Geometric Stiffness Effects in 2D and 3D Frames

Similar documents
Review of Strain Energy Methods and Introduction to Stiffness Matrix Methods of Structural Analysis

Finite Element Analysis Prof. Dr. B. N. Rao Department of Civil Engineering Indian Institute of Technology, Madras. Module - 01 Lecture - 13

Aircraft Structures Beams Torsion & Section Idealization

Module 4 : Deflection of Structures Lecture 4 : Strain Energy Method

General elastic beam with an elastic foundation

Lecture 7: The Beam Element Equations.

14. *14.8 CASTIGLIANO S THEOREM

External Work. When a force F undergoes a displacement dx in the same direction i as the force, the work done is

Chapter 2: Deflections of Structures

LATERAL BUCKLING ANALYSIS OF ANGLED FRAMES WITH THIN-WALLED I-BEAMS

10. Applications of 1-D Hermite elements

k 21 k 22 k 23 k 24 k 31 k 32 k 33 k 34 k 41 k 42 k 43 k 44

Quintic beam closed form matrices (revised 2/21, 2/23/12) General elastic beam with an elastic foundation

MAE 323: Lecture 1. Review

ENGN2340 Final Project: Implementation of a Euler-Bernuolli Beam Element Michael Monn

Mathematical Properties of Stiffness Matrices

2 marks Questions and Answers

Strain Energy in Linear Elastic Solids

Chapter 5 Structural Elements: The truss & beam elements

Module 3. Analysis of Statically Indeterminate Structures by the Displacement Method

BEAM DEFLECTION THE ELASTIC CURVE

Consider an elastic spring as shown in the Fig.2.4. When the spring is slowly

Basic Energy Principles in Stiffness Analysis

Section 6: PRISMATIC BEAMS. Beam Theory

FINAL EXAMINATION. (CE130-2 Mechanics of Materials)

MAE4700/5700 Finite Element Analysis for Mechanical and Aerospace Design

Axial force-moment interaction in the LARSA hysteretic beam element

BOUNDARY EFFECTS IN STEEL MOMENT CONNECTIONS

Chapter 5 Elastic Strain, Deflection, and Stability 1. Elastic Stress-Strain Relationship

TORSION INCLUDING WARPING OF OPEN SECTIONS (I, C, Z, T AND L SHAPES)

Part D: Frames and Plates

Finite Element Analysis Prof. Dr. B. N. Rao Department of Civil Engineering Indian Institute of Technology, Madras. Module - 01 Lecture - 11

Lecture 15 Strain and stress in beams

Mechanics PhD Preliminary Spring 2017

Survey of Wave Types and Characteristics

UNIT- I Thin plate theory, Structural Instability:

Mechanical Design in Optical Engineering

ME 323 Examination #2 April 11, 2018

torsion equations for lateral BucKling ns trahair research report r964 July 2016 issn school of civil engineering

Structural Dynamics Lecture Eleven: Dynamic Response of MDOF Systems: (Chapter 11) By: H. Ahmadian

Chapter 6 2D Elements Plate Elements

Aircraft Structures Structural & Loading Discontinuities

Lecture Slides. Chapter 4. Deflection and Stiffness. The McGraw-Hill Companies 2012

Institute of Structural Engineering Page 1. Method of Finite Elements I. Chapter 2. The Direct Stiffness Method. Method of Finite Elements I

Shafts: Torsion of Circular Shafts Reading: Crandall, Dahl and Lardner 6.2, 6.3

Calculus of the Elastic Properties of a Beam Cross-Section

Chapter 4 Deflection and Stiffness

Slender Structures Load carrying principles

3 Relation between complete and natural degrees of freedom

Stress and Strain ( , 3.14) MAE 316 Strength of Mechanical Components NC State University Department of Mechanical & Aerospace Engineering

Methods of Analysis. Force or Flexibility Method

Presented By: EAS 6939 Aerospace Structural Composites

Rigid and Braced Frames

7.5 Elastic Buckling Columns and Buckling

Structural Analysis III Compatibility of Displacements & Principle of Superposition

Assumptions: beam is initially straight, is elastically deformed by the loads, such that the slope and deflection of the elastic curve are

Indeterminate Analysis Force Method 1

Discontinuous Distributions in Mechanics of Materials

Aircraft Structures Kirchhoff-Love Plates

CHAPTER 5. Beam Theory

Ph.D. Preliminary Examination Analysis

Due Tuesday, September 21 st, 12:00 midnight

D : SOLID MECHANICS. Q. 1 Q. 9 carry one mark each.

