Uncertainty Analysis of Power Distribution for NESTOR Based on the Double Latin Hypercube Sampling Method

Due to the complexity of the reactor system, many approximations are used in the nuclear design and calculations inevitably. The accuracy of the nuclear design software is closely related to the safety of the reactor design and operation. Besides, improving the accuracy is an effective way to excavating the economy of nuclear power plants. To analyze the uncertainty of power distribution for NESTOR nuclear design software, we suggested an uncertainty analysis method based on the double Latin Hypercube Sampling (LHS) method and the random sampling statistical analysis (RSSA) method, and built an uncertainty analysis process based on LHS. The uncertainty of physical model and the uncertainty of the change of parameters were both taken into consideration with the double samplings, and 3481 core states were generated by the double samplings. Therefore, the uncertainty of power distribution could be directly analyzed through modeling computation, and the uncertainty of radial power distribution was achieved as ±3.856% under the condition of 95% confidence coefficient and 95% probability. Meanwhile, according to deviation transmission idea, we obtained the uncertainty of power distribution from physical models and the change of parameters based on the measured method. The result shows that the accuracy using the double sampling method is nearly the same to which achieved by the deviation transmission idea, and more conservative.