Exact static element stiffness matrices of non‐symmetric thin‐walled curved beams

An evaluation procedure of exact static stiffness matrices for curved beams with non‐symmetric thin‐walled cross section are rigorously presented for the static analysis. Higher‐order differential equations for a uniform curved beam element are first transformed into a set of the first‐order simultaneous ordinary differential equations by introducing 14 displacement parameters where displacement modes corresponding to zero eigenvalues are suitably taken into account. This numerical technique is then accomplished via a generalized linear eigenvalue problem with non‐symmetric matrices. Next, the displacement functions of displacement parameters are exactly calculated by determining general solutions of simultaneous non‐homogeneous differential equations. Finally an exact stiffness matrix is evaluated using force–deformation relationships. In order to demonstrate the validity and effectiveness of this method, displacements and normal stresses of cantilever thin‐walled curved beams subjected to tip loads are evaluated and compared with those by thin‐walled curved beam elements as well as shell elements. Copyright © 2004 John Wiley & Sons, Ltd.     

On curved finite elements for the analysis of circular arches

This paper proposes a straightforward criterion to warrant the displacement functions being used in the finite element approximation of circular arches. The criterion was established by studying the natural shape function, i.e. the exact solution of the deformed shape, of the circular arch element. The exact stiffness matrix [ K ]_exact is derived from the natural shape and is confirmed to be the inverse of the well-known flexibility matrix [ F ]_exact in the curved beam theory. The present paper compares the inverse [ K ]^?1 of the stiffness matrix derived from the assumed displacement function with the [ F ]_exact. It is shown that the procedure also guarantees the implicit inclusion of rigid-body modes in the pertinent stiffness matrix [ K ]. Case studies on typical approximate displacement functions assure the appropriateness as well as the ease of application of the proposed method.     

A spatially curved‐beam element with warping and Wagner effects

This paper presents a new spatially curved‐beam element with warping and Wagner effects that can be used for the non‐linear large displacement analysis of members that are curved in space. The non‐linear behaviour of members curved in space shows that the Wagner effects are substantial in the large twist rotation analysis. Most existing finite beam element models, such as ABAQUS and ANSYS cannot predict the non‐linear large displacement response of members curved in space correctly because the Wagner effects, viz. the Wagner moment and the corresponding finite strain terms, have not been considered in these finite beam elements. As a consequence, these finite beam elements do not provide correct predictions for the out‐of‐plane buckling and postbuckling behaviour of arches as well. In this paper, the symmetric tangent stiffness matrix has been derived based on the finite rotations parameterized by the conventional displacements. The warping and Wagner effects: both the Wagner moment and the corresponding finite strain terms and their constitutive relationship, are included in the spatially curved‐beam element. Two components of the initial curvature, the initial twist and their interactions with the displacements are also considered in the spatially curved‐beam element. This ensures that the large twist rotation analysis for the members curved in space is accurate. Comparisons with existing experimental, analytical and numerical results show that the spatially curved‐beam element is accurate and efficient for the non‐linear elastic analysis of curved members, buckling and postbuckling analysis of arches, and in its ability to predict large deflections and twist rotations in more arbitrarily curved members. Copyright © 2005 John Wiley & Sons, Ltd.     

A new approach for displacement functions of a curved Timoshenko beam element in motions normal to its initial plane