Using the finite element method of structural analysis, determine displacements at nodes 1 and 2.

SLOPE-DEFLECTION METHOD

Example 3.7 Consider the undeformed configuration of a solid as shown in Figure 3.60.

D : SOLID MECHANICS. Q. 1 Q. 9 carry one mark each. Q.1 Find the force (in kn) in the member BH of the truss shown.

4.5 The framework element stiffness matrix

Unit 13 Review of Simple Beam Theory

Stability Analysis of Laminated Composite Thin-Walled Beam Structures

Physics 8 Monday, November 20, 2017

REVIEW FOR EXAM II. Dr. Ibrahim A. Assakkaf SPRING 2002

1-1 Locate the centroid of the plane area shown. 1-2 Determine the location of centroid of the composite area shown.

Unit - 7 Vibration of Continuous System

LECTURE 13 Strength of a Bar in Pure Bending

ON THE DESIGN CURVES FOR BUCKLING PROBLEMS

Conceptual question Conceptual question 12.2

Basic principles of steel structures. Dr. Xianzhong ZHAO

MECHANICS OF MATERIALS REVIEW

Module 3. Analysis of Statically Indeterminate Structures by the Displacement Method

Chapter 11. Displacement Method of Analysis Slope Deflection Method

A METHOD OF LOAD INCREMENTS FOR THE DETERMINATION OF SECOND-ORDER LIMIT LOAD AND COLLAPSE SAFETY OF REINFORCED CONCRETE FRAMED STRUCTURES

March 24, Chapter 4. Deflection and Stiffness. Dr. Mohammad Suliman Abuhaiba, PE

Local Maximums and Local Minimums of Functions. f(x, y) has a local minimum at the point (x 0, y 0 ) if f(x 0, y 0 ) f(x, y) for

MMJ1153 COMPUTATIONAL METHOD IN SOLID MECHANICS PRELIMINARIES TO FEM

A PROTOCOL FOR DETERMINATION OF THE ADHESIVE FRACTURE TOUGHNESS OF FLEXIBLE LAMINATES BY PEEL TESTING: FIXED ARM AND T-PEEL METHODS

FLEXIBILITY METHOD FOR INDETERMINATE FRAMES

Seismic Pushover Analysis Using AASHTO Guide Specifications for LRFD Seismic Bridge Design

3 2 6 Solve the initial value problem u ( t) 3. a- If A has eigenvalues λ =, λ = 1 and corresponding eigenvectors 1

Finite elements for plates and shells. Advanced Design for Mechanical System LEC 2008/11/04

Chapter 13 TORSION OF THIN-WALLED BARS WHICH HAVE THE CROSS SECTIONS PREVENTED FROM WARPING (Prevented or non-uniform torsion)

DESIGN OF FIXED CIRCULAR ARCHES WITH TUBE CROSS-SECTIONS UNDER CONCENTRATED LOADS ACCORDING TO EC3

CHAPTER 8. TRANSVERSE VIBRATIONS-IV: MULTI-DOFs ROTOR SYSTEMS

Eigenvalues of Trusses and Beams Using the Accurate Element Method

CHAPTER 14 BUCKLING ANALYSIS OF 1D AND 2D STRUCTURES

Lecture 8. Stress Strain in Multi-dimension

If the number of unknown reaction components are equal to the number of equations, the structure is known as statically determinate.

