中图分类号:O34
通讯作者:
收稿日期:2017-10-9
接受日期:2018-02-24
网络出版日期:2018-02-10
版权声明:2018 《力学学报》编辑部 《力学学报》编辑部 所有
基金资助:
展开
摘要
求解含裂纹等不连续问题一直是计算力学的重点研究课题之一,以偏微分方程为基础的连续介质力学方法处理不连续问题时面临很大的困难. 近场动力学方法是一种基于积分方程的非局部理论,在处理不连续问题时有很大的优越性. 本文提出了求解含裂纹热传导问题的一种新的近场动力学与有限元法的耦合方法. 结合近场动力学方法处理不连续问题的优势以及有限元方法计算效率高的优势,将求解区域划分为两个区域,近场动力学区域和有限元区域. 包含裂纹的区域采用近场动力学方法建模,其他区域采用有限元方法建模. 本文提出的耦合方案实施简单方便,近场动力学区域与有限元区域之间不需要设置重叠区域. 耦合方法通过近场动力学粒子与其域内所有粒子(包括近场动力学粒子和有限元节点)以非局部方式连接,有限元节点与其周围的所有粒子以有限元方式相互作用. 将有限元热传导矩阵和近场动力学粒子相互作用矩阵写入同一整体热传导矩阵中,并采用Guyan缩聚法进一步减小计算量. 分别采用连续介质力学方法和近场动力学方法对一维以及二维温度场算例进行模拟,结果表明,本文的耦合方法具有较高的计算精度和计算效率. 该耦合方案可以进一步拓展到热力耦合条件下含裂纹材料和结构的裂纹扩展问题.
关键词:
Abstract
To accurately model discontinuous problems with cracks is one important topic in computational mechanics. It is very difficult to solve discontinuous problems using continuum mechanics methods based on partial differential equations. However, peridynamics (PD), a non-local theory based on integral equations, has great advantages in solving these problems. In this paper, a new method is proposed to solve heat conduction problems with cracks using coupled PD and finite element method (FEM). This method has both the advantage of the computational efficiency of FEM and the advantage of PD in solving discontinuous problems. The computational domain can be partitioned into two regions, PD region and FEM region. The region containing the crack is modeled by PD, and the other region is modeled by FEM. Application of the coupling scheme proposed in this paper is simple and convenient, since there is no need to introduce an overlapping region between PD region and FEM region. As for the coupling approach, the PD particle is connected non-locally to all particles (PD particles and finite element nodes) within its horizon, whereas the finite element node interacts with other nodes in the finite element manner. The heat conduction matrices of FEM and the matrices of the interaction between PD particles are combined to be a global heat conduction matrix. The Guyan reduction method is used to further reduce the computational cost. The temperature fields of a one-dimensional bar and a two-dimensional plate obtained by the coupling approach are compared with classical solutions. Results show that the proposed coupling method is accurate and efficient. The coupling scheme can be extended to solve crack propagation problems with the thermo-mechanical load.
Keywords:
损伤与破坏问题的准确建模是计算力学的重点研究课题之一[1,4]. 近场动力学方法(peridynamics, PD)是基于非局部积分思想发展的数值方法,该方法基于粒子间的长程作用建模,通过近场范围体现结构特征尺度,避免了不连续处变量空间导数不存在而导致的奇异性问题,适用于分析含缺陷、裂纹等不连续问题的力学模型[5]. 作为一种新兴理论,PD 自问世起便备受关注,近年来许多专家学者开始从事这方面的研究[6,15].
