A New Explicit Time Integration Algorithm for Nonlinear Seismic Response Analysis of Site
-
摘要: 针对线弹性结构动力学方程,作者已提出一种具有良好稳定性的二阶精度单步显式时间积分算法。本文将该方法推广到求解材料非线性结构动力学方程中,采用带误差控制的修正欧拉算法计算单元应力,提高显式时间积分算法的精度。将求解非线性问题的显式算法应用于地震波垂直入射时非线性地震反应分析中,使用黏性边界模拟场地土层底部半空间基岩的辐射阻尼,并考虑地震动输入。与中心差分法计算结果进行对比,以表明新显式算法的有效性。Abstract: For the linear-elastic structure dynamics equation, the author has proposed a second-order precision single-step explicit time integration algorithm with good stability. In this paper, the method is extended to solve the nonlinear structural dynamics equations, and the modified Euler algorithm with error control is used to calculate the element stress and improve the accuracy of the explicit time integration algorithm. The explicit algorithm for solving nonlinear problems is applied to nonlinear site seismic response analysis. The viscous boundary is used to simulate the radiation damping of the bedrock underlying the soil layer, and the ground motion input is considered. The comparison with the calculation results of the central difference method shows the effectiveness of the new explicit algorithm.
-
引言
关于场地地震反应的分析已有大量研究成果,研究表明土壤在地震作用下会表现出材料非线性效应ADDIN EN.CITE.DATA(Joyner等,1975;Huang等,2001;Arslan等,2006;Hosseini等,2012)。等效线性化方法ADDIN EN.CITE.DATA(Schnabel等,1972;Idriss等,1992;Bardet等,2000;王笃国等,2016)是一种频域方法,通过在不同土体应变条件下选择等效阻尼比和剪切模量,将非线性问题转化为线性问题。当采用材料非线性本构模型描述土体非线性时,需采用时间积分算法求解非线性动力有限元方程。时间积分算法可分为隐式方法和显式方法。隐式算法每时刻需求解线性代数方程组,计算效率相对较低,如Wilson-θ法和Newmark法等。显式算法无需求解线性代数方程组,适合于强非线性和自由度数目较大的问题。研究者已提出多种显式时间积分算法ADDIN EN.CITE.DATA(Chung等,1994;王进廷等,2002;Belytschko等,2014)。作者近期提出一种二阶精度的单步显式算法,该算法适合变时步问题,在线弹性范围内稳定性较好。本文将该算法推广至求解非线性动力有限元方程中,并将其应用于地震波垂直入射时非线性地震反应分析。
1. 非线性动力有限元方程的显式时间积分算法
设已知非线性体系第${t_i}$时步的受力状态,求解第${t_{i + 1}}$时步的非线性结构动力学方程:
$${\boldsymbol{M}}{{\boldsymbol{\ddot u}}_{i + 1}}{\boldsymbol{ + C}}{{\boldsymbol{\dot u}}_{i + 1}} + {\boldsymbol{f}}_{i + 1}^S{\boldsymbol{ = }}{{\boldsymbol{f}}_{i + 1}}$$ (1) 式中M、C、${{\boldsymbol{f}}^S}$和${\boldsymbol{f}}$分别表示非线性体系的质量矩阵、阻尼矩阵、内力向量和外荷载向量;u表示位移,点号对时间t求导,i+1表示第${t_{i + 1}}$时刻。第i+1时刻时间步长为:
$${\boldsymbol{\Delta }}{t_i} = {t_{i + 1}} - {t_i}$$ (2) 文献显式方法求解非线性方程(1)的过程如下,第i+1时刻位移${{\boldsymbol{u}}_{i + 1}}$为:
$${{\boldsymbol{u}}_{i + 1}} = {{\boldsymbol{u}}_i} + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{t_i}{{\boldsymbol{\dot u}}_i} + \frac{{\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{t_i}^2}}{2}{{\boldsymbol{\ddot u}}_i}$$ (3) 第i+1时刻位移增量$\mathit{\Delta }{{\boldsymbol{u}}_i}$、内力增量$\mathit{\Delta }{\boldsymbol{f}}_i^S$和内力全量${\boldsymbol{f}}_{i + 1}^S$分别为:
$$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{{\boldsymbol{u}}_i} = {{\boldsymbol{u}}_{i + 1}} - {{\boldsymbol{u}}_i}$$ (4) $$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{f}}_i^S = {\boldsymbol{f}}(\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{{\boldsymbol{u}}_i})$$ (5) $${\boldsymbol{f}}_{i + 1}^S = {\boldsymbol{f}}_i^S + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{f}}_i^S$$ (6) 第i+1时刻预估速度${{\boldsymbol{\dot {\tilde u}}}_{i + 1}}$、预估加速度${{\boldsymbol{\ddot {\tilde u}}}_{i + 1}}$、速度${{\boldsymbol{\dot u}}_{i + 1}}$和加速度${{\boldsymbol{\ddot u}}_{i + 1}}$分别为
