Beam Deflection

Span matters most: double it, and the deflection grows eight- or sixteenfold.

Before this Lesson

Lay a ruler across two books, with a few inches of it hanging out past one of them, and press down on the middle of the gap. What does the end hanging out do? Your finger pushes down, so down seems like the safe guess. Try it before you read on. Whatever the ruler does, the reason is written in its moment diagram, and reading a beam's shape off that diagram is where this Lesson starts.

Each piece bends by its own moment

In Bending Stress, a moment tilted the faces of a short slice toward each other. Line the slices up end to end, and those small tilts add up to a curve. Bend a short piece into an arc of radius ρ\rho: the neutral axis keeps its length, and a fiber a distance cc below it, on a circle of radius ρ+c\rho + c, stretches by the fraction c/ρc/\rho, its strain.

ρ neutral axis c shortens stretches M M
A short piece of beam bent into an arc

The flexure formula says that same strain is σ/E=Mc/(EI)\sigma/E = Mc/(EI). Set the two equal, and cc drops out:

1ρ=MEI\frac{1}{\rho} = \frac{M}{EI}

The Handbook opens its Deflection of Beams section with this. 1/ρ1/\rho is the curvature, large for a tight bend. The moment bends the beam, and EIEI, the flexural rigidity, resists it. And the relation is local: each piece bends by the moment it carries, whatever the rest of the beam is doing. So the moment diagram, divided by EIEI, charts how sharply the beam bends all along its length.

The bends are gentle. A W18X50 under 200 kip⋅ft\text{kip}\cdot\text{ft} bends into an arc about 800 ft in radius. For a curve that flat, the curvature is the second derivative of the deflection yy, measured upward from the beam's original line, and the Handbook writes:

EId2ydx2=MEI\frac{d^2y}{dx^2} = M

The curve that yy traces is the elastic curve. This yy is a deflection, not the yy of σ=My/I\sigma = My/I.

EIEI is positive, so d2y/dx2d^2y/dx^2 has the sign of MM. A positive moment makes the beam concave up, a smile, which is exactly how the Handbook's sign convention from Internal Shear and Moment in Beams defines it. A negative moment makes a frown, and zero moment makes no bend at all.

The moment diagram is a map of the curve

So you can sketch a deflected shape from the moment diagram without integrating anything. Four rules do it:

  1. Positive moment bends the beam concave up, and negative moment concave down.
  2. Where the moment is zero over a stretch, that stretch stays straight.
  3. The curve passes through every pin and roller at zero deflection.
  4. The curve is smooth, with no kinks, because the beam is one piece.

Try them on a beam with an overhang under a uniform load. The moment rises from zero at AA to a positive peak, falls back through zero, and reaches its most negative value over BB, where the overhang hangs off the support.

w A B C M + − V = 0 M = 0 inflection point concave up concave down
The moment's sign sets the bend, and the supports pin the curve

From AA, the beam sags in a smile. At the zero crossing, the bend reverses, and from there to the tip the beam is a frown. It has to curve back up to meet BB, and past BB the same frown carries the overhang down. Whether the tip ends up below its start, as here, or above it depends on the proportions. A shorter overhang can be tilted up by the sag beside it. The moment fixes only the bends.

Now you can answer the ruler.

Checkpoint
1 of 2
P A B C
A beam with an overhang and one point load

A beam is pinned at AA, rests on a roller at BB, and overhangs to a free end at CC. It carries a point load PP in the middle of ABAB, like your finger on the ruler. Which sketch shows its deflected shape?

A.
A B C
B.
A B C
C.
A B C
D.
A B C

Points of inflection

Where the moment changes sign, the bend reverses, and the elastic curve has a point of inflection. That's the Primer's definition, carried through EI d2y/dx2=MEI\,d^2y/dx^2 = M: the second derivative changes sign exactly where the moment does. The Handbook doesn't make that link for you. On a moment diagram, look for where it crosses its zero line. Three kinds of point look like candidates and aren't:

  • Supports. A support holds the deflection at zero, which says where the beam is, not how it bends. Over BB in the figure, the beam bends as hard as anywhere.
  • Points of zero shear. Shear is the slope of the moment diagram, so zero shear marks a peak of the moment, where the bend is at its sharpest nearby.
  • Ends where the moment is zero. At a free end, or an end pin such as AA, the moment is zero but doesn't change sign. The beam just stops.
Checkpoint
w A B C D E F G + − − M
A beam overhanging both supports, with its moment diagram

This beam rests on supports at BB and FF and overhangs both of them, under a uniform load. Which labeled points are points of inflection?

A.
A
B.
B
C.
C
D.
D
E.
E
F.
F
G.
G

Pinning the curve in place

The moment diagram fixes how the beam bends, but not where it sits. To get numbers, the Handbook integrates twice:

EIdydx=∫M(x) dxEIy=∫[∫M(x) dx]dxEI\frac{dy}{dx} = \int M(x)\,dx \qquad EIy = \int\left[\int M(x)\,dx\right]dx

Each integration adds a constant, so the result is like a bent wire hovering over the beam: you can slide it up or down, or tip it, without changing its shape. The supports hold it still. The Handbook says only that the constants "can be determined from the physical geometry of the beam." Here's what that geometry says:

  • Pin or roller: holds the height but lets the beam rotate, so y=0y = 0.
  • Fixed end: holds the height and the angle, so y=0y = 0 and dy/dx=0dy/dx = 0.
  • Free end: holds nothing, so it gives no condition.
x y Simply supported y = 0, free to rotate y = 0 Cantilever y = 0 and dy/dx = 0 free: no condition
What each support holds still

Two constants need two conditions: a simple beam's two supports, or a cantilever's wall. The Elastic Curve column of the Handbook's beam tables is this integration, already done for each case, so the exam asks you to name the conditions, not to integrate.

