【作者】 侯春林；

【导师】 李小军；

【作者基本信息】 中国地震局工程力学研究所 ， 结构工程， 2012， 博士

【摘要】 中国已运行和在建的核电厂涉及的技术种类不一，有些使用国内技术,有些引进国外技术，且引进的国家及技术均有所不同，这就造成现有中国核电厂设计依据的法规和标准有所不同。对于核电厂的抗震分析，核电厂设计依据不同必然对一定阶段下核电厂抗震设计的深度和实际核电厂抗震设防水准造成一定的影响，本文便基于实际核电厂工程抗震设计过程中因引用不同的规范造成核电厂许可过程和设计过程方面差异的一系列问题进行了比较、分析和研究，并给出了一些可供实际工程应用的研究结果，同时为更精确的开展计算分析，发展和提出了一种拟合多阻尼反应谱的人工地震动时程优化算法。研究工作为我国核电厂抗震设计、安全评估以及相关法规标准的完善和修订提供有益的参考。本文的主要工作如下：(1)以AP1000核电机组在中美两国申请许可证过程为例，比较分析了中国和美国核电厂许可证管理程序的差异，展示了对于同一核电机组依据不同的规范审评也可能对许可证管理过程造成一定的差异。(2)利用ASCE4-98和RCC-G两种不同的规范方法进行了CPR1000安全壳结构的阻抗函数计算分析，研究了不同阻抗函数计算方法在核电厂结构地震方面应用引起的差异性。在考虑地基土剪切波速、泊松比及地基土密度三种影响因素的情况下，分别探讨了不同的地基参数对应的两种规范方法得到的计算结果，给出了两种规范分析方法计算的地基刚度和阻尼的差异，验证了地基土剪切波速是影响地基阻抗函数大小的关键因素。并基于上述计算结果，分析CPR1000安全壳结构的地震反应，研究了地基土剪切波速、泊松比及地基土密度三种影响因素变化的情况下核电厂结构的反应规律。计算结果说明：ASCE4-98计算地基阻抗函数的方法较于RCC-G方法保守；地基土密度对核电厂结构响应的影响较小；地基土泊松比对核电厂结构响应的影响可以忽略；随着地基土剪切波速的增大，核电厂结构响应变化幅度降低，当剪切波速从1900m/s增加至2200m/s时，安全壳结构响应已出现拐点，考虑到整个计算过程的不确定性，建议当地基土剪切波速大于等于1900m/s时，CPR1000安全壳结构可不考虑地基与结构之间的相互作用。(3)基于AP1000核岛结构设计地基的场地参数模型，分析选取不同地震输入界面时同一特定场地的地表地震动峰值加速度和反应谱的差异。计算结果表明：地震输入界面的不同，AP1000核岛结构设计地基的同一场地土层模型地表地震动峰值加速度差异高达2.25倍，地表地震动加速度反应谱频谱特性发生较大的变化，建议在AP1000核电厂地震安全性评价中应基于剪切波速为2438m/s的基岩层作为土层地震反应分析的地震输入界面，而当一定深度内地基剪切波速达不到剪切波速2438m/s，则应基于地基与结构敏感性分析选择地震输入界面。同时说明在特定厂址评价中应基于同一标高比较厂址特定反应谱和AP1000标准设计地震反应谱的大小。并基于考虑地基与结构之间相互作用的不同方法，给出设计基岩处和基础底部的AP1000设计谱。计算结果说明：基础底部AP1000设计谱不能包络核岛结构抗震设计常用的0.20g标定的RG1.60谱。(4)基于不同核电厂法规涉及人工合成地震动方面的差异，对比现行中国可引用的法规之间在输入地震动的时程方面的异同，并结合工程实践给出了相关的评述和应用建议。同时，基于核电实际工程需要，提出了一种拟合多阻尼反应谱的人工地震动时程优化算法，实现了提高设计地震动时程对更多多阻尼目标反应谱的更高的拟合精度。该方法基于真实地震记录或传统方法获取初始加速度时程，然后从全局角度来考虑添加的信号，使得添加的信号对反应的影响总是在可控范围之内，这种可控性通过对优化的目标函数施加约束来实现。算例分析结果表明，该方法可实现满足5个不同阻尼比下的目标谱的拟合精度的设计地震动时程合成，同时较容易地将设计地震动时程的反应谱与多阻尼目标谱的最大误差降低至5%左右。(5)基于不同规范规定的核电厂抗震分析方法不同，且中国核电厂抗震规范中无复频响应分析方法，使用时程动力法和复频响应分析法计算同一核电厂结构的响应。计算结果表明：对于核电厂结构，因在抗震设计时其结构体系等效为线性体系，满足叠加原理要求，故在计算核电厂结构的响应的时域法或频域法是等价的，计算结果一致；由于频率在振动过程中物理概念的重要性，在频域中处理振动问题有许多优点，且频域解在计算上已无大困难。基于此，建议中国核电厂抗震设计规范中应增加复频响应分析法。(6)基于不同规范涉及竖向地震动方面的规定不同，对比分析竖向地震动选取的方式不同得到的竖向反应谱的差异，并基于AP1000核电厂结构分析竖向反应谱的差异可能对同一核电厂抗震响应造成的影响。计算结果说明：不同方法选取的竖向反应谱对核电厂结构水平向响应的影响可以忽略；峰值相同，反应谱谱形不同的竖向设计反应谱对核电厂结构的竖向峰值加速度的影响可以忽略，但需考虑对竖向相对最大位移和竖向楼层反应谱的影响；在核电厂抗震设计过程中，竖向峰值加速度取水平向峰值加速度的2/3偏于不保守。最后，在总结论文工作的基础上，提出了有待进一步研究的问题。更多还原

【Abstract】 The NPPs (nuclear power plants) that had been built and will be built inChina are involve many types technique, some are come from domestic, somewere introduced from foreign countries, and the countries that the techniquewas used for design NPPs are difference. So we could use many codes thatwere published in other countries, but this status will lead to NPPs in Chinacould accord different codes to design. For the NPPs seismic analysis, if weuse the different regulations, then we maybe reach to different depths ofdesign in one same phase and receive two distinct design results for one sameNPP which is according to two different codes. In this paper, we do series ofstudies on this discrepancy between codes of NPPs seismic design that couldbe used in China, and we give some results of studies that could be used inactual NPPs project, and provide theoretical guidance on NPPs design, safereview of design reports and renovate code. These studies could help theconstruction and research of the later NPPs.The works in this study include:(1)Based on the process of AP1000licensing process in China andAmerican, comparison and analysis of the difference between the manageprograms of nuclear power plant in China and American. We gained that thereare also some differences in same nuclear power plant under two disparatecodes.