Indiacon CivilSubscribe

Textbook › Structural Analysis › Chapter 7 › Lesson 7.1

7.1 Flexibility matrix method

Structural Analysis Chapter 7: Matrix methods Under review — not yet three-checked Author: Indiacon

Learning objectives

After this lesson you should be able to:

  • Explain the force (flexibility) method in matrix form: choose redundants, release them, and write compatibility as [F]{X} = {D} − {D_L}.
  • Build flexibility coefficients from deflection formulas or from the unit-load method, and assemble a structure flexibility matrix from member flexibility matrices.
  • Solve beams with two redundants, including a support settlement, and check the answer by equilibrium and by the symmetry of the flexibility matrix.

Theory

Chapter 3 solved indeterminate structures by consistent deformations: remove enough restraints to make the structure statically determinate, find the gaps that the loads open at the removed restraints, and find the redundant forces that close those gaps again. With one redundant this is one equation. With two or more redundants the equations are coupled, because each redundant moves every released point. Writing the equations as a matrix keeps the bookkeeping clean and lets a computer do the solving. Sign convention for this chapter (the usual matrix convention): global x to the right and y upward; displacements and forces are positive along these axes; rotations and moments are positive anticlockwise, both for joints and for member-end moments. Bending moment diagrams are still drawn with sagging positive. Chapter 4 used clockwise-positive member-end moments, so a member-end moment found here has the opposite sign to the same moment written in Chapter 4; the physical answer is the same.

The flexibility coefficient f_ij is the displacement at released point i caused by a unit value of redundant j acting alone on the released (primary) structure. Collected together they form the flexibility matrix [F]. The vector {D_L} holds the displacements at the released points caused by the applied loads on the primary structure, and {D} holds the displacements that actually happen there in the real structure (zero at an unyielding support, the settlement at a support that sinks). Compatibility then reads [F]{X} + {D_L} = {D}, with every displacement measured in the direction of its own redundant.

Released structure, two redundantsw per metreX₁ = R_BX₂ = R_CABCDspans 5 m each, EI constantIndiacon original, not to scale
Figure 7.1: The three-span beam of Example 1 released to a simple beam AD, with the interior reactions X1 and X2 as redundants. Indiacon original, not to scale.

By Maxwell’s reciprocal theorem f_ij = f_ji, so [F] is symmetric; it is also positive definite for a stable primary structure, which means it can always be inverted. A non-symmetric flexibility matrix is a sure sign of an arithmetic or sign slip. The coefficients can be read from standard deflection formulas for simple cases, or computed for any frame with the unit-load method: f_ij = Σ∫ m_i m_j dx/(EI), where m_i is the bending moment due to a unit value of redundant i.

For members, a convenient form uses the member flexibility matrix. A prismatic member treated as simply supported, with end moments M_1 and M_2 applied, rotates at its ends by θ = [f_m]{M} with f_m = (L/6EI)[[2, 1], [1, 2]]. If the internal moments at the member ends are related to the redundants by a force-transformation matrix [B], the structure flexibility matrix is [F] = [B]ᵀ[f][B], where [f] collects the member matrices on its diagonal. This is the matrix version of the three-moment equation, and it leads naturally to the stiffness method of the next lesson, where the roles of forces and displacements are swapped.

