A Study On Sealing Design And Reliability Analysis Of Reactor Pressure Vessel

The reactor pressure vessel(RPV)is an important part of the pressure boundary of the reactor coolant system,which is used to contain the reactor core and provides a secondary barrier for the release of radioactive fission products.In the analysis and design of RPV,it is necessary not only to ensure that the structure is not damaged and broken,but also to ensure that there is no leakage of radioactive substances under operating conditions.However,the RPV has long service in the environment of high temperature,high pressure and high radioactivity,and bears large fluctuation on pressure and temperature.The sealing of its flange connection is relatively weak,and high-pressure fluid leakage may occur.Therefore,the sealing performance design and reliability analysis of RPV are significant in the analysis and design process.Based on the research method of combining experiment,numerical analysis and theory,systematic study on the reliability of RPV sealing structure was conducted,and the sensitive factors affecting the sealing performance was numerically analyzed by the variable-controlling method.Combined with Bayesian network scoring algorithm based on expert knowledge,the influence of structural parameters is comprehensively evaluated.Carrying out the application research of RPV seal digital design,and the optimal design and reliability evaluation of RPV seal structure were relized.Firstly,the strain and displacement of spherical head reactor pressure vessel under hydrostatic test condition were experimentally studied,and the evolution laws of flange axial separation,bolt load and flange angle were obtained.Based on three-dimensional finite element method and test,the state response of structural parameters to spherical top cover RPV and flat top cover RPV under design conditions and start-up and shutdown transient conditions were comprehensively explored.The results show that the spherical head is obviously better than the flat head in terms of sealing performance.The starting point of ramp,the ramp angle,the pitch diameter of the inner and outer sealing rings,the thickness of the upper flange and the head,and the bolt preload all have obvious effects on the sealing performance.Moreover,the surfacing material has obvious < 001 > texture,and the elastic modulus of the material measured in the laboratory is significantly lower than that provided in the RCC-M specification,while the yield stress measured is significantly higher than that provided in the specification.Low elastic modulus and high yield stress will degrade the sealing performance of pressure vessels.Then,a Bayesian network scoring algorithm based on expert knowledge was developed through information fusion technology,and was used to comprehensively analyze the structural parameters such as axial separation of flange and state parameters such as flange angle,and identify key sensitive factors.At the same time,a new weighted loss function was proposed as a measurement function of seal reliability,which integrated the importance of quantifying sensitivity parameters and the effect of deviation on seal reliability,and provided a basis for more reasonable reliability analysis and design of RPV.Finally,the sealing performance analysis software system of RPV seal structure is established based on advanced data management,finite element simulation and artificial intelligence.Furthermore,based on the embedded machine learning model and sealing reliability scoring function,the RPV sealing structure optimization and sealing performance evaluation of parametric design are realized.