虽然PD在模拟裂纹扩展时有很大的优势. 但作为非局部理论,PD计算效率远低于有限元法(finite element method, FEM). 为了发挥PD处理非连续问题以及FEM计算效率高的优势,许多学者在PD与FEM结合方面做了很多研究[16,25]. Macek等[12]将模型划分为PD区域与FEM区域,对两个区域重叠区域中的PD物质点与FEM结点采用位移协调的方法, PD物质点的位移可通过FEM结点位移进行插值后确定,但此过程不可逆,故边界条件只能在FEM区域实施. 文献[21,22,23]发展了混合函数方法,该方法在局部模型和非局部模型的重合区域采用混合函数实现以上两种模型的平滑过渡, 用局部模型的应变能密度与非局部模型的应变能密度混合得到耦合区域的应变能密度,从而得到耦合区域的本构模型. Shojaei等[24]将FEM和PD耦合求解动态裂纹扩展问题,首先将求解区域划分为3个区域,FEM区域、PD区域以及耦合区域. 其中危险区域采用PD粒子进行离散,其他区域采用有限元节点计算,该方案的优势在于在耦合区域内不需要设置混合函数,实施简单.
近场动力学在求解位移场以及破坏问题方面已有广泛的应用,但很少有文章提及近场动力学在热传导问题中的应用. Bobaru等[26,27]根据能量平衡推导出键型PD热传导方程,算例表明解析解与PD热传导方程得到的结果相吻合. 此外,文章分析了近场范围
本文提出了一种求解稳态热传导问题的PD与FEM耦合方法,该方案不设置PD与FEM的重合区域,故不需要引入混合函数, 实施简单方便. 将PD与FEM写入同一个总体热传导矩阵,再采用Guyan缩聚法,进一步减小计算量.
键型PD热传导方程[28]可以写为
$\rho c_v \dot {\varTheta }\left( {{\pmb x},t} \right) = \int_H f_h \Big( \varTheta \left( {\pmb x}',t \right) - \varTheta \left( {{\pmb x},t} \right),{\pmb x}' - {\pmb x},t \Big ) d V_{x}' + \rho S_{\rm b} \left( {{\pmb x},t} \right) (1) $
式中, $\rho $ 代表材料密度,$c_v $为材料比热容,$H$表示以$\delta $为半径的粒子${\pmb x}$的邻域,$f_{\rmh}$代表热流密度函数,$S_{\rm b}$为内热源. $t$和$\varTheta $分别表示时间和温度,$V_{x}'$代表粒子${x}'$所占的体积.
对于不计内热源稳态热传导方程可写为
$ \int_H {f_{\rm h} \left( {\varTheta \left( {{\pmb x}',t} \right) - \varTheta \left( {{\pmb x},t} \right),{\pmb x}' - {\pmb x},t} \right)} d V_{x}'=0 (2) $
将式(2)写为离散形式
$ \sum_{x}' {f_{\rm h} \left( {\varTheta \left( {{\pmb x}',t} \right) - \varTheta \left( {{\pmb x},t} \right),{\pmb x}' - {\pmb x},t} \right)} V_{x}'=0 (3) $
$ f_{\rm h} = \dfrac{\kappa }{\left| \xi \right|}\left( {\varTheta \left( {{\pmb x}'} \right) - \varTheta \left( {\pmb x} \right)} \right) (4)$
式中,$\left| \xi \right| = \left| {{\pmb x}' - {\pmb x}} \right|$,$ \kappa $代表微热传导系数. 为了与长程力更好地匹配,
$ \left.\!\!\begin{align} \kappa = \dfrac{2k}{A\delta ^2}\left[ {1 - \left( {\dfrac{\left| \xi \right|}{\delta }} \right)^2} \right]^2 , \ \hbox{one-dimensional} \\ \kappa = \dfrac{6k}{\pi h\delta ^3}\left[ {1 - \left( {\dfrac{\left| \xi \right|}{\delta }} \right)^2} \right]^2 , \ \hbox{two-dimensional} \\ \kappa = \dfrac{6k}{\pi \delta ^4}\left[ {1 - \left( {\dfrac{\left| \xi \right|}{\delta }} \right)^2} \right]^2 , \ \hbox{three-dimensional} \end{align}\!\!\right\} (5) $
式中,A代表横截面积,h表示平面厚度,k为热传导系数,
对于处于一种材料中的两个PD粒子,微热传导系数
对于处于两种材料中的两个PD粒子的微热传导系数
$ \kappa = \dfrac{l_1 + l_2 }{{l_1 } / {\kappa _1 } + {l_2 } / {\kappa _2 }} (6) $
式中,$l_1 $代表粒子${\pmb x}'$与${\pmb x}$的距离$\left| \xi \right|$在材料1中的部分(图中实线部分),其对应的微热传导系数为$\kappa _{1} $;$l_{2} $代表粒子${\pmb x}'$与${\pmb x}$的距离$\left| \xi \right|$在材料2中的部分(图中虚线部分),其对应的微热传导系数为$\kappa_{2} $. 从图2不难看出$\left| \xi \right| = l_1 + l_2 $,$\kappa_{1} $和 $\kappa _{2} $分别由对应材料的热导率$k_{1}$, $k_{2} $代入式(5)得到.