Derivation

  1. Choose n redundants X_1 … X_n such that removing them leaves a stable, statically determinate primary structure (n = degree of static indeterminacy).
  2. Apply the loads alone to the primary structure and find the displacements D_Li at each release, measured in the positive direction of X_i.
  3. Apply X_j = 1 alone and find the displacement f_ij at every release i. By superposition, the displacement at release i under the loads plus all redundants is D_Li + Σ_j f_ij X_j.
  4. Compatibility: the displacement at release i must equal the real displacement D_i there (zero for a rigid support, the settlement for a yielding one). In matrix form [F]{X} = {D} − {D_L}, so {X} = [F]⁻¹({D} − {D_L}).
  5. For the unit-load method, the displacement at i due to X_j = 1 is f_ij = Σ∫ m_i m_j dx/(EI); the integrand is symmetric in i and j, which proves f_ij = f_ji.
  6. With member-end moments as internal forces, the complementary energy of the members is ½{M}ᵀ[f]{M} with {M} = [B]{X} + {M_L}; minimising it with respect to {X} gives [B]ᵀ[f][B]{X} = −[B]ᵀ[f]{M_L} + {D}, the same compatibility equations with [F] = [B]ᵀ[f][B].
Formulas, symbols and SI units
  • Compatibility: [F]{X} + {D_L} = {D}; {X} = [F]⁻¹({D} − {D_L}) (X in kN or kN m; D in m or rad)
  • Flexibility coefficient: f_ij = Σ∫ m_i m_j dx/(EI) (m/kN, rad/kN or rad/(kN m))
  • Maxwell: f_ij = f_ji, so [F] is symmetric and positive definite
  • Simple beam, unit load at a, deflection at the same point: a²b²/(3EIL) (m/kN), with b = L − a
  • Simple beam, uniform load w: deflection at x = w x (L³ − 2L x² + x³)/(24EI) (m)
  • Member flexibility (simply supported member, end moments): f_m = (L/6EI)[[2, 1], [1, 2]] (rad per kN m)
  • Structure flexibility from members: [F] = [B]ᵀ[f][B]
  • Final forces by superposition: any action S = S_L + Σ s_j X_j
Assumptions and limits
  • Linear elastic material and small displacements, so superposition holds and [F] does not depend on the loads.
  • Bending deformations govern beams and frames; axial and shear deformations are neglected unless stated (they can be added as extra terms in f_ij).
  • The primary structure is stable and statically determinate; a badly chosen release (for example one that makes a mechanism) gives a singular [F].
  • Supports are rigid unless a settlement or a spring is stated; settlements enter through {D}, not through [F].
  • Limits: the force method needs a new choice of redundants for every structure, so it is hard to automate; for large structures and for computer programs the stiffness method of 7.2 and 7.3 is used instead. Results are first-order: buckling and large-deflection effects are not included.

Codes and standards

Indian Standards are cited by number and clause only; no clause text, tables or figures are reproduced. Other standards are listed for comparison, not ranked.

  • Flexibility analysis is a result of mechanics; no design code governs the mechanics of this lesson, and no Indian Standard clause is cited here.
  • Indian, European and American practice, side by side and not ranked, all expect a structure to be analysed by a recognised elastic method with stiffness assumptions stated, and require that software used for analysis is checked against hand methods such as this one; their load and safety values are not reproduced here.

Worked examples

Example 1 (exam style: two redundants)

Problem. A continuous beam ABCD has three equal spans of 5 m, supports at A, B, C and D, and constant EI. It carries 12 kN/m over the whole 15 m. Using the interior reactions R_B and R_C as redundants, find all reactions and the support moments.

  1. Primary structure: a simple beam AD of span 15 m. Take X_1 = R_B and X_2 = R_C, positive upward; measure displacements upward.
  2. Unit load at B (a = 5 m, b = 10 m): f_11 = a²b²/(3EIL) = 25 × 100/(3 × 15) = 55.56/EI; by symmetry f_22 = 55.56/EI. Deflection at C due to a unit load at B: measure x from D, so the point C is at x = 5 m and the unit load at B is 10 m from D and b = 5 m from A: f_21 = b x (L² − b² − x²)/(6EIL) = 5 × 5 × (225 − 25 − 25)/(6 × 15) = 48.61/EI = f_12 (symmetric, as Maxwell requires).
  3. Load deflection at B and C (downward): w x (L³ − 2Lx² + x³)/(24EI) = 12 × 5 × (3 375 − 750 + 125)/24 = 6 875/EI, so D_L = −6 875/EI at both points (upward positive).
  4. Compatibility with rigid supports (D = 0): [55.56 48.61; 48.61 55.56]{X} = {6 875; 6 875}. Adding the equations: 104.17 X = 6 875, X_1 = X_2 = 66.0 kN.
  5. Equilibrium: R_A = R_D = (12 × 15 − 2 × 66)/2 = 24.0 kN. Moment at B from the left: 24 × 5 − 12 × 5²/2 = −30.0 kN m (hogging), and the same at C.

Answer. R_A = R_D = 24 kN, R_B = R_C = 66 kN (1.1wL); M_B = M_C = −30 kN m (−0.1wL²)

Example 2 (site style: member flexibility, three-moment form)

Problem. The floor beam ABC of lesson 4.1 is fixed at A, on a roller at B and pinned at C. Span AB = 6 m carries 20 kN/m and span BC = 4 m carries a central 40 kN load; EI is constant. Take the hogging moments M_A and M_B as redundants and solve with member flexibility matrices.