In existing literature, either analytical methods or numerical methods, the formulations for free vibration analysis of circularly curved beams normal to its initial plane are somewhat complicated, particularly if the effects of both shear deformation (SD) and rotary inertia (RI) are considered. It is hoped that the simple approach presented in this paper may improve the above‐mentioned drawback of the existing techniques. First, the three functions for axial (or normal to plane) displacement and rotational angles about radial and circumferential (or tangential) axes of a curved beam element were assumed. Since each function consists of six integration constants, one has 18 unknown constants for the three assumed displacement functions. Next, from the last three displacement functions, the three force–displacement differential equations and the three static equilibrium equations for the arc element, one obtained three polynomial expressions. Equating to zero the coefficients of the terms in each of the last three expressions, respectively, one obtained 17 simultaneous equations as functions of the 18 unknown constants. Excluding the five dependent ones among the last 17 equations, one obtained 12 independent simultaneous equations. Solving the last 12 independent equations, one obtained a unique solution in terms of six unknown constants. Finally, imposing the six boundary conditions at the two ends of an arc element, one determined the last six unknown constants and completely defined the three displacement functions. By means of the last displacement functions, one may calculate the shape functions, stiffness matrix, mass matrix and external loading vector for each arc element and then perform the free and forced vibration analyses of the entire curved beam. Good agreement between the results of this paper and those of the existing literature confirms the reliability of the presented theory. Copyright © 2005 John Wiley & Sons, Ltd.     

一种新的集成非线性杆件单元刚度矩阵的方法

对于非线性杆件单元,本文提出一种新的简便有效的集成单元刚度矩阵的算法。该方法直接从结构力学中的位移法的概念出发,通过解析积分或数值积分求解积分算子,由积分算子线性组合,能快速求解考虑弯、剪、扭、轴压等各种非线形刚度的杆件单元的刚度矩阵。该方法具有广泛的普适性,能适用于所有多项式、插值多项式、解析式、离散点描述的变刚度、变截面直杆的单元刚度矩阵集成计算。文中通过求解线性直杆单元刚度矩阵验证了该方法的正确性。     

Free vibrations of shear‐flexible and compressible arches by FEM

The purpose of this paper is to analyse free vibrations of arches with influence of shear and axial forces taken into account. Arches with various depth of cross‐section and various types of supports are considered. In the calculations, the curved finite element elaborated by the authors is adopted. It is the plane two‐node, six‐degree‐of‐freedom arch element with constant curvature. Its application to the static analysis yields the exact results, coinciding with the analytical ones. This feature results from the use of the exact shape functions in derivation of the element stiffness matrix. In the free vibration analysis the consistent mass matrix is used. It is obtained on the base of the same functions. Their coefficients contain the influences of shear flexibility and compressibility of the arch. The numerical results are compared with the results obtained for the simple diagonal mass matrix representing the lumped mass model. The natural frequencies are also compared with the ones for the continuous arches for which the analytically determined frequencies are known. The advantage of the paper is a thorough analysis of selected examples, where the influences of shear forces, axial forces as well as the rotary and tangential inertia on the natural frequencies are examined. Copyright © 2001 John Wiley & Sons, Ltd.     

Explicit free‐floating beam element

A two‐node free‐floating beam element capable of undergoing arbitrary large displacements and finite rotations is presented in explicit form. The configuration of the beam in three‐dimensional space is represented by the global components of the position of the beam nodes and an associated set of convected base vectors (directors). The local constitutive stiffness is derived from the complementary energy of a set of six independent deformation modes, each corresponding to an equilibrium state of constant internal force or moment. The deformation modes are characterized by generalized strains, formed via scalar products of the element related vectors. This leads to a homogeneous quadratic strain definition in terms of the generalized displacements, whereby the elastic energy becomes at most bi‐quadratic. Additionally, the use of independent equilibrium modes to set up the element stiffness avoids interpolation of kinematic variables, resulting in a locking‐free formulation in terms of three explicit matrices. A set of classic benchmark examples illustrates excellent performance of the explicit beam element. Copyright © 2014 John Wiley & Sons, Ltd.     