(2)The earthquake response analysis of power plant structures underdifferent impendence functions of regulations considering the sensitivity ofparameters of ground soil. Based on impendence functions of foundation thatare calculated by two methods of different regulations, we analyze thestructure earthquake response of CPR1000NPP and give the difference ofearthquake response of NPP building by RCC-G and ASCE4-98impendencefunction. Considering the sensitivity of the impendence function of groundsoil, we study the response behavior of three parameters of velocity of ground,Poisson’s ratio of ground, and density of ground change. The results showthat：the computational method of impendence functions of ASCE4-98isconservative compared with the method in RCC-G；The density of ground haslittle impact on the response of NPP; the influence of Poisson’s ratio ofground could be neglected to the response of NPP; the ground soil could reduce the response of earthquakes under some conditions； With theincreasing of the shear wave velocity of ground, the response of NPP changeslittle, and when the velocity varies from1900m/s to2200m/s, there is only2%difference between two horizontal peak acceleration responses of the highestpoint in CPR1000NPP. Considering the uncertainty of the computation, wesuggest that the effect of foundation-structure interaction could be neglectedwhen the shear wave velocity of ground is greater than or is equal to1900m/s.(3)In the process of seismic hazard assessment of nuclear power plant,there are some different requirements in determination of seismic bedrock inthe site model by the related China and America code. In this study, theseismic bedrocks with different shear wave velocities of700m/s,1100m/s,and2438m/s were selected based on the site models for AP1000nuclearpower plant design, and seismic responses of the site models, including thepeak ground accelerations and response spectra, were analyzed by softwareswere widely used in China and America. The results indicated that differentshear wave velocities of seismic bedrocks could lead to a strong difference ofcalculated ground motions of sites for AP1000nuclear power plant design,such as more than2.25times variation of the peak ground accelerations.Therefore, it was suggested that the rock layer with shear wave velocity of2438m/s should be regarded as the seismic bedrock in the seismic safetyanalysis of AP1000nuclear power plant site. When the shear wave velocity ofsoil in certain depth still could not reach2438m/s，then we should select theseismic bedrock based on soil-structure interaction analysis. Based on theintroduction of AP1000Certified Seismic Design Response Spectra，we givethe relationship of difference Response Spectra of definite by differentRegulations, and illuminate that we should compare the site specific ResponseSpectra and AP1000CSDRS on one same ground level. Then, According totwo methods in SSI analysis, utilizing one artificial earthquake waves thatsimulating five damping rations of AP1000CSDRS, calculating two levelsCSDRS of design rock and foundation base on the basis of AP1000designground. We get that the calculated two levels CSDRS of design rock andfoundation base could not envelope RG1.60anchored by0.2g.