  1. Release: put hinges at A and B, so both spans become simply supported. The redundants are the hogging moments X_1 = M_A and X_2 = M_B; the matching displacements are the relative rotations at A and B.
  2. Member flexibilities: AB, (L/6EI)[2 1; 1 2] = (1/EI)[2 1; 1 2]; BC, (4/6EI)[2 1; 1 2], of which only the end-B term 4/(3EI) is needed because C is a pin with zero moment.
  3. Assemble [F] = [B]ᵀ[f][B]: f_11 = 2/EI, f_12 = f_21 = 1/EI, f_22 = 2/EI + 4/(3EI) = 10/(3EI).
  4. Load rotations of the released spans (opening of the hinges): at A, wL³/(24EI) = 20 × 216/24 = 180/EI; at B, 180/EI from AB plus PL²/(16EI) = 40 × 16/16 = 40/EI from BC, total 220/EI.
  5. Compatibility (hinges must close): 2X_1 + X_2 = 180 and X_1 + (10/3)X_2 = 220. Solving: X_2 = 780/17 = 45.88 kN m, X_1 = 1 140/17 = 67.06 kN m.
  6. Reactions by statics of each span: span AB, R_A = 60 + (67.06 − 45.88)/6 = 63.53 kN; span BC, R_C = 20 − 45.88/4 = 8.53 kN; R_B = 120 + 40 − 63.53 − 8.53 = 87.94 kN. These match the slope-deflection answer of lesson 4.1.

Answer. M_A = −67.06 kN m, M_B = −45.88 kN m (hogging); R_A = 63.53 kN, R_B = 87.94 kN, R_C = 8.53 kN

Example 3 (site style: support settlement)

Problem. The beam of Example 1 has EI = 40 000 kN m². During construction support C settles by 10 mm while B, A and D stay put. Find the reactions and the support moments under the same 12 kN/m load.

  1. Only {D} changes: D_1 = 0 at B and D_2 = −0.010 m at C (downward). Multiply compatibility by EI: [55.56 48.61; 48.61 55.56]{X} = 6 875{1; 1} + EI{D}.
  2. EI × D_2 = 40 000 × (−0.010) = −400 kN m³, so the right-hand side becomes {6 875; 6 475}.
  3. Determinant 55.56² − 48.61² = 723.4. X_1 = (6 875 × 55.56 − 48.61 × 6 475)/723.4 = 92.88 kN; X_2 = (55.56 × 6 475 − 48.61 × 6 875)/723.4 = 35.28 kN.
  4. Equilibrium: moments about D give 15R_A = 180 × 7.5 − 92.88 × 10 − 35.28 × 5, so R_A = 16.32 kN; then R_D = 180 − 16.32 − 92.88 − 35.28 = 35.52 kN.
  5. Moments: M_B = 16.32 × 5 − 12 × 25/2 = −68.4 kN m; M_C = 35.52 × 5 − 150 = +27.6 kN m. A 10 mm settlement has more than doubled the hogging at B and turned C into a sagging region.

Answer. R_A = 16.32, R_B = 92.88, R_C = 35.28, R_D = 35.52 kN; M_B = −68.4 kN m, M_C = +27.6 kN m (was −30 and −30 without settlement)

Step-by-step procedure

  1. Find the degree of static indeterminacy n and choose n redundants that leave a stable, determinate primary structure.
  2. Define a positive direction for each redundant and measure every displacement at that release in the same direction.
  3. Compute {D_L}: the displacements at the releases due to the loads on the primary structure (formula tables or unit-load integrals).
  4. Compute [F] column by column with X_j = 1; check that it is symmetric.
  5. Put settlements, lack of fit or temperature effects into {D}, solve [F]{X} = {D} − {D_L}, and keep the signs of the X values.
  6. Find every reaction and moment by superposition S = S_L + Σ s_j X_j, then check overall equilibrium and that the moments are zero at pins.