STATIC AND DYNAMIC APPLICATIONS OF A FIVE NODED HORIZONTALLY CURVED BEAM ELEMENT WITH SHEAR DEFORMATION

A five-noded thirteen DOF horizontally curved beam element with or without an elastic base is presented. One set of fourth-degree Lagrangian polynomials in natural co-ordinates is used for interpolation of beam geometry and vertical displacement while the angles of transverse rotation and twist are interpolated by another set of third-degree polynomials. For elastic subgrade, the reactive forces offered at any point are assumed to be proportional to the corresponding displacements at that point. The effect of shear deformation is accounted for in the stiffness matrix. In mass matrix evaluation, for dynamic problems, translational as well as rotary intertias have been considered and studied separately. For numerical integration of the stiffness matrix, a four-point Gaussian scheme has been found to be adequate. Numerical results for a number of sample problems and their comparison with analytical solutions have been presented for circular as well as for non-circular curved beams. Displacements, bending moment and torque for static loading with or without elastic foundation, as well as natural frequencies and mode shapes are computed for different cases. Examples include the problem of a cantilever beam of spiral geometry with different parametric values of the spiral and the agreement with the analytical results establishes the efficacy of the element. The performance of the element has been found be be excellent in both static and dynamic conditions. Sufficient details are presented so that the formulation may be readily used. It is hoped that the large number of numerical illustrations will elucidate the validity and the range of applicability of the element and will also serve as benchmark for future researchers. © 1997 by John Wiley & Sons, Ltd.     

Hierarchical universal matrices for triangular finite elements with varying material properties and curved boundaries

The integration required to find the stiffness matrix for a triangular finite element is inexpensive if the polynomial order of the element is low. Higher‐order elements can be handled efficiently by universal matrices provided they are straight‐edged and the material properties are uniform. For curved elements and elements with varying material properties (e.g. non‐linear B–H curves), Gaussian integration is generally used, but becomes expensive for high orders. Two new methods are proposed in which the high‐order part of the integrand is integrated exactly and the results stored in pre‐computed universal matrices. The effect of curved edges and varying material properties is approximated via interpolation. The storage requirement of the procedure is kept to a minimum by using specifically devised basis functions which are hierarchical and possess the three‐fold symmetry of a triangular element. Care has been taken to maintain the conditioning of the basis. One of the new methods is hierarchical in nature and suitable for use in an adaptive integration scheme. Results show that, for a given required accuracy, the new approaches are more efficient than Gauss quadrature for element orders of 4 or greater. The computational advantage increases rapidly with increasing order. Copyright © 1999 John Wiley & Sons, Ltd.     

AN EFFICIENT CURVED BEAM FINITE ELEMENT

The plane two-node curved beam finite element with six degrees of freedom is considered. Knowing the set of 18 exact shape functions their approximation is derived using the expansion of the trigonometric functions in the power series. Unlike the ones commonly used in the FEM analysis the functions suggested by the authors have the coefficients dependent on the geometrical and physical properties of the element. From the strain energy formula the stiffness matrix of the element is determined. It is very simple and can be split into components responsible for bending, shear and axial forces influences on the displacements. The proposed element is totally free of the shear and membrane locking effects. It can be referred to the shear-flexible (parameter d) and compressible (parameter e) systems. Neglecting d or e yields the finite elements in all necessary combinations, i.e. curved Euler–Bernoulli beam or curved Timoshenko beam with or without the membrane effect. Applying the elaborated element in the calculations a very good convergence to the analytical results can be obtained even with a very coarse mesh without the commonly adopted corrections as reduced or selective integration or introduction of the stabilization matrices, additional constraints, etc., for the small depth–length ratio. © 1997 John Wiley & Sons, Ltd.     