A.R. Tusnin, M. Prokic. Behavior of symmetric steel I-sections under combined bending and torsion actions in inelastic range

MITOCW MITRES2_002S10linear_lec07_300k-mp4

Stability Of Structures: Continuous Models

Transcription:

Geometric Stiffness Effects in D and 3D Frames CEE 41. Matrix Structural Analsis Department of Civil and Environmental Engineering Duke Universit Henri Gavin Fall, 1 In situations in which deformations are not infinitesimall 1 small, linear elastic analses ma not capture the true structural response. For large-strain problems, results of finite deformation analsis is significantl more accurate than linear, infinitesimal deformation analsis. Incorporating the details of finite deformation, the analsis ma also be applied to a buckling analsis of the structural sstem. In the derivation of the linear elastic stiffness matrix for frame elements, the potential energ function includes strain energ due to bending, axial and shear deformation effects. Axial effects are decoupled from shear and bending effects in the resulting linear elastic stiffness matrices. 3 In finite deformation analsis, on the other hand, the potential energ function includes additional terms, which accounts for the interaction between the axial load effects on the frame element and the lateral deformation of the frame element. These effects are often called P effects. We will separate the potential energ function U into an elastic part U E (which contains the infinitesimal strain energ) and a geometric part, U G (which includes the interaction of lateral deformations and axial loads). The linear elastic strain energ results in the same frame element stiffness matrices k E that we have found previousl. So, this document focuses onl on the geometric component of the potential energ function. From this geometric part of the potential energ, we will derive the geometric stiffness matrix k G. As in the finite deformation analsis of trusses, we need to know the deformation of the structure in order to find the internal axial loads, but we need to know the internal axial loads to determine the geometric stiffness matrix and the deformations. This chicken-andegg problem can be solved with the same tpe of Newton-Raphson iteration approach which we used previousl for finite deformation analsis of trusses. 1 infinitesimal means arbitraril close to zero, as in infinitesimal calculus. finite means neither infinite nor infinitesimal, as in a finite distance. 3 The rows and columns corresponding to axial effects (1st and 4th) have non-zero elements onl in the 1st and 4th columns and rows. Also, the rows and columns corresponding to bending and shear effects have non-zero elements onl in the nd, 3rd, 5th, and 6th columns and rows.

CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin 1 Deformed shape of a frame element in bending To start with, we need to introduce the deformed shape of a frame element. The deformed shape of a frame element, h(x), subjected to end-forces, q, is a cubic polnomial. A cubic polnomial ma be written in a power-polnomial form: h(x) = a + a 1 x + a x + a 3 x 3. (1) ikewise, the slope of the beam, h (x) ma be expressed h (x) = a 1 + a x + 3a 3 x. () u 6 u u3 u 5 h(x) x= x= x Figure 1. The deformed shape of a beam, h(x) is assumed to be a cubic function of x. The four polnomial coefficients, a,, a 3, are found b matching h(), h (), h() and h () to the specified end displacements and rotations of the beam. Assuming small rotations, tan θ θ, and neglecting shear deformation effects, h() = u a = u h () = u 3 a 1 = u 3 h() = u 5 a + a 1 + a + a 3 3 = u 5 h () = u 6 a 1 + a + 3a 3 = u 6 These four equations with four unknowns (a,, a 3 ) have the solution a = u (3) a 1 = u 3 (4) a = 3(u 5 u )/ (u 3 + u 6 )/ (5) a 3 = (u 5 u )/ 3 + (u 3 + u 6 )/. (6) CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 3 You should be able to confirm this solution for the polnomial coefficients. Note that the cubic deformation function h(x) ma also be written as a weighted sum of cubic polnomials. h(x) = u b (x) + u 3 b 3 (x) + u 5 b 5 (x) + u 6 b 6 (x), (7) The weights u i are simpl the set of local element displacements and the functions b i (x) are each cubic functions in x. These cubic shape functions represent beam deformations due to a unit applied displacement in the corresponding coordinate (onl). Neglecting shear deformation effects, the frame element shape functions are the Hermite polnomials, b (x) = 1 3 (x/) + (x/) 3 b () = 1 b () = b () = b () = b 3 (x) = x (1 x/) b 3 () = b 3() = 1 b 3 () = b 3() = b 5 (x) = 3 (x/) (x/) 3 b 5 () = b 5() = b 5 () = 1 b 5() = b 6 (x) = (x/) (x/ 1) b 6 () = b 6() = b 6 () = b 6() = 1 Equations (1) and (7) are two different was of expressing exactl the same equation, h(x). The finite element method makes use of polnomials in the form of equation (7). To complete the picture, for axial deformations, (which contribute to transverse deformations onl through the axial load in the geometric stiffness matrix), b 1 (x) = (1 x/) b 4 (x) = (x/). You should confirm that with the given definitions of b i (x), and the coefficients a i, that equations (1) and (7) are equivalent. We now relate this polnomial for the assumed deformation shape to a potential energ function. Recall the internal elastic strain energ of a beam due to bending effects, U B = 1 M (x) EI dx. (8) Since the curvature of the beam is M(x)/(EI) and assuming infinitesimal deformation, h (x) is practicall the same as the curvature, and U B = 1 M(x)h (x) dx (9) = 1 EI h (x) h (x) dx. (1) Equation (8) is an expression of the strain energ in terms of the internal bending moment; equation (9) is an expression of the strain energ in terms of the internal bending moment and the assumed cubic deformation function; and equation (1) is an expression of the strain energ in terms of the assumed cubic deformation function onl. In the finite element method it is common to express the potential energ using forms like equation (1). If the internal bending moment M(x) does indeed generate the assumed deformation function h(x) then all three forms of the elastic strain energ are exactl equivalent and completel interchangeable. CC BY-NC-ND December 8, 17, H.P. Gavin

