Indiacon CivilSubscribe

Textbook › Strength of Materials › Chapter 5 › Lesson 5.2

5.2 Double integration and Macaulay's method

Strength of Materials Chapter 5: Deflection of beams Under review — not yet three-checked Author: Indiacon

Learning objectives

After this lesson you should be able to:

  • Apply double integration to beams whose bending moment changes form along the span, and see why it needs many constants when done segment by segment.
  • Use Macaulay's bracket notation to write one moment equation for a whole beam with point loads, part-span distributed loads and applied couples.
  • Find slopes, deflections under loads and the position and size of the maximum deflection for simply supported beams carrying several loads.

Theory

Lesson 5.1 integrated E I y″ = M for beams where one expression for M holds over the whole span. Most real beams carry several loads: a main beam supports secondary beams at two or three points, a lintel carries a wall over part of its length, a bridge girder carries axle loads. The bending moment then has a different formula in each segment between loads. Integrating each segment separately works, but every segment adds two constants, so a beam with three loads has four segments and eight constants, fixed by two support conditions and six continuity conditions (slope and deflection equal on both sides of each load). The algebra is long and easy to get wrong.

W. H. Macaulay published a neat fix in 1919. Write the moment at a section x in the last segment, taking moments of all forces to the left, and put each term that only starts beyond some point a inside special brackets ⟨x − a⟩. The rule is that a bracket is zero when its content is negative and is an ordinary number when it is positive or zero. One expression then gives the correct moment everywhere: for sections to the left of a load, its bracket simply vanishes.

P₁P₂R_AR_BxabM = R_A x − P₁⟨x − a⟩ − P₂⟨x − b⟩⟨x − a⟩ = 0 when x < a; integrate as ⟨x − a⟩²/2Indiacon original
Figure 5.2: Macaulay's method: one moment equation, written for a section beyond the last load, covers the whole beam because each bracket is zero until its load is passed. Indiacon original, not to scale.

The brackets are integrated as a whole: ∫⟨x − a⟩ⁿ dx = ⟨x − a⟩ⁿ⁺¹/(n + 1). Done this way, the slope and deflection expressions automatically match on both sides of each load, so the continuity conditions are satisfied without being written. Only two constants of integration remain, whatever the number of loads, and they come from the two support conditions. A couple M₀ at a enters as M₀⟨x − a⟩⁰, a term that switches on at a.

A distributed load that stops before the right end of the beam needs one trick. A bracket term can only switch on, never off, so the load is extended to the right end and an equal and opposite load is added over the extension to cancel it. For a load w from a to b, the moment terms are −w⟨x − a⟩²/2 + w⟨x − b⟩²/2. With this, Macaulay's method handles nearly every statically determinate beam met in practice, and it also supplies the deflection equations used later for propped and fixed beams.