单轴对称截面圆弧拱平面外稳定性研究

根据变分原理推导了任意开口薄壁截面曲梁的稳定平衡方程。单轴对称截面圆弧拱在均布径向荷载(均匀受压)或两端作用大小相等、方向相反的端弯矩(均匀弯曲)作用下,平衡方程中曲率平面内的变形和曲率平面外的变形相互独立,故要么发生曲率平面内弯曲失稳,要么发生曲率平面外的弯扭失稳。给出单轴对称截面圆弧拱在这两种受力情况下平面外屈曲荷载的理论解答。通过一些无碍结果的近似使所得公式形式简洁,便于在工程中应用。最后给出了计算实例,与已有的文献进行比较,并使用通用有限元软件ANSYS 进行了模拟,分析结果与该文计算结果吻合,证明了所得公式的正确性。     

含裂纹损伤杆系结构的动态特性研究

运用动刚度有限元法,研究了含裂纹损伤杆系结构的动态特性。提出了一种含裂纹的杆单元,基于断裂力学的线弹簧模型,导出了相应的动刚度矩阵。在此基础上,对含裂纹的悬臂梁和平面框架进行了数值计算,并与已有的实验值和解析解进行了比较。结果表明:损伤位置和损伤程度的不同均会导致结构动态特性发生改变,因而在结构分析中应考虑损伤的影响;而该单元能够方便地用于含裂纹损伤杆系结构的动态特性分析,并具有很好的精度。     

基于场一致性的可带铰空间梁元几何非线性分析

虽然关于几何非线性分析的空间梁单元研究成果较多,但这些单元均是基于几何一致性得到的单元刚度矩阵,而基于场一致性的单元研究则较少,该文基于局部坐标系(随转坐标系)下扣除结构位移中的刚体位移得到的结构变形与结构坐标系下的总位移的关系,直接利用微分方法导出两者增量位移之间的关系,再基于场一致性原则,最终获得空间梁单元在大转动、小应变条件下的几何非线性单元切线刚度矩阵,在此基础上根据带铰梁端受力特征,导出了能考虑梁端带铰的单元切线刚度矩阵表达式,利用该文的研究成果编制了程序,对多个梁端带铰和不带铰的算例进行了空间几何非线性分析,计算结果表明这种非线性单元列式的正确性,实用价值较强。     

梁杆结构二阶效应分析的一种新型梁单元

推导了一种计及梁杆二阶效应的新型两结点梁单元。首先依据插值理论构造了三结点Euler-Bernoulli梁单元的位移场:使用五次Hermite插值函数建立梁单元的侧向位移场,二次Lagrange插值函数建立梁单元的轴向位移场,进而由非线性有限元理论推导了单元的线性刚度矩阵和几何刚度矩阵,然后使用静力凝聚方法消除三结点梁单元中间结点的自由度,从而得到一种考虑轴力效应的新型两结点梁单元。实例分析表明,此新型梁单元具有很高的计算精度,使用此单元进行梁杆结构分析可获得相当准确的二阶位移和内力。     

同时考虑切、法向抗力的弹性地基园拱单元

本文从控制微分方程以及内力与位移之间的关系出发,同时考虑切向和法向地基抗力的影响,推导了弹性地基园拱的6×6阶精确单元刚度矩阵。该单元可以用于断面含有园弧段的地下结构物分析。     

三维杆系结构的几何非线性有限元分析

为了更准确地描述杆系结构的几何非线性性能,建立了一种基于三维梁单元有限元分析的计算方法。引入了考虑两方向曲率和扭转角变化的坐标转换矩阵来描述任意增量下的单元平移和转动;采用了包括轴向变形和扭转的非线性项的刚度矩阵来考虑高阶非线性项的影响。应用广义位移控制法进行增量迭代,编制了相应的三维梁单元非线性计算程序NL_Beam3D。通过对几个例子进行的分析,验证了该方法可较好地考虑结构几何非线性。     

计及二阶效应的梁杆系统动力响应分析方法研究

基于动力刚度法和有限元理论提出了一种考虑二阶效应计算梁杆动力响应的新方法。通过求解轴向力作用下Bernoulli-Euler 梁横向和轴向挠度自由振动微分方程,利用位移边界条件反解出待定系数,得到了动态精确形函数;使用经典有限元方法推导了考虑截面自身旋转惯量的质量阵和考虑二阶效应的刚度阵,该质量阵和刚度阵各元素均为轴力和圆频率的超越函数;建立了杆系结构瞬态动力学分析的动力平衡方程,给出了稳定和高效的求解方案。对几个典型的算例进行了计算分析,并与通用软件ANSYS 的计算结果进行了比较。计算结果表明:该分析梁杆系统动力响应的新方法具有较高的计算精度和效率,特别是能够准确地计入轴力对于梁杆动力响应的影响。     