对于一维热传导问题,FEM单元热传导矩阵可以写为
$ {\pmb K}_{\rm FEM} = \dfrac{kA}{l}\left[ \!\! \begin{array}{cc} 1 & { - 1} \\ { - 1} & 1 \end{array} \!\! \right] = a\left[\!\! \begin{array}{cc} 1 & { - 1} \\ { - 1} & 1 \end{array} \!\! \right] (7) $
式中,$a = \dfrac{kA}{l}$.
根据式(4)和式(5),键型PD中粒子
$ {\pmb K}_{\rm PD} = \dfrac{\kappa }{l}V_x V_{x'} \left[ \!\!\begin{array}{cc} 1 & { - 1} \\ { - 1} & 1 \end{array} \!\! \right] = b\left[\!\!\begin{array}{cc} 1 & { - 1} \\ { - 1} & 1 \end{array} \!\! \right] (8) $
式中,$b = \dfrac{\kappa }{l}V_x V_{x'} $,$ V_{x}$代表粒子${\pmb x}$所占的空间,$V_{x'} $代表粒子${\pmbx}'$所占的空间,对于一维情况,$V_x = A\Delta x$,$A$代表一维模型的横截面积.
对于一维模型,FEM与PD的耦合方法如图2所示. 图中圆点代表FEM节点,方块代表PD粒子. 直线代表有限元节点的作 用,曲线代表PD粒子的作用. 其中,FEM节点与其左右两边的节点(包括FEM节点和PD粒子)以有限元方式相互作用, PD粒子与其域内的节点(包括FEM节点和PD粒子) 以PD方式相互作用. 对于
$\left[\!\!\begin{array}{ccccccccccccc} a & -a & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \cdots \\ -a & {2a} & -a & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \\ 0 & -a & {2a} & -a & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \\ 0 & 0 & -a & {2a} & -a & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \\ 0 & -b/3 & -b/2 & -b & {8b /3} & -b & -b/2 & -b/3 & 0 & 0 & 0 & 0 & \\ 0 & 0 & -b/3 & -b/2 & -b & {8b / 3} & -b & -b/2 & -b/3 & 0 & 0 & 0 & \\ 0 & 0 & 0 & -b/3 & - b/ 2 & -b & 8b /3 & -b & -b/2 & -b/3 & 0 & 0 & \cdots \\ 0 & 0 & 0 & 0 & -b/3 & -b/2 & -b & {8b / 3} & -b & -b/2 & - b/ 3 & 0 & \\ 0 & 0 & 0 & 0 & 0 & - b /3 & -b/2 & -b & {8b/ 3} & -b & -b/2 & -b/3 & \\ \cdots & & & & & & \cdots & & & \cdots & & & \cdots \\ \cdots & & & & & & \cdots & & & & \cdots & & \cdots \\ \cdots & & & & & & \cdots & & & & & \cdots & \cdots \\ \cdots & & & & & & \cdots & & & & & & \cdots \end{array}\!\! \right]\cdot \left[\!\! \begin{array}{c} {\varPhi _{1} } \\ {\varPhi _{2} } \\ {\varPhi _{3} } \\ {\varPhi _{4} } \\ {\varPhi _{5} } \\ {\varPhi _{6} } \\ {\varPhi _{7} } \\ {\varPhi _{8} } \\ {\varPhi _{9} } \\ \cdots \\ \cdots \\ \cdots \\ \cdots \\ \end{array} \!\! \right] = \left[\!\! \begin{array}{c} {P_{1} } \\ {P_{2} } \\ {P_{3} } \\ {P_{4} } \\ {P_{5} } \\ {P_{6} } \\ {P_{7} } \\ {P_{8} } \\ {P_{9} } \\ \cdots \\ \cdots \\ \cdots \\ \cdots \end{array}\!\! \right] (9)$
从式(9)可以看出,耦合后的总体热传导矩阵不再对称,式中第1至第4行为FEM区域对总体热传导矩阵的贡献,第5至第9行为PD区域对总体热传导矩阵的贡献. 通过式(9)可以清晰的看出,近场动力学区域的宽度比有限元区域的宽. 式(9)第4行对应图2中FEM节点
与一维情况类似,二维情况下FEM与PD的耦合方法如图3所示.