4 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin Power polnomial basis Hermite polnomial basis 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ 1 x/ CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 5 Finite deformation effects in transversel-displaced frame elements Frame elements carring large axial loads or undergoing large displacements have nonlinear behavior arising from the internal moments that are the product of the axial loads P and the displacements transverse to the loads,. This kind of nonlinearit is sometimes called the P effect. For frame elements with transverse displacements h(x), but without axial deformation, the frame element has displacements u(x) along the length of the element. Figure shows how transverse displacements h(x) relate to axial displacement u(x). These can be simpl analzed as: (dx) = (dx du) + (dh) (dx) = (dx) (du)(dx) + (du) + (dh) (du)(dx) = (du) + (dh) du dx = 1 ( ) du + 1 ( ) dh dx dx For frame elements in which axial strain is much less than transverse rotations, which is alwas the case in elastic frame elements made from common structural materials, (u ) (h ) and u (x) (1/)(h (x)). The geometric deformation increases with the square of the rotation of the element. N(x) h(x) dx h(x) + dh du N(x) dx h (x) dx dh du dx Figure. Axial end displacements due to transverse displacements, without axial deformation. To determine the strain energ associated with axial forces N(x) and geometric deformation u (x) = (1/)(h (x)), recall that the work of a constant force F moving through a displacement d, is simpl F d. Since N(x) does not increase linearl with h(x), the associated internal strain energ of a force N(x) distributed along the length of a frame element and geometric deformation u (x) = (1/)(h (x)) associated with transverse displacements h(x) U G = N(x) du(x) dx dx = 1 N(x) ( ) dh(x) dx. (11) dx CC BY-NC-ND December 8, 17, H.P. Gavin

6 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin 3 Geometric stiffness of frame elements The previous section shows that the potential energ due to axial loads, N(x) and transverse displacements, h(x), is U G = 1 N(x) ( ) dh dx. (1) dx If the axial load N(x) is constant over the length of the beam, N(x) = T, and U G = 1 T (h (x)) dx. (13) Substituting the beam shape functions, equation (7), and carring out the integral leads to the potential energ function in terms of transverse end displacements, u and u 5, and end rotations, u 3 and u 6, U G = T 3 ( u 3u 6 3u 5 u 3 3u 5 u 6 + 3u u 3 + 3u u 6 +18u 5 36u 5 u + 18u + u 3 + u 6). (14) Invoking Castigliano s theorem, the partial derivative of the potential energ function with respect to a displacement coordinate is the force in the direction of that displacement coordinate. The end forces due to geometric stiffness effects can then be found as follows: q = U G u = T 3 (36u + 3u 3 36u 5 + 3u 6 ) q 3 = U G u 3 = T 3 (3u + 4 u 3 3u 5 u 6 ) q 5 = U G u 5 = T 3 ( 36u 3u 3 + 36u 5 3u 6 ) q 6 = U G u 6 = T 3 (3u u 3 3u 5 + 4 u 6 ). Writing these expressions in matrix form, we arrive at the geometric stiffness matrix for a frame element: q 1 q q 3 q 4 q 5 q 6 = T 6 5 1 1 6 1 5 15 1 3 6 5 1 6 5 1 1 3 1 15 u 1 u u 3 u 4 u 5 u 6, (15) CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 7 where the tension in the beam is given b T = EA(u 4 u 1 )/. The geometric stiffness matrix for a D (planar) frame element in local coordinates is: 6 6 k G = T 5 1 5 1 1 15 1 3 6 5 1 6 1 5 3 1 1 15. (16) The coordinate transformation process is identical to the process carried out before for the elastic element stiffness matrix. The coordinate transformation matrix, T, is T = c s s c 1 c s s c 1, (17) where s and c are the sine and cosine of the counter-clockwise angle from global element coordinate number 1 to the frame element. Here we are making the approximation that the deformed inclination of the frame element is approximatel the same as the original inclination of the frame element. The element stiffness matrix in global coordinates is found b appling the coordinate transformation matrix. 6 6 s sc s 6 5 5 1 5 s 6sc s 5 1 6 K G = T T k G T = T sc 6 5 5 c c 6 sc 6 1 5 5 c c 1 s c s c 1 1 15 1 1 3 6 5 s 6sc s 6 6 s sc s 5 1 5 5 1 6 sc 6 c c 6 sc 6 c c 5 5 1 5 5 1 s c 1 1 3 s c 1 1 15 It is not hard to confirm this expression for K G, and ou should feel encouraged to do so. The assembl of the structural stiffness matrix K s with elastic and geometric effects proceeds exactl as with the elastic stiffness matrix. (18) CC BY-NC-ND December 8, 17, H.P. Gavin