Checkpoint
1 of 2

A beam is pinned at x=0x = 0 and rests on a roller at x=Lx = L. Which conditions fix the constants when you integrate EI d2y/dx2=MEI\,d^2y/dx^2 = M?

A.
dy/dx=0 at x=0dy/dx = 0 \text{ at } x = 0
B.
y=0 at x=0y = 0 \text{ at } x = 0
C.
dy/dx=0 at x=Ldy/dx = 0 \text{ at } x = L
D.
y=0 at x=Ly = 0 \text{ at } x = L

The Handbook's four cases

The Handbook's tables Simply Supported Beam Slopes and Deflections and Cantilevered Beam Slopes and Deflections give the results case by case. Four rows cover most single-case questions:

Maximum deflections and slopes from the Handbook's beam tables
Beam and loadvmaxv_{\text{max}}θmax\theta_{\text{max}}
Simple, uniform ww−5wL4384EI-\dfrac{5wL^4}{384EI}−wL324EI-\dfrac{wL^3}{24EI}
Simple, PP at center−PL348EI-\dfrac{PL^3}{48EI}−PL216EI-\dfrac{PL^2}{16EI}
Cantilever, PP at free end−PL33EI-\dfrac{PL^3}{3EI}−PL22EI-\dfrac{PL^2}{2EI}
Cantilever, uniform ww−wL48EI-\dfrac{wL^4}{8EI}−wL36EI-\dfrac{wL^3}{6EI}

Adapted from the FE Reference Handbook, Simply Supported Beam Slopes and Deflections and Cantilevered Beam Slopes and Deflections

The tables call the deflection vv, measured upward like yy, so a sag prints with a minus sign. Drop it and say "down." θ\theta is the slope, in radians. A simple beam's vmaxv_{\text{max}} is at midspan and its θmax\theta_{\text{max}} at the ends, and a cantilever has both at its free end. The table doesn't print the cantilever's end load as its own row. It gives a load at a distance aa from the wall, and a=La = L turns it into the end-load row above.

The tables print no units. In USCS, use kips, inches, and ksi, with the load in kip/in, which is kip/ft divided by 12. In SI, use N, mm, and MPa. A load in kN/m is the same number in N/mm. The Handbook's Table 1 - Typical Material Properties gives EE: 29.0 Mpsi (200 GPa) for steel, 10.0 Mpsi (69.0 GPa) for aluminum, and 1.6 Mpsi (11.0 GPa) for wood (fir). Adding cases for a beam with several loads waits for Beam Deflection by Superposition.

Worked Example The sag of a steel floor beam
1 of 3

A W18X50 spans 24 ft, simply supported, under a uniform load of 2 kip/ft. Find its maximum deflection.

Why span matters most

Every row has the same anatomy: a load, a power of LL, and EIEI underneath. The power comes from the curvature. Double the span under the same point load. The load sits on twice the lever arm, so the moment doubles, and every piece bends twice as sharply. That bending runs over twice the length, so the tilt it builds doubles again. And the tilt carries over twice the distance to reach midspan, so the sag doubles a third time: 2×2×2=82 \times 2 \times 2 = 8, the L3L^3 in PL3/(48EI)PL^3/(48EI). A uniform load adds one more factor of LL, because a longer span at the same ww also carries more load.

The rest is simpler. Deflection is proportional to the load, and EE and II sit underneath. For a rectangle, I=bh3/12I = bh^3/12, so depth counts three times and width once. That's the plank from Bending Stress again: on edge, it's five times as strong, but its II is 25 times as large, so it sags a twenty-fifth as far.

Checkpoint
1 of 3

A simply supported beam carries a uniform load. Its span doubles, with the same load per foot and the same section. What happens to its maximum deflection?

A.
It doubles.
B.
It grows 8 times.
C.
It grows 16 times.
D.
It grows 4 times.

Recap

  • EI d2y/dx2=MEI\,d^2y/dx^2 = M: positive moment bends a beam concave up, negative moment concave down, and zero moment leaves it straight.
  • Sketch a deflected shape from the moment diagram's signs, with the curve pinned at zero at every support. An unloaded overhang stays straight and tilts up when the span beside it sags.
  • Points of inflection are where the moment changes sign, not at supports, points of zero shear, or free ends.
  • A pin or roller gives y=0y = 0. A fixed end gives y=0y = 0 and dy/dx=0dy/dx = 0.
  • Match one table case, in kip/in, in., and ksi, or N/mm, mm, and MPa. Deflection grows with the load and with L3L^3 or L4L^4, and shrinks with EIEI. A rectangle's depth counts three times.