两区域交界附近的PD粒子
$ \left[\!\! \begin{array}{ccccc} & \vdots & & \vdots & \\ \cdots & \sum_j^N {k_{cj}^{11} } & \cdots & {k_{cd}^{12} } & \cdots \\ & \vdots & & \vdots & \\ \cdots & {k_{dc} } & \cdots & {\sum_i^M {k_{dd} } } & \cdots \\ & \vdots & & \vdots & \end{array} \!\! \right] \cdot \left\{ \!\!\begin{array}{c} \vdots \\ {\varTheta _c } \\ \vdots \\ {\varTheta _d } \\ \vdots \end{array} \!\! \right\} = \left\{ \!\! \begin{array}{c} \vdots \\ {P_c } \\ \vdots \\ {P_d } \\ \vdots \end{array} \!\! \right\} (10)$
式中,$P_c $和$P_d $代表节点的热流.
$ P_c = -\sum_i^M\dfrac {1} {2}q S (11) $
其中,
其中
$ {\pmb K}_{\rm PD}^{cd} = \dfrac{\kappa }{l}V_c V_d \left[ \!\! \begin{array}{cc} 1 & { - 1} \\ { - 1} & 1 \end{array} \!\! \right]{ = }\left[\!\! \begin{array}{cc} {k_{cd}^{11} } & {k_{cd}^{12} } \\ {k_{cd}^{21} } & {k_{cd}^{22} } \end{array} \!\! \right] (13) $
其中,
式(10) 中的整体热传导矩阵由FEM 的单元热传导矩阵与PD 粒子热传导矩阵装配得到,其中PD 粒子热传导矩阵的装配方法与FEM 的相同.
将式(10) 写为分块格式
$ \left[\!\! \begin{array}{cc} {k_{ii} } & {k_{io} } \\ {k_{oi} } & {k_{oo} } \end{array} \!\! \right]\left[\!\! \begin{array}{c} {\varTheta _i } \\ {\varTheta _o } \end{array} \!\! \right] = \left[\!\! \begin{array}{c} {P_i } \\ {P_o } \end{array} \!\! \right] (14)$
式中,下标i代表从自由度,下标o代表主自由度,采用Guyan缩聚法[34],式(14)可写为
$ k_{c0} \varTheta _o = P_{c0} (15) $
式(15)为缩聚后的有限元方程,$k_{c0} $的阶数与$k_{oo} $的相同,其中
$k_{c0} { = }k_{oo} - k_{oi} k_{ii}^{ - 1} k_{io} (16)$
$P_{c0} { = }P_o - k_{ii}^{ - 1} k_{io} P_i (17)$
通过式(15) 可以得到主自由度的温度值. 从自由度的温度可以由式(14) 第1 行得到
$ \Theta _i { = } - k_{ii}^{ - 1} \left( {P_i - k_{io} \varTheta _o } \right) (18) $
采用Guyan法将PD和FEM耦合求解热传导问题的计算效率进一步提高.
如图4所示的一维热传导模型,模型长
分别取两类边界条件,对比不同边界下解析解与本文耦合方法的计算结果.
第一类边界条件
第二类边界条件
其中,
本文耦合方法与解析解的计算结果如图5所示. 从图中可以看出,采用本文耦合方法得到的结果与解析解吻合很好.