8 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin 4 Derivation of stiffness coefficients directl from the strain energ function It is common to derive the coefficients of a stiffness matrix directl from the strain energ function. Note that the i, j component of the stiffness matrix is k ij = u j q i = k ji = and that the i th component of the end force, q i, is q i = u i U. Therefore, the stiffness coefficients ma be written k ij = U u i u j. u i q j, If the stiffness matrix to be determined is for bending effects onl, then, as seen before, U = U B = 1 EI h (x) h (x) dx. Now, since integration and differentiation are both linear operations, it does not matter which is done first, integration or differentiation. Therefore, the stiffness coefficient ma be written, k Eij = 1 EI h (x) u i h (x) u j dx. The elastic stiffness matrix incorporating bending effects onl ma be determined directl from this expression. ikewise, the geometric stiffness matrix ma be determined directl from k Gij = 1 T h (x) u i h (x) u j dx. The coefficients for the elastic stiffness matrix and the geometric stiffness matrix for frame elements in three dimensions are derived using this method in the remaining sections of this document. CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 9 5 Cubic shape functions for beams including shear deformations Consider the twelve local coordinates of a three dimensional frame element. The transverse Figure 3. The twelve local coordinates of a three-dimensional frame element. displacements in the local x plane, h (x), of an elastic beam ma be separated into a shear-related component, h s (x) and a bending-related component h b (x), h (x) = h s (x) + h b (x). (19) The shear force at the end of the beam in the local direction, q ma be found in terms of the beam end-displacements in the local direction and the end-rotations about the local z axis, q = k u + k 6 u 6 + k 8 u 8 + k 1 u 1. The effective shear strain is simpl h s(x) = q GA s = 1 GA s (k u + k 6 u 6 + k 8 u 8 + k 1 u 1 ). () The internal bending moment, M z (x), due to the effects of the end displacements and end rotations is simpl M z (x) = q x q 6, and the curvature is approximatel h b(x) = 1 EI z [ (k u + k 6 u 6 + k 8 u 8 + k 1 u 1 ) x (k 6 u + k 66 u 6 + k 68 u 8 + k 6 1 u 1 ) ]. (1) CC BY-NC-ND December 8, 17, H.P. Gavin