Derivation

  1. Simply supported beam of span L with a point load P at distance a from the left support A and b = L − a from the right support B. Reactions: R_A = P b/L, R_B = P a/L.
  2. Moment at x, taking moments of forces to the left: M = (P b/L) x − P⟨x − a⟩, where ⟨x − a⟩ = 0 for x < a.
  3. E I y″ = (P b/L) x − P⟨x − a⟩.
  4. E I y′ = P b x²/(2L) − P⟨x − a⟩²/2 + C₁.
  5. E I y = P b x³/(6L) − P⟨x − a⟩³/6 + C₁ x + C₂.
  6. y(0) = 0: the bracket is zero there, so C₂ = 0.
  7. y(L) = 0: P b L²/6 − P b³/6 + C₁ L = 0, so C₁ = −P b (L² − b²)/(6L).
  8. Deflection under the load (x = a, bracket zero): E I y = P b a³/(6L) − P b (L² − b²) a/(6L) = −P a² b²/(3L), so δ = P a² b²/(3 E I L).
  9. Maximum deflection: for a > b it lies in the segment 0 < x < a, where y′ = 0 gives x₀ = √((L² − b²)/3), and δ_max = P b (L² − b²)^(3/2)/(9√3 E I L). For a central load, a = b = L/2 and both results reduce to P L³/(48 E I).
  10. Part-span load w from a to b: M contains −w⟨x − a⟩²/2 + w⟨x − b⟩²/2, which integrate to −w⟨x − a⟩⁴/24 + w⟨x − b⟩⁴/24 in E I y.
Formulas, symbols and SI units
  • Macaulay bracket: ⟨x − a⟩ⁿ = 0 for x < a; (x − a)ⁿ for x ≥ a
  • Integration: ∫⟨x − a⟩ⁿ dx = ⟨x − a⟩ⁿ⁺¹/(n + 1)
  • Point load P at a: moment term −P⟨x − a⟩
  • Uniform load w from a onward: −w⟨x − a⟩²/2; stopping at b: add +w⟨x − b⟩²/2
  • Couple M₀ (clockwise) at a: +M₀⟨x − a⟩⁰
  • Point load P at a on simply supported span L (b = L − a): δ under load = P a² b²/(3 E I L)
  • δ_max = P b (L² − b²)^(3/2)/(9√3 E I L) at x₀ = √((L² − b²)/3) measured from A, when a ≥ b
  • Units: kN, m and kN·m² give deflection in m; multiply by 1000 for mm
Assumptions and limits
  • All assumptions of lesson 5.1 apply: linear elastic material, small slopes, shear deformation neglected, and the same sign convention (y upward, sagging moment positive, so downward deflections come out negative).
  • E I is constant along the beam; for a stepped beam, M/(E I) must be written segment by segment or the moment-area method (lesson 5.3) used.
  • Moments are always taken of the forces to the left of the section, with the origin at the left end, so that every bracket switches on as x increases.
  • Brackets are never expanded before integration; (x − a)² expanded and integrated term by term gives different constants and breaks the continuity property.
  • Limits: the method gives the bending deflection of determinate beams directly; for indeterminate beams the redundant reactions are kept as unknowns and found from extra boundary conditions.

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.

  • Macaulay's method is mechanics, and no code governs it; codes use its results when deflections are checked against their limits.
  • The limits explained in lesson 5.1 apply to beams with several loads too: IS 456:2000 cl. 23.2 for reinforced concrete members and IS 800:2007 Table 6 for steel members, with the load case each code asks for.
  • Side by side, not ranked: EN 1993-1-1, ACI 318 and AISC 360 practice likewise accept deflections from any valid elastic analysis, by hand integration or by software, and then compare them with the limits for the member's use.

Worked examples

Example 1 (exam style: off-centre point load)

Problem. A simply supported steel beam of span 6 m carries a point load of 40 kN at 2 m from the left support. E I = 2.4 × 10¹³ N·mm² (24 000 kN·m²). Find the end slopes, the deflection under the load, and the maximum deflection and its position.

  1. Reactions: R_A = 40 × 4/6 = 26.67 kN, R_B = 13.33 kN.
  2. E I y″ = 26.67 x − 40⟨x − 2⟩ (kN, m). E I y′ = 13.33 x² − 20⟨x − 2⟩² + C₁; E I y = 4.444 x³ − 6.667⟨x − 2⟩³ + C₁ x + C₂.
  3. y(0) = 0 gives C₂ = 0. y(6) = 0: 4.444 × 216 − 6.667 × 64 + 6 C₁ = 0, so C₁ = −88.89 kN·m².
  4. End slopes: θ_A = C₁/(E I) = −88.89/24 000 = −3.70 × 10⁻³ rad; at B, E I y′ = 13.33 × 36 − 20 × 16 − 88.89 = 71.11, θ_B = 2.96 × 10⁻³ rad.
  5. Under the load: E I y = 4.444 × 8 − 88.89 × 2 = −142.2, so y = −142.2/24 000 = −5.93 mm (check: P a² b²/(3 E I L) = 40 × 4 × 16/(3 × 24 000 × 6) = 5.93 mm).
  6. The maximum is in the longer segment, where y′ = 0: 13.33 x² − 20 (x − 2)² − 88.89 = 0 gives x = 2.734 m from A (3.266 m from B, matching √((L² − a²)/3)), and δ_max = 6.45 mm.