图 5 一维温度场模型模拟结果
Fig. 5 Simulation results for one-dimensional temperature field model
表1为一维温度场模型正则化的绝对误差以及相对误差. 绝对误差由下式计算
$ \left| {e_{\rm abs} } \right| = \sqrt {\sum_{i = 1}^{N_{1} } {e_{{\rm abs},i}^2 } } (21) $
式中,
相对误差的表达式为
$ \left| {e_{\rm rel} } \right| = \sqrt {\sum_{i = 1}^{N_{1} } {e_{{\rm rel},i}^2 } } (22) $
式中,
针对二维热传导模型如图6所示,模型取长
图6 FEM/PD区域划分及子结构划分示意图
Fig. 6 FEM/PD region division and substructure division diagram
接下来讨论本文耦合算法的收敛性. 对
$ \left. \varTheta \right|_{x = 0} = \left. \varTheta \right|_{x = a} = \left. \varTheta \right|_{y = 0} = 0 {\rm K} , \ \ \left. \varTheta \right|_{y = b} = 100 {\rm K} (23) $
采用本文耦合方法得到的结果与ABAQUS计算得到的FEM结果对比.图7为不同
图7 不同
Fig. 7 Simulation results of temperature field of coupling model with different
对于
图8 不同
Fig. 8 Simulation results of temperature field of coupling model with different
通过以上
$ \left. \varTheta \right|_{x = 0} = \left. \varTheta \right|_{x = a} = \left. \varTheta \right|_{y = 0} = 0 {\rm K} , \ \ \left. q \right|_{y = b} = 1 {\rm W} /{\rm m}^{2} (24) $
分别采用经典有限元理论和本文耦合方法对本节的模型进行求解(均采用显式计算). 模拟结果如图9所示,其中左边一组为 有限元模拟结果(ABAQUS),右边一组为本文耦合方法的结果. 由图9可知,本文耦合方法与有限元法的模拟结果吻合很好. 图形的上半部分温度梯度较大,下半部分温度 梯度很小. 对于第一类边界条件,采用两种方法计算的温度最大值均为100 K;对于第二类边界条件,采用有限元法计算得到温度最大值为190.83 K,采用本文方法得到温度最大值为190.98 K,相对误差为0.08%.
在2.2节二维热传导模型的基础上,在模型中间开一个裂纹(如图10所示),裂纹位置由下式表达
$ 0.25a < x < 0.75a , \ \ y = 0.5b (25) $
边界条件仍按式(19)、式(20)选取. PD区域引入初始裂纹有两种方法,一种为破坏经过裂纹线上的所有键, 如图11(a)所示. 另一种方法为移除裂纹线上的粒子同时破坏经过裂纹线上的所有键[35],如图11(b)所示,本文采用第二种形式.
图9 二维温度场模型模拟结果
Fig. 9 Temperature field model simulation results of two-dimensional
图9 二维温度场模型模拟结果(续)
Fig. 9 Temperature field model simulation results of two-dimensional (continued)
分别采用传统FEM理论、本文耦合方案和PD理论对含裂纹二维温度场进行模拟,采用式(23) (第一类边 界条件)的模拟结果如图12所示,采用式(24) (第二类边界条件)的模拟结果如图13所示. 从图12和图13可以看出,在以上两类边界条件下,裂纹上下表面的温度梯度最大,这是因为裂纹的存在阻断了热量的传播. 由于裂纹的阻隔以及左右边界的散热,热量几乎不能传到裂纹下侧.
图12 第一类边界条件下含裂纹二维温度场模拟结果
Fig. 12 Temperature field results of two-dimensional problem with crack under the first kind of boundary condition
图13 第二类边界条件下含裂纹二维温度场模拟结果
Fig. 13 Temperature field results of two-dimensional problem with crack under the second kind of boundary condition
对比图9与图12,图13可知,模型中的裂纹对温度场分布造成了很大的影响. 此外,本文耦合方案的结果与传统FEM理论、PD理论的结果相吻合.