1 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin in which small angles are assumed. B computing the potential energ function for shear and bending deformations, U = 1 EI z (h b(x)) dx + 1 GA s (h s(x)) dx, and taking the partial derivatives of the potential energ function with respect to the displacements coordinates u, u 6, u 8 and u 1, rows, 6, 8, and 1 of the elastic stiffness matrix ma be computed. The shape of the deformed beam ma therefore b found b integrating equations () and (1) to obtain h s (x) and h b (x) and b solving for the constants of integration using the end conditions. So doing, h (x) = h s(x) + h b(x) = 1 q + 1 ( ) 1 GA s EI z q x q 6 x + C 1. Inserting the end condition h () = u 6 the constant of integration, C 1, is EI z u 6. Integrating again, h (x) = h s (x) + h b (x) = 1 q x + 1 ( 1 GA s EI z 6 q x 3 1 ) q 6x + u 6 x + C. Inserting the boundar condition h() = u, and noting that 1/(GA s ) = Φ /(1EI z ), the deformed shape of the beam becomes h (x) = 1 ( 1 EI z 6 q x 3 1 q 6 x 1 1 Φ q x) + u 6 x + u. The transverse deformation shape function for a beam with bending and shear deformation in the x plane. h (x) = 1 3 1 1+Φ { [x 3 3x Φ x + 3 (1 + Φ )] u + [x 3 ( + Φ /)x + 3 (1 + Φ /)x] u 6 + [ x 3 + 3x + Φ x] u 8 + [x 3 (1 Φ /)x Φ 3 x/] u 1 }. () For bending and shear deformations in the x z plane, the shape function ma be found using an analogous method, while respecting the right-hand coordinate sstem. The transverse deflection, h z (x) will consist of shear and bending components, h z (x) = h s (x) + h b (x). The end-shear force and the end-bending moment arise for end displacements in the local z direction and end moments about the local axis, q 3 = k 33 u 3 + k 35 u 5 + k 39 u 9 + k 3 11 u 11, CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 11 and The effective shear strain is q 5 = k 53 u 3 + k 55 u 5 + k 59 u 9 + k 5 11 u 11. h s(x) = q 3 GA sz = 1 GA sz ( k 33 u 3 + k 35 u 5 + k 39 u 9 + k 3 11 u 11 ), and the bending curvature is approximatel h b(x) = 1 EI (q 3 x + q 5 ) = 1 EI [ (k 33 u 3 + k 35 u 5 + k 39 u 9 + k 3 11 u 11 ) x +(k 53 u 3 + k 55 u 5 + k 59 u 9 + k 5 11 u 11 ) ], where small angles are again assumed. Integrating h b(x) and combining with h s(x), h z(x) = 1 q 3 + 1 ( ) 1 GA sz EI q 3 x + q 5 x + C 1. Inserting the end condition, h z() = u 5 gives C 1 = EI u 5. Integrating again, h z (x) = 1 q 3 x + 1 ( 1 GA sz EI 6 q 3 x 3 + 1 ) q 5 x u 5 x + C. Now inserting the end condition h z () = u 3 gives C = u 3 and noting that 1/(GA sz ) = Φ z /(1EI ) the deformed shape ma be written h z (x) = 1 ( 1 EI 6 q 3 x 3 + 1 q 5 x 1 1 Φ z q 3 x) u 5 x + u 3. Finall, the shape function for a frame element bending and shear in the local x z plane is h z (x) = 1 1 3 1+Φ z { [x 3 3x Φ z x + 3 (1 + Φ z )] u 3 + [ x 3 + ( + Φ z /)x 3 (1 + Φ z /)x] u 5 + [ x 3 + 3x + Φ z x] u 9 + [ x 3 + (1 Φ z /)x + Φ z 3 x/] u 11 }. (3) Note that this expression is equivalent to equation () except for the fact that Φ z replaces Φ and that the signs of the u 5 and u 6 shape functions are reversed, as are the signs of the u 11 and u 1 shape functions. This is consistent with the right-hand coordinate sstem. For axial displacements, the shape function is the same as for a truss, ( h x (x) = 1 x ) u 1 + x u 7. (4) ikewise, for torsional displacements, the shape function is analogous to the axial displacement shape function ( h θx (x) = 1 x ) u 4 + x u 1. (5) CC BY-NC-ND December 8, 17, H.P. Gavin

