incorporates with the field inspection data. In this study, visual inspection was carried out on the aged bridge(SKbridge) by professional visual inspectors, during which all variations of the inspection results were evaluated using a five-step question-naire. As a result, it was found that the soundness score (safety index) and remaining life predictions were influenced by the learning (supervised) data selection. Additionally, the predicted remaining lives were verified through concrete core tests extracted from main girders and deck slabs, and cross-section cutting-off girder tests.

In this study, a novel VR-based system for inspection of deteriorating concrete bridges is proposed to support of the crack and spalling detection which is a key role in existing bridge maintenance. The proposed system, called a VR-based support system for bridge inspection, was applied to an aged bridge for confirming the representing ability of both cracking and spalling damages as a specific example. Experimental results showed that the VR images data with hammering sounds (audio) data have a sufficient ability of representing both crack width over 0.2 mm and a certain spalling area near concrete surface. In this system, the three-dimensional image(photo) data was used to recognize the cracking damage for professional visual inspector. On the other hand, as spalling damage inside the concrete surface couldn’t be identified accurately from only image(photo) data, then the hammering sounds (audio) data was incorporated into the system as an additional function for recoginittion of the spalling damage.