需要指出,本文的耦合方案以及PD模拟结果由编制的MATLAB程序实现,程序采用显式计算,对以上两个算例均采用1000个 增量步进行计算,计算时间如表2所示
从表2可以看出,本文耦合方案和PD模型相比可以在保证计算精度的前提下将计算效率提高5倍左右,这充分说明了本文耦合方案的在提高计算效率上的优势.
本文计算采用计算机的硬件条件如下:
处理器: i5-4590 CPU @3.30 GHz
内存: 16.00 GB
操作系统: Windows 10 Enterprise 64 bit
本文给出了有效处理含裂纹结构热传导问题的一种新的有限元与近场动力学耦合方法. 该方法结合了近场动力学处理含裂纹问题和有限单元法处理连续问题的优势,并应用Guyan缩聚法进一步减小计算量. 通过数值算例表明,本文耦合方法得到的温度场结果准确. 利用该耦合方案可以进一步拓展到热力耦合条件下含裂纹材料和结构的裂纹扩展问题.
The authors have declared that no competing interests exist.
[1] |
爆炸与冲击问题的大规模高精度计算 .Large scale high precision computation for explosion and impact problems . |
[2] |
第七届全国固体力学青年学者学术研讨会报告综述 .
Review of the seventh national symposium on solid mechanics for young scholars .
|
[3] |
含层理页岩气藏水力压裂裂纹扩展规律解析分析 .
Hydraulic fracture propagation in shale gas bedding reservoir analytical analysis .
|
[4] |
静动组合载荷下混凝土率效应机理及强度准则 .Study on strain rate effect and strength criterion of concrete under static-dynamic coupled loading . |
[5] |
Peridynamic theory of solid mechanics . |
[6] |
Origin and effect of nonlocality in a composite . |
[7] |
Wave dispersion analysis and simulation method for concrete SHPB test in peridynamics .
|
[8] |
Peridynamic Theory and Its Application . |
[9] |
冲击载荷作用下颗粒材料动态力学响应的近场动力学模拟 .
Peridynamics simulation for dynamic response of granular materials under impact loading,
|
[10] |
近场动力学方法及其应用 .
A review on peridynamics method and its application .
|
[11] |
基于近场动力学理论的层压板损伤分析方法 .
Damage analysis method for laminates based on peridynamic theory .
|
[12] |
Peridynamics for fatigue life and residual strength prediction of composite laminate . |
[13] |
Peridynamic modeling of delamination growth in composite laminates .
|
[14] |
A constructive peridynamic kernel for elasticity .
|
[15] |
Meshless modeling framework for fiber reinforced concrete structures . |
[16] |
Peridynamics via finite element analysis . |
[17] |
Coupling of peridynamic theory and the finite element method .
|
[18] |
A coupling approach of discretized peridynamics with finite element method .
|
[19] |
An effective way to couple FEM meshes and peridynamics grids for the solution of static equilibrium problems .
|
[20] |
A force-based coupling scheme for peridynamics and classical elasticity .
|
[21] |
A morphing strategy to couple non-local to local continuum mechanics . |
[22] |
A morphing framework to couple non-local and local anisotropic continua . |
[23] |
A morphing approach to couple state-based peridynamics with classical continuum mechanics . |
[24] |
A coupled meshless finite point/Peridynamic method for 2D dynamic fracture analysis . |
[25] |
A coupling approach of state-based peridynamics with node-based smoothed finite element method . |
[26] |
The peridynamic formulation for transient heat conduction . |
[27] |
A peridynamic formulation for transient heat conduction in bodies with evolving discontinuities . |
[28] |
Peridynamic thermal diffusion . |
[29] |
Selecting the kernel in a peridynamic formulation: A study for transient heat diffusion . |
[30] |
近场动力学中内核参数对非均匀材料热传导数值解的影响研究 .Effects of kernel parameters of peridynamic theory on heat conduction numerical solution for non-homogeneous material . |
[31] |
Reformulation of elasticity theory for discontinuities and long-range forces . |
[32] |
A peridynamic model for dynamic fracture in functionally graded materials . |
[33] |
|
[34] |
|
[35] |
Studies of dynamic crack propagation and crack branching with peridynamics . |
/
〈 | 〉 |