$${{\boldsymbol{\dot {\tilde u}}}_{i + 1}} = {{\boldsymbol{\dot u}}_i} + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{t_i}{{\boldsymbol{\ddot u}}_i}$$ (7) $${{\boldsymbol{\ddot {\tilde u}}}_{i + 1}} = {{\boldsymbol{M}}^{ - 1}}({{\boldsymbol{f}}_{i + 1}} - {\boldsymbol{C\dot {\tilde u}}}_{i + 1}^{} - {\boldsymbol{f}}_{i + 1}^S)$$ (8) $${{\boldsymbol{\dot u}}_{i + 1}} = {{\boldsymbol{\dot u}}_i} + \frac{{\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{t_i}}}{2}({{\boldsymbol{\ddot u}}_i} + {{\boldsymbol{\ddot {\tilde u}}}_{i + 1}})$$ (9) $${{\boldsymbol{\ddot u}}_{i + 1}} = {{\boldsymbol{M}}^{ - 1}}({{\boldsymbol{f}}_{i + 1}} - {\boldsymbol{C\dot u}}_{i + 1}^{} - {\boldsymbol{f}}_{i + 1}^S)$$ (10) 式(3)—式(10)为求解式(1)的显式算法。算法中需由位移增量计算内力增量,目前常用的应力计算方法包括向前欧拉法、向后欧拉法和完全隐式计算法等ADDIN EN.CITE.DATA(Sloan等,1992;2001;Ahadi等,2003)。下面给出式(5)由位移增量计算内力增量的过程,即一种带误差控制的修正欧拉算法。
对于每个有限单元,由位移增量$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{u}}_i^e$计算应变增量$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e$的表达式为:
$$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e = {{\boldsymbol{B}}^e}\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{u}}_i^e$$ (11) 式中Be为应变矩阵。将ti时刻单元应变增量$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e$赋值给子步应变增量$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e$,ti时刻单元应力${\boldsymbol{ \pmb{\mathit{ σ}} }}_i^e$赋值给${\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e$,初始化子步应变增量和应力状态分别为:
$$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e \leftarrow \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e$$ (12) $${\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e \leftarrow {\boldsymbol{ \pmb{\mathit{ σ}} }}_i^e$$ (13) 每个子步中应力增量计算思路见图 1,具体计算公式如下:
$${\boldsymbol{D}}_1^e = {\boldsymbol{D}}({\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e)$$ (14) $$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_1^e = {\boldsymbol{D}}_1^e\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e$$ (15) $${\boldsymbol{D}}_2^e = {\boldsymbol{D}}({\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_1^e)$$ (16) $$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_2^e = {\boldsymbol{D}}_2^e\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e$$ (17) $$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_s^e = \frac{{\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_1^e + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_2^e}}{2}$$ (18) 式中${{\boldsymbol{D}}^e}$为单元应力-应变关系矩阵。判断每个子步中应力增量$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{{\boldsymbol{ \pmb{\mathit{ σ}} }}_s}$是否符合精度要求的误差判断式为:
$${e_r} = \frac{{\left\| {\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_1^e - \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_2^e} \right\|}}{{\left\| {{\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_s^e} \right\|}}$$ (19) 判断误差er是否小于预先给定的判断值st,条件不满足时,缩小子步应变增量为:
$$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e \leftarrow A\sqrt {{{{s_t}} \mathord{\left/ {\vphantom {{{s_t}} {{e_r}}}} \right. } {{e_r}}}} \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e$$ (20) 式中A为误差峰值系数。采用缩小的子步应变增量重新进行式(14)—式(19)的计算与判断,循环直至满足精度要求,更新剩余应变增量和应力状态分别为:
$$\mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e \leftarrow \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_i^e - \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ ε}} }}_s^e$$ (21) $${\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e \leftarrow {\boldsymbol{ \pmb{\mathit{ σ}} }}_{i + 1}^e + \mathit{\boldsymbol{ \boldsymbol{\varDelta} }}{\boldsymbol{ \pmb{\mathit{ σ}} }}_s^e$$ (22) 利用更新剩余应变增量和应力状态循环执行式(14)—式(20),直至剩余应变增量小于等于零结束。