Answer. θ_A = 3.70 × 10⁻³ rad, θ_B = 2.96 × 10⁻³ rad; 5.93 mm under the load; δ_max = 6.45 mm at 2.734 m from the left

Example 2 (site style: main beam carrying two secondary beams)

Problem. A main steel beam of span 8 m receives two secondary beams: 60 kN at 2 m and 40 kN at 5 m from the left support. E I = 6.0 × 10¹³ N·mm² (60 000 kN·m²). Find the deflection at each load and at midspan, and the maximum deflection.

  1. Reactions: R_B = (60 × 2 + 40 × 5)/8 = 40 kN; R_A = 100 − 40 = 60 kN.
  2. E I y″ = 60 x − 60⟨x − 2⟩ − 40⟨x − 5⟩; E I y = 10 x³ − 10⟨x − 2⟩³ − 6.667⟨x − 5⟩³ + C₁ x (C₂ = 0 from y(0) = 0).
  3. y(8) = 0: 5120 − 2160 − 180 + 8 C₁ = 0, so C₁ = −347.5 kN·m².
  4. At x = 2 m: E I y = 80 − 695 = −615, y = −10.25 mm. At x = 4 m: E I y = 640 − 80 − 1390 = −830, y = −13.83 mm. At x = 5 m: E I y = 1250 − 270 − 1737.5 = −757.5, y = −12.63 mm.
  5. Maximum, between the loads where y′ = 0: 30 x² − 30 (x − 2)² − 347.5 = 0 gives 120 x = 467.5, x = 3.896 m; δ_max = 13.84 mm, only 0.01 mm more than at midspan.
  6. Limit span/300 = 26.7 mm, so the beam is well within it under these loads.

Answer. 10.25 mm at 2 m, 13.83 mm at midspan, 12.63 mm at 5 m; δ_max = 13.84 mm at 3.896 m

Example 3 (part-span load: wall on half a lintel beam)

Problem. A simply supported beam of span 6 m carries 10 kN/m over the left 3 m only (a wall ending at midspan). E I = 24 000 kN·m². Find the deflection at midspan and the maximum deflection.

  1. Reactions: total 30 kN acting at 1.5 m, so R_B = 30 × 1.5/6 = 7.5 kN and R_A = 22.5 kN.
  2. Extend the load to the right end and cancel it: E I y″ = 22.5 x − 5 x² + 5⟨x − 3⟩².
  3. E I y = 3.75 x³ − (5/12) x⁴ + (5/12)⟨x − 3⟩⁴ + C₁ x, with C₂ = 0.
  4. y(6) = 0: 810 − 540 + 33.75 + 6 C₁ = 0, so C₁ = −50.625 kN·m².
  5. At midspan, x = 3 (bracket zero): E I y = 101.25 − 33.75 − 151.875 = −84.375, y = −3.52 mm.
  6. Setting y′ = 0 in the loaded part (11.25 x² − (5/3) x³ − 50.625 = 0) gives x = 2.76 m and δ_max = 3.54 mm, close to midspan as is usual for simply supported beams.

Answer. δ at midspan = 3.52 mm; δ_max = 3.54 mm at 2.76 m from the loaded end

Step-by-step procedure

  1. Place the origin at the left end and find the reactions.
  2. Write one moment equation for a section in the last segment, taking forces to the left; put every term that starts at x = a inside ⟨x − a⟩.
  3. Extend part-span distributed loads to the right end and add equal and opposite loads to cancel the extension.
  4. Integrate twice without expanding the brackets, adding only C₁ and C₂.
  5. Apply the support conditions, dropping any bracket whose content is negative at that point, to find C₁ and C₂.
  6. Evaluate y at the points needed; find the maximum where y′ = 0, first guessing the segment from the shape of the beam and then confirming the root lies in it.

