Aerobic and anaerobic general corrosion analysis models for carbon steel over pack for deep geological repository were derived using Artificial Intelligence technology. The derived models were utilized for trace analysis of experimental corrosion data and statistical corrosion analysis of analogue data from Yamato 6th tumulus. This study proposed AI technology utilization for corrosion analysis models, and presented procedure of statistical corrosion analysis and procedure to trace analogue data using corrosion analysis models. Furthermore, issues that underlying in analysis models derived from AI technology was discussed and presented possible direction of AI technology utilization for this purpose.

A fracture mechanics evaluation of the initiation and propagation of stress corrosion cracking, SCC, in high-strength carbon steel plates, HT780, was performed using notched four-point bending test specimens at 2 V （vs. Pt） in liquid ammonia. A semi-elliptical SCC was obtained with a stress intensity factor, K, of 3.2 MPam^0.5. Further tests using lower initial K-values and longer test periods are needed to verify whether this is the critical K-value, K_ISCC, or not. Although relationships between SCC growth rate and K were obtained, their SCC growth characteristics were significantly accelerated. It is necessary to reduce SCC acceleration and to understand the effects of SCC acceleration conditions on SCC growth characteristics for SCC growth prediction in the actual liquid ammonia tank environment.