1 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin 6 The 3D elastic stiffness matrix for frame elements including shear and bending effects Differentiating and integrating the shape functions derived above, the three-dimensional elastic stiffness matrix for frame elements in local coordinates including bending and shear deformation effects is: k E = EA 1EI z 6EI 3 (1+Φ ) z 1EI 6EI 3 (1+Φ z) (1+Φ z) GJ 6EI (4+Φ z)ei (1+Φ z) (1+Φ ) (1+Φ z) (4+Φ )EI z (1+Φ ) 6EI z (1+Φ ) EA 1EI z 6EI 3 (1+Φ ) z 1EI 6EI 3 (1+Φ z) (1+Φ z) GJ 6EI ( Φ z)ei (1+Φ z) 6EI z (1+Φ ) (1+Φ ) (1+Φ z) ( Φ )EI z (1+Φ ) EA 1EI z 6EI 3 (1+Φ ) z 1EI 6EI 3 (1+Φ z) (1+Φ z) GJ 6EI ( Φ z)ei (1+Φ z) (1+Φ ) (1+Φ z) ( Φ )EI z (1+Φ ) 6EI z (1+Φ ) EA 1EI z 6EI 3 (1+Φ ) z 1EI 6EI 3 (1+Φ z) (1+Φ z) GJ 6EI (4+Φ z)ei (1+Φ z) 6EI z (1+Φ ) (1+Φ ) (1+Φ z) (4+Φ )EI z (1+Φ ), where Φ = 1EI z GA s, and Φ z = 1EI GA sz. CC BY-NC-ND December 8, 17, H.P. Gavin

Geometric Stiffness Effects in D and 3D Frames 13 7 Formulation of the geometric stiffness matrix from the cubic shape functions For the elements of the geometric stiffness matrix in rows,6,8, and 1, the potential energ function is U G = 1 T h (x) h (x) dx, where h (x) is given b equation (). ikewise, for rows 3,5,9, and 11, the potential energ function is U Gz = 1 T h z(x) h z(x) dx, where h z (x) is given b equation (3). For rows 1 and 7, the potential energ function is where h x (x) is given b equation (4). U Gx = 1 T h x(x) h x(x) dx, 8 Torsion In the torsion of non-circular sections, torsional displacements result in axial deformation (warping) of the cross-section. In such cases, the work of the axial tension, T, moving through the warping displacements provides the potential energ function for rows 4 and 1, U Gθ = 1 T J x h A θx(x) h θx(x) dx, x where h θx (x) is given b equation (5) and J x is the torsional moment of inertia. The geometric stiffness coefficients ma then be found b forming the Hessian of the appropriate potential energ function, k Gij = U G u i u j. CC BY-NC-ND December 8, 17, H.P. Gavin

14 CEE 41. Matrix Structural Analsis Duke Universit Fall 1 H.P. Gavin 9 The 3D geometric stiffness matrix for frame elements including shear and bending effects Differentiating and integrating the shape functions as described above, the three-dimensional geometric stiffness matrix for frame elements in local coordinates including axial, bending, shear and torsional warping effects is: k G = T 6/5+Φ +Φ (1+Φ ) 6/5+Φ z+φ z (1+Φ z) /1 (1+Φ ) /1 (1+Φ z) J x A x /15+ Φ z/6+ Φ z /1 /1 /1 (1+Φ z) (1+Φ z) /15+ Φ /6+ Φ /1 (1+Φ ) (1+Φ ) 6/5 Φ Φ (1+Φ ) 6/5 Φ z Φ z (1+Φ z) /1 (1+Φ ) /1 (1+Φ z) Jx A x /3 Φ z/6 Φ z /1 /1 (1+Φ z) /1 (1+Φ ) (1+Φ z) /3 Φ /6 Φ /1 (1+Φ ) 6/5 Φ Φ (1+Φ ) 6/5 Φ z Φ z (1+Φ z) /1 (1+Φ ) /1 (1+Φ z) Jx A x /3 Φ z/6 Φ z /1 /1 /1 (1+Φ z) (1+Φ z) /3 Φ /6 Φ /1 (1+Φ ) (1+Φ ) 6/5+Φ +Φ (1+Φ ) 6/5+Φ z+φ z (1+Φ z) /1 (1+Φ ) /1 (1+Φ z) J x A x /15+ Φ z/6+ Φ z /1 /1 (1+Φ z) /1 (1+Φ ) (1+Φ z) /15+ Φ /6+ Φ /1 (1+Φ ), where T = EA(u 7 u 1 )/, Φ = 1EI z GA s and Φ z = 1EI GA sz. CC BY-NC-ND December 8, 17, H.P. Gavin