利用求得的第i+1时刻单元应力可得到单元应力增量和内力增量分别为:
$$ \Delta \boldsymbol{\sigma }_i^e = \boldsymbol{\sigma }_{i + 1}^e - \boldsymbol{\sigma }_i^e $$ (23) $$ \Delta {\boldsymbol{f}}_i^S{\rm{ = }}\sum\limits_e {\int {{{\boldsymbol{B}}^{e{\rm{T}}}}\boldsymbol{\Delta }{\boldsymbol{\sigma }}_i^e{\bf{d}}A} } $$ (24) 2. 地震波垂直入射时场地非线性地震反应分析
本节将上述非线性有限元方程的显式时间积分算法应用于地震波垂直入射时场地非线性地震反应分析中。假定基岩为线弹性半空间,考虑基岩上覆土层的材料非线性,不考虑土体阻尼。在土层下部设置黏性边界条件模拟半空间基岩的辐射阻尼,并在该处以等效结点力的方式实现地震动输入。
计算模型见图 2,选取A点作为观测点。土体非线性材料本构模型选取邓肯-张模型,土体线弹性参数见表 1,未给出配套的非线性参数,故算例中的非线性参数参考实际情况选取,后续研究中将使用更真实表现土体非线性行为的本构模型及真实工程场地参数。算例中的大气压参数取100kPa,内摩擦角增量取0°。入射地震动分别选取狄拉克脉冲和实测地震动(Gilroy Array #3,Coyote Lake, 1979)。入射狄拉克脉冲见图 3,观测点结果见图 4,实测地震动见图 5,观测点结果见图 6。图 4、图 6中给出采用中心差分法的计算结果作为参考解,由图 4、图 6可知,本文算法与中心差分法计算结果吻合较好,说明本文算法的有效性。
表 1 土层参数Table 1. Parameters of soils土质 深度/
m$\rho $/
(g/cm3)cs /
(m/s)v
-EN
-Rf
-c/
(MPa)θ/(°) D
-F
-人工填土 0—1.0 1.9 140 0.33 0.33 0.758 0.084 26.9 1.06 0.021 全新世砂土 1.0—5.1 1.9 140 0.32 0.33 0.758 0.084 26.9 1.06 0.021 全新世砂土 5.1—8.3 1.9 170 0.32 0.36 0.768 0.120 31.0 1.11 0.015 更新世粘土 8.3—11.4 1.9 190 0.40 0.44 0.822 0.188 28.4 1.01 0.012 更新世粘土 11.4—17.2 1.9 240 0.30 0.44 0.822 0.188 28.4 1.01 0.012 更新世砂土 17.2—22.2 2.0 330 0.26 0.51 0.840 0.300 30.0 1.02 0.011 基岩 >22.2 2.0 330 0.26 - - - - - - 表 1中ρ、cs、v、EN、Rf、c、θ为模型参数,分别表示密度、剪切波速、泊松比、无量纲幂次、破坏比、土的内聚力、土的摩擦角。D、F为试验常数。
3. 结论
本文发展一种求解材料非线性结构动力学方程的显式时间积分算法,并应用于地震波竖直入射时非线性地震反应分析中,通过算例验证了该方法的有效性。该显式算法具有无需对角阻尼矩阵、单步、稳定性良好等优点。本文考虑了邓肯-张非线性弹性本构模型,下步研究可考虑将该显式算法扩展到弹塑性本构模型及更能反映土层真实变形的本构模型中。
-
表 1 土层参数
Table 1. Parameters of soils
土质 深度/
m$\rho $/
(g/cm3)cs /
(m/s)v
-EN
-Rf
-c/
(MPa)θ/(°) D
-F
-人工填土 0—1.0 1.9 140 0.33 0.33 0.758 0.084 26.9 1.06 0.021 全新世砂土 1.0—5.1 1.9 140 0.32 0.33 0.758 0.084 26.9 1.06 0.021 全新世砂土 5.1—8.3 1.9 170 0.32 0.36 0.768 0.120 31.0 1.11 0.015 更新世粘土 8.3—11.4 1.9 190 0.40 0.44 0.822 0.188 28.4 1.01 0.012 更新世粘土 11.4—17.2 1.9 240 0.30 0.44 0.822 0.188 28.4 1.01 0.012 更新世砂土 17.2—22.2 2.0 330 0.26 0.51 0.840 0.300 30.0 1.02 0.011 基岩 >22.2 2.0 330 0.26 - - - - - - -
杜修力, 李洋, 许成顺等, 2016.1995年日本阪神地震大开地铁车站震害原因及成灾机理分析研究进展.岩土工程学报, 40(2):223-236. http://d.old.wanfangdata.com.cn/Periodical/ytgcxb201802002 栾茂田, 林皋, 1992.场地地震反应一维非线性计算模型.工程力学, 9(1):94-103. http://www.cnki.com.cn/Article/CJFDTotal-GCLX199201015.htm 王笃国, 赵成刚, 2016.地震波斜入射时二维成层介质自由场求解的等效线性化方法.岩土工程学报, 38(3):554-561. http://d.old.wanfangdata.com.cn/Periodical/ytgcxb201603020 王进廷, 杜修力, 2002.有阻尼体系动力分析的一种显式差分法.工程力学, 19(3):109-112. http://d.old.wanfangdata.com.cn/Periodical/gclx200203022 尹候权, 2015.地震波斜入射时成层半空间场地反应分析方法及其应用.北京: 北京工业大学. Ahadi A., Krenk S., 2003. Implicit integration of plasticity models for granular materials. Computer Methods in Applied Mechanics and Engineering, 192(31-32):3471-3488. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dd253a3e8dddeed056e6c5a8b1aa1498 Arslan H., Siyahi B., 2006. A comparative study on linear and nonlinear site response analysis. Environmental Geology, 50(8):1193-1200. doi: 10.1007/s00254-006-0291-4 Bardet J. P., Ichii K., Lin C. H., 2000. EERA-A computer program for equivalent-linear earthquake site response analyses of layered soil deposits. Belytschko T., Liu W. K., Moran B., et al., 2014. Nonlinear finite elements for continua and structures. 