Use of equivalent mass method for free vibration analyses of a beam carrying multiple two‐dof spring–mass systems with inertia effect of the helical springs considered

This paper investigates the free vibration characteristics of a beam carrying multiple two‐degree‐of‐freedom (two‐dof) spring–mass systems (i.e. the loaded beam). Unlike the existing literature to neglect the inertia effect of the helical springs of each spring–mass system, this paper takes the last inertia effect into consideration. To this end, a technique to replace each two‐dof spring–mass system by a set of rigidly attached equivalent masses is presented, so that the free vibration characteristics of a loaded beam can be predicted from those of the same beam carrying multiple rigidly attached equivalent masses. In which, the equation of motion of the loaded beam is derived analytically by means of the expansion theorem (or the mode superposition method) incorporated with the natural frequencies and the mode shapes of the bare beam (i.e. the beam carrying nothing). In addition, the mass and stiffness matrices including the inertia effect of the helical springs of a two‐dof spring–mass system, required by the conventional finite element method (FEM), are also derived. All the numerical results obtained from the presented equivalent mass method (EMM) are compared with those obtained from FEM and satisfactory agreement is achieved. Because the equivalent masses of each two‐dof spring–mass system are dependent on the magnitudes of its lumped mass, spring constant and spring mass, the presented EMM provides an effective technique for evaluating the overall inertia effect of the two‐dof spring–mass systems attached to the beam. Furthermore, if the total number of two‐dof spring–mass systems attached to the beam is large, then the order of the overall property matrices for the equation of motion of the loaded beam in EMM is much less than that in FEM and the computer storage memory required by the former is also much less than that required by the latter. Copyright © 2005 John Wiley & Sons, Ltd.     

Higher-order inverse mass matrices for the explicit transient analysis of heterogeneous solids

New methods are presented for the direct computation of higher-order inverse mass matrices (also called reciprocal mass matrices) that are used for explicit transient finite element analysis. The motivation of this work lies in the need of having appropriate sparse inverse mass matrices, which present the same structure as the consistent mass matrix, preserve the total mass, predict suitable frequency spectrum and dictate sufficiently large critical time step sizes. For an efficient evaluation of the reciprocal mass matrix, the projection matrix should be diagonal. This condition can be satisfied by adopting dual shape functions for the momentum field, generated from the same shape functions used for the displacement field. A theoretically consistent derivation of the inverse mass matrix is based on the three-field Hamilton principle and requires the projection matrix to be evaluated from the integral of these shape functions. Unfortunately, for higher-order FE shape functions and serendipity FE elements, the projection matrix is not positive definitive and can not be employed. Therefore, we study several lumping procedures for higher order reciprocal mass matrices considering their effect on total-mass preserving, frequency spectra and accuracy in explicit transient simulations. The article closes with several numerical examples showing suitability of the direct inverse mass matrix in dynamics.     

On artificial strain of thin curved elements

This paper presents an attempt to clarify further the nature of the overstiffness of thin curved structural elements. It follows the approach presented by Ashwell and co-workers^1,2 who based their study on shape functions of arch elements. Such elements display the same stiff behaviour as complex shell elements but are simpler to study and understand. An examination of several discretizations for a two-dimensional curved beam-column element for linear elastic analysis of arches is performed. This examination and the numerical results obtained from these approximations provide a new interpretation of the artificial strains based on the mapping variables of shape functions. Two new methods to reduce the artificial stiffness of the element are proposed.