Risk-informed approach for safety improvement of domestic research reactor

Abstract

Abstract This paper describes the effort to improve the safety level of a domestic research reactor and optimize its operation with regard to safety by conducting a probabilistic safety assessment (PSA) under full-power operating conditions. The PSA was undertaken to assess the level of safety for an operating research reactor in Korea, evaluate whether it is probabilistically safe to operate and reliable to use, and obtain insights about the procedure and design improvements for achieving increased safe operation. The scope of the PSA reported here is a Level 1 PSA, which addresses the risks associated with core damage. The technical objectives of this study were to identify the accident sequences causing core damage and to derive insights, such as design and procedure improvements, from the dominant accident sequences by conducting sensitivity analyses. The event tree (ET)/fault tree (FT) linking approach was used as the basic method for accident sequence quantification for the PSA. The accident sequence models followed the small ET/large FT approach. The total core damage frequency (CDF) was estimated as the sum of the frequencies of the individual ET sequences causing core damage. A truncation value of 1.0E−15/yr was applied for all the FTs or accident sequences. The PSA results indicated a point estimate of 4.55E−06/yr for the overall CDF, attributable to the internal initiating events for the Korea research reactor. Based on the dominant accident sequences from the PSA, eight types of sensitivity analyses were performed, and relevant insights for improved safe operation were derived. When the insights from these sensitivity analyses were applied to the reactor design and the operating procedure, the risk was reduced by approximately ten times compared to that before, and its safety was significantly improved. The PSA methodology is very effective in improving reactor safety in the full-power operating phase, and, in particular, it is a highly suitable approach to determine the deficiencies of a reactor at full power and improve the reactor safety by overcoming them.