2nd updated and extended ed. John Wiley & Sons Inc. Chopra A. K., 2009. Dynamics of structures: Theory and applications to earthquake engineering (3rd edn). Tsinghua University Press: Beijing. Chung J., Lee J. M., 1994. A new family of explicit time integration methods for linear and non-linear structural dynamics. International Journal for Numerical Methods in Engineering, 37(23):3961-3976. doi: 10.1002/nme.1620372303 Crisfield M. A., 1991. Nonlinear finite element analysis of solids and structures. Journal of Engineering Mechanics, 17(6):1504-1505. http://d.old.wanfangdata.com.cn/OAPaper/oai_arXiv.org_1110.5321 Duncan J. M., Chang C. Y., 1970. Nonlinear analysis of stress and strain in soils. Journal of the Soil Mechanics and Foundations Division ASCE, 96(5):1629-1653. http://ci.nii.ac.jp/naid/10007805651 Hashash Y. M. A., Phillips C., Groholski D. R., 2010. Recent advances in non-linear site response analysis. Fifth International Conference on Recent Advances in Geotechnical Earthquake Engingeering and Soil Dynamics and Symposium in Honor of Professor I.M. Idriss. San Diego, California. Hosseini S. M. M. M., Pajouh M. A., 2012. Comparative study on the equivalent linear and the fully nonlinear site response analysis approaches. Arabian Journal of Geosciences, 5(4):587-597. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=8661fab143f2e354f593dd1ff81acd3c Huang H. C., Shieh C. S., Chiu H. C., 2001. Linear and nonlinear behaviors of soft soil layers using Lotung downhole array in Taiwan. Terrestrial, Atmospheric and Oceanic Sciences, 12(3):503-524. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=CC027479014 Idriss I. M., Sun J. I., 1992. User's Manual for SHAKE91: A computer program for conducting equivalent linear seismic response analysis of horizontally layered soil deposits. Center for Geotechnical Modeling, Department of Civil and Environmental Engineering, University of California. Joyner W. B., Chen A. T. F., 1975. Calculation of nonlinear ground response in earthquakes. Bulletin of the Seismological Society of America, 65(5):1315-1336. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=Doaj000003603734 Schnabel P. B., Lysmer J., Seed H. B., 1972. SHAKE a computer program for earthquake response analysisi of horizontally layered sites. Report No. UBC/EERC72-12. Berkeley, USA: Earthquake Research Center, University of California. Sloan S. W., Abbo A. J., Sheng D., 2001. Refined explicit integration of elastoplastic models with automatic error control. Engineering Computations, 18(1-2):121-194. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=76c28dcc34c077dea2b16ae6d6674092 Sloan S. W., Booker J. R., 1992. Integration of tresca and mohr-coulomb constitutive relations in plane strain elastoplasticity. International Journal for Numerical Methods in Engineering, 33(1):163-196. doi: 10.1002/nme.1620330112 Zhao M., Li H., Cao S. et al. 2019. An explicit time integration algorithm for linear and non-linear finite element analyses of dynamic and wave problems. Engineering Computations, 36(1):161-177. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=497e43137d940ce8f0a85b06178f0517 -