Nuclear Power Floor Spectrum Parameter Uncertainty And Sensitivity Analysis

With the increasingly serious environmental problems, nuclear power as a clean energy has been paid more and more attention, and the development of nuclear power in China is entering a golden period. To ensure the safety of nuclear power structure is a prerequisite for the development of nuclear power, it is necessary for the nuclear structure to carry on the reasonable seismic response analysis to make the seismic safety evaluation results reliable. In the dynamic analysis of nuclear structure two aspects is the key:one is how to consider the interaction of structure and foundation, because the lumped mass model specification gives the convenient spring and damper method, but the finite element of three-dimensional model analysis is applied to more complex artificial boundary, and this is not suitable for engineering applications; the second is that the parameters are uncertainty due to the soil and structure’s material randomness and assumption of the simplified model, so that it needs to analysis the structure considering the parameters’uncertainty to get the floor response spectrum with statistical significance and summarize the rules of the influence of parameter uncertainty on floor response spectrum. Based on the content above, this article mainly doing the research from the following several aspects:(1) By3D finite element analysis of nuclear structure, the lumped mass spring damper model method is applied. Through the torque equivalent, it will be ok to establish the nuclear power three-dimensional soil structure spring damper model, which is simple and convenient for engineering applications. The correctness will be improved by the comparison of modal and modal mass between the3D spring and damper model and the lumped mass model, which are coincide with each other.(2)It is able to analyse the nuclear structure considering parameter uncertainty by using MATLAB and ANSYS software based on the theory of Monte Carlo and explore the parameter uncertainty on the floor response spectrum. The example demonstrates that the90%guarantee rate of floor response spectrum analysis result compared with the deterministic analysis result exist the consistent "peak" and "peak extension" change with the specification description.But in order to save computing time, it is generally take the small sample cycle, so that the calculation accuracy can’t be guaranteed. In order to solve this problem, we change the MATLAB into parallel computing by using the principle of computer and software of multi-processor; and proposed the method of parallel modeling and unified operation mode to realize parallel loops on ANSYS software. The numerical example shows that, after the two improved algorithms, it is able to achieve the Monte Carlo sample analysis to ensure the accuracy of the statistical results of floor response spectrum, and greatly improves the computation efficiency; at the same time, the correctness of two kinds of software also has been demonstrated.(3) Through the nuclear structure parameter uncertainty analysis, it can effectively grasp the overall random rules, but cannot be delved into stochastic theory about specific parameters. In this paper, on the basis of global analysis, we will explore each random parameter’ influence on floor response spectrum, and the floor response spectrum’change with the change of shear modulus of the foundation from the perspective of parsing, and then explores the floor spectrum is more sensitive to which performance parameter. The results show that the peak "drop" and "peak extension" are common due to one or several parameters; floor spectral sensitivity to random parameters varied with the site conditions and floor spectrum frequency, and will obvious fluctuate in the peak frequency.