Cubic Spline Method 최적화를 통한 노심보호연산기 성능향상 연구

Abstract

The optimization of the cubic spline method employed in the Core Protection Calculator (CPC) has been performed to predict an accurate axial power distribution. The conventional cubic spline method shows a technical limitation to the synthesis of the axial power distribution because it uses a restricted and simplified model. Especially, the decline of the synthesis performance occurs along the progress of the nuclear fuel burnup. In this study, the some factors influencing the synthesis performance were analyzed. From the factors analyzed, the optimal cubic spline method was constructed. It was also confirmed that the accuracy of synthesized axial power distribution was improved by the optimal method. Through the evaluation of 18,424 function sets and 50 basis functions, the cubic spline method using the optimal function sets and basis functions was suggested. In addition, the synthesis performance for predicting the axial power distributions, which are frequently generated during reactor operation period, was improved through the evaluation of the segmentation of the selection algorithm. The results reconstructed by the conventional cubic spline method exceed more than 8% Root Mean Square (RMS) error, which is permissible standard. On the other hand, the optimal method reduces the RMS errors by 5% below in the asymmetric axial power distributions such as bottom peak. It is verified that the optimal method keeps the RMS errors below 0.84% for one fuel cycle in the typical OPR-1000 reactor. It is thus expected that the CPC could obtain more accurate axial power distributions using the optimal cubic spline method proposed in this study. In addition, the unnecessary reactor trips would be avoided while improving the safety and economic of nuclear power plant by employing the optimal synthesis method. |노내 축방향출력분포의 정확한 예측을 위하여 노심보호연산기에서 사용하는 Cubic Spline Method의 최적화를 수행하였다. 기존 노심보호연산기에서는 다양한 형태를 가지는 축방향출력분포를 예측하기 위하여 비교적 단순한 모델을 사용함으로써 축방향출력분포를 정확하게 합성하지 못한다. 특히 노심 연소가 진행 될수록 합성능력이 저하되는 현상을 보인다. 본 연구에서는 축방향출력분포 합성에 미치는 요소를 분석하여 최적의 Cubic Spline Method를 구축/적용함으로써 기존 합성방법보다 축방향출력분포가 정확하게 예측되는 것을 확인하였다. 본 논문에서는 축방향출력분포 합성능력에 영향을 미치는 요소들의 분석을 통해 18,424개의 함수셋과 50개의 Basis Function을 출력분포 합성에 각각 적용하였다. 요소들의 분석 결과를 토대로 최적의 함수셋과 Basis Function을 적용한 Cubic Spline Method를 제시하였다. 또한 알고리즘 세분화를 통하여 많이 사용되는 축방향출력분포 구간에서의 합성능력을 더욱 개선시켰다. 기존 방법으로 합성 시 특정오차(8%)를 모두 초과했던 비대칭형 축방향출력분포 형태에서는 최적의 Cubic Spline Method를 적용한 결과, 합성에 인한 오차가 5% 이하로 평가 되었다. OPR-1000 원자로의 모든 연소도의 축방향출력분포 합성 시 오차가 0.84% 이하였다. 본 논문에서 제시한 최적의 Cubic Spline Method 적용한 노심보호연산기는 기존보다 정확한 축방향출력분포 예측이 가능 할 것이다. 또한 정확한 축방향출력분포 예측을 통하여 불필요한 원자로 정지를 최소화할 수 있으며, 원자로의 가동률을 향상시킴으로써 경제성 또한 확보 할 수 있을 것으로 기대된다.; The optimization of the cubic spline method employed in the Core Protection Calculator (CPC) has been performed to predict an accurate axial power distribution. The conventional cubic spline method shows a technical limitation to the synthesis of the axial power distribution because it uses a restricted and simplified model. Especially, the decline of the synthesis performance occurs along the progress of the nuclear fuel burnup. In this study, the some factors influencing the synthesis performance were analyzed. From the factors analyzed, the optimal cubic spline method was constructed. It was also confirmed that the accuracy of synthesized axial power distribution was improved by the optimal method. Through the evaluation of 18,424 function sets and 50 basis functions, the cubic spline method using the optimal function sets and basis functions was suggested. In addition, the synthesis performance for predicting the axial power distributions, which are frequently generated during reactor operation period, was improved through the evaluation of the segmentation of the selection algorithm. The results reconstructed by the conventional cubic spline method exceed more than 8% Root Mean Square (RMS) error, which is permissible standard. On the other hand, the optimal method reduces the RMS errors by 5% below in the asymmetric axial power distributions such as bottom peak. It is verified that the optimal method keeps the RMS errors below 0.84% for one fuel cycle in the typical OPR-1000 reactor. It is thus expected that the CPC could obtain more accurate axial power distributions using the optimal cubic spline method proposed in this study. In addition, the unnecessary reactor trips would be avoided while improving the safety and economic of nuclear power plant by employing the optimal synthesis method.