Site notes

  • The force method is the natural hand check for a two- or three-span floor beam: two redundants are quick to solve and catch input slips in a software model.
  • Settlement enters only on the right-hand side, so one [F] serves for many settlement cases; a settlement table for a building can be produced from one inversion.
  • A beam on soft ground or on a transfer girder is more sensitive to settlement than to load changes when it is stiff (large EI): the extra moment grows in proportion to EI.
  • When a test or monitoring record shows a support has moved, use the measured movement in {D} and re-check the critical sections rather than assuming the design moments still hold.
  • Keep the release choice simple: interior reactions for continuous beams, end moments for fixed ends; a well-conditioned [F] usually has diagonal terms that dominate the off-diagonal ones.
Common mistakes
  • Measuring D_L and f_ij in different directions, which gives a redundant of the wrong sign; every displacement at a release must use the direction of its own redundant.
  • Choosing releases that make the primary structure a mechanism, so [F] cannot be inverted.
  • Putting a settlement into [F] instead of into {D}, or using the settlement with the wrong sign.
  • Mixing EI-scaled and real values: if [F] is written as numbers times 1/EI, the settlement term must be multiplied by EI before solving.
  • Forgetting that a non-symmetric [F] means an error: f_12 and f_21 must agree.
Quick check

For n redundants, [F] must be symmetric with positive diagonal terms, and for every pair of redundants f_ii f_jj > f_ij² (a property of any positive definite matrix); in Example 1, 55.56² > 48.61². For a three-span continuous beam of equal spans under a full uniform load the interior reactions are 1.1wL and the support moments −0.1wL²; a matrix answer far from these needs checking.

Summary

  • The flexibility method writes consistent deformations in matrix form: [F]{X} = {D} − {D_L}.
  • f_ij is the displacement at release i for X_j = 1; Maxwell makes [F] symmetric.
  • Coefficients come from deflection formulas or from f_ij = Σ∫ m_i m_j dx/(EI); member matrices assemble as [B]ᵀ[f][B].
  • Settlements and lack of fit enter through {D} and can change the moments a great deal in stiff members.
  • The method suits hand checks with a few redundants; programs use the stiffness method instead.

Practice questions

Tap a question to see the answer.

1A cantilever of span 4 m carries 15 kN/m and is propped at its free end. Using the prop force as the redundant, find it.

f = L³/(3EI) = 21.33/EI; D_L = wL⁴/(8EI) = 480/EI downward; X = 480/21.33 = 22.5 kN (= 3wL/8).

2A two-span beam has equal spans of 5 m, simple end supports and a central support, and carries 8 kN/m throughout. Find the central reaction by the flexibility method.

Primary simple beam of 10 m: f = L³/(48EI) = 20.83/EI; D_L = 5wL⁴/(384EI) = 1 041.7/EI; X = 1 041.7/20.83 = 50 kN (= 1.25 × 8 × 5).

3Write the member flexibility matrix of a 6 m member with EI = 30 000 kN m².

(L/6EI)[2 1; 1 2] = [6.667 3.333; 3.333 6.667] × 10⁻⁵ rad per kN m.

4A propped cantilever of span 5 m with EI = 40 000 kN m² has its prop lowered by 10 mm with no load. Find the prop force.

f = L³/(3EI); X = 3EIΔ/L³ = 3 × 40 000 × 0.010/125 = 9.6 kN, acting downward on the beam (the beam is pulled down to the lowered prop).

5Why must the flexibility matrix be symmetric?

Because of Maxwell’s reciprocal theorem: the displacement at i due to a unit force at j equals the displacement at j due to a unit force at i, so f_ij = f_ji.

6What size is [F] for a beam fixed at both ends and continuous over two intermediate supports, loaded vertically?

Four redundants (for example the two end moments and two interior reactions, with axial effects ignored), so [F] is 4 × 4.

References

Standard texts, listed in no order of merit. Our explanations, examples and diagrams are original; nothing is copied from these books.

  • R. C. Hibbeler, Structural Analysis, Pearson, 10th edition, 2018.
  • C. S. Reddy, Basic Structural Analysis, Tata McGraw Hill (now McGraw Hill Education (India)), 3rd edition, 2011.
  • D. Menon, Structural Analysis, Narosa Publishing House, 2nd edition.
  • B. J. Goodno, J. M. Gere, Mechanics of Materials, Cengage Learning, 9th edition, 2018.
Suggest a correction

If something on this page is wrong, tell the editor. Your message is sent privately and is not published.

Sent privately to the Indiacon editor; not published.