(4) The main technical requirements in several domestic and foreigncodes involved in the synthesis of design ground motion are interpreted, andthe suggestions could be employed in actual project in China. Based on thepractical engineering requirements of nuclear power plants, we propose anoptimization algorithm for the time history generation to match the multi-damping design spectra. The algorithm could enhance the fittingprecision between the response spectra of time history to target spectra undermulti-damping ratios. In our method, the seed ground motion history could beobtained from actual seismic records or from traditional methods, and then weconsider to superimpose an adjusted ground motion in time domain from aglobal view point, which make the influences of the superimposed signal tothe response spectra be under control. The controllable property isimplemented through adding constraints on the objective function whichneeds to be optimized. In a numerical example, we demonstrate that theground motions generated by our method could match the design spectraunder5different damping ratios in a high precision. The max differencebetween response spectra of artificial time history and design spectra is about5%, and all the response spectra envelope the design spectra.(5)The main technical requirements in several domestic and foreignstandards involved in the methods of seismic analysis of NPPs are different,Based on there is no complex frequency response method in Code for SeismicDesign of Nuclear Power Plants of China, and we do analysis in AP1000nuclear island structure by complex frequency response method and timehistory method. The results of analysis show that the two methods are similaron the basis of the linear analysis of nuclear island. Besides, with thedevelopment of computer science, the complex frequency response methodbecomes much more fast and convenience, and then we suggest that we couldadd complex frequency response method in Code for Seismic Design ofNuclear Power Plants of China.(6)The requirements of vertical ground response of seismic analysis ofNPPs in several domestic and foreign standards are compared，and we do theseismic analysis on the basis of AP1000nuclear island structure and AP1000design response spectrum. The results of analysis show that different methodsof obtaining vertical ground response in seismic evaluation of one site on thehorizontal response of nuclear island structure effect could be neglected；thepeaks of different vertical ground accelerations are same and the shapes ofvertical response spectrum are different which have a little effect on the peaksof response accelerations of nuclear island structure，the effect on responsespectrums of nuclear island structure but need to be considered. In seismicanalysis of building of nuclear power plant, the vertical peak of designacceleration by scaling the corresponding ordinates of the horizontal peak ofdesign acceleration by two-thirds is un-conservative. In the end, we summarized our studied work, and then pointed out theproblems that need to be studied more deeply.更多还原