Site notes

  • Main beams that carry secondary beams at a few points are a classic Macaulay case; the maximum deflection is usually within a few per cent of the midspan value, which is why midspan is often checked first.
  • Walls or stacked materials over part of a beam, such as a brick wall stopping short of a door opening, produce part-span loads; the trick of extending and cancelling the load is the hand route, and software does the same internally.
  • A beam that deflects more than expected under a concentrated load near one support is often due to an assumed fixity that the actual connection does not provide; a simple seat connection behaves close to a simple support.
  • Deflections from hand calculations are a useful check on software output: if a frame analysis gives a deflection very different from a Macaulay estimate for the same beam, check supports, units and the I used.
Common mistakes
  • Expanding a bracket such as (x − 2)³ before integrating, which changes the constants and gives wrong deflections.
  • Keeping a bracket term when its content is negative (for example ⟨x − 5⟩ at x = 2), instead of setting it to zero.
  • Forgetting to cancel a part-span load that was extended to the end of the beam.
  • Taking moments of forces to the right of the section in some terms and to the left in others.
  • Searching for the maximum deflection in the wrong segment and accepting a root that lies outside it.
Quick check

With Macaulay brackets, write one moment equation from the left end, integrate brackets as a whole and use only two constants. For a single point load, δ under the load = P a² b²/(3 E I L); for a central load this becomes P L³/(48 E I). On a simply supported beam the maximum deflection rarely differs from the midspan value by more than a few per cent; a much larger difference signals an error.

Summary

  • Segment-by-segment double integration needs two constants per segment and continuity conditions at every load.
  • Macaulay brackets ⟨x − a⟩ switch a term on at x = a and are zero before it, giving one moment equation for the whole beam.
  • Integrating brackets as a whole keeps slope and deflection continuous, leaving just two constants from the supports.
  • Part-span loads are extended to the end and cancelled; couples enter as M₀⟨x − a⟩⁰.
  • The maximum deflection lies where y′ = 0, normally close to midspan for simply supported beams.

Practice questions

Tap a question to see the answer.

1What value does ⟨x − 3⟩² take at x = 2 m and at x = 5 m?

At x = 2 m the content is negative, so the bracket is zero. At x = 5 m it is (5 − 3)² = 4 m².

2Using Macaulay's method, show that a central load P on a simply supported span L gives δ = P L³/(48 E I).

E I y = P x³/12 − P⟨x − L/2⟩³/6 + C₁ x; y(L) = 0 gives C₁ = −P L²/16. At x = L/2: E I y = P L³/96 − P L³/32 = −P L³/48.

3A 20 kN load acts 1 m from the left support of a simply supported 4 m span. E I = 1.0 × 10¹³ N·mm² (10 000 kN·m²). Find the deflection under the load.

δ = P a² b²/(3 E I L) = 20 × 1² × 3²/(3 × 10 000 × 4) = 1.5 × 10⁻³ m = 1.5 mm.

4A uniform load w acts from x = 2 m to x = 4 m on a 6 m simply supported beam. Write its terms in the Macaulay moment equation.

−w⟨x − 2⟩²/2 + w⟨x − 4⟩²/2; the second term cancels the load beyond 4 m.

5A clockwise couple M₀ acts at the midspan of a simply supported beam. What is the deflection at midspan, and why?

Zero. The loading is antisymmetric about midspan, so the elastic curve is antisymmetric too: one half deflects down and the other up, crossing zero at the centre.

6Why does Macaulay's method need only two constants of integration however many loads there are?

Because bracket terms are zero up to their starting point and are integrated as a whole, the slope and deflection expressions are automatically continuous at every load, so the continuity conditions that would fix extra constants are already satisfied.

References

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

  • B. J. Goodno, J. M. Gere, Mechanics of Materials, Cengage Learning, 9th edition, 2018.
  • R. C. Hibbeler, Mechanics of Materials, Pearson, 10th edition, 2016.
  • F. P. Beer, E. R. Johnston, J. T. DeWolf, D. F. Mazurek, Mechanics of Materials, McGraw Hill, 8th edition, 2020.
  • E. P. Popov, Engineering Mechanics of Solids, Prentice Hall, 2nd edition, 1998.
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.