Shear in Reinforced Concrete Beams

Concrete cracks diagonally near the supports, and stirrups stitch the cracks closed.

The beam from Reinforced Concrete Beam Strength, 12 in. wide with four #8 bars at d=20d = 20 in. in 4,000 psi normal-weight concrete, has ϕMn=251.4\phi M_n = 251.4 ft-kip. On an 18 ft span under a factored 6.0 kip/ft, bending checks out: Mu=6.0(18)2/8=243M_u = 6.0(18)^2/8 = 243 ft-kip.

But each support has to push up Vu=6.0(18)/2=54V_u = 6.0(18)/2 = 54 kips, and the code credits the concrete with only about 23 of them. A beam short on shear fails near a support, where the moment is small, along a crack that slants up toward midspan, and it goes all at once. This Lesson finds those 23 kips, and the steel that carries the rest.

Shear stretches a diagonal

Near a support the moment is small but the shear is at its largest, and shear does something sneaky. Draw a square on the side of a deck of cards and push the top of the deck sideways. The square leans into a diamond: one diagonal gets longer, the other shorter. Shear on a square is a pull along one diagonal and a push along the other, each as large as the shear stress. That's the pure-shear case from Principal Stresses and Mohr's Circle.

Concrete shrugs off the push but cracks under a pull of a few hundred psi, so it cracks across the stretched diagonal, at about 45°. Low in the beam, bending tension adds to the pull, so a crack often starts vertical at the bottom and leans over as it climbs toward midspan.

τ tension crack w midspan diagonal cracks flexural cracks tension steel A square near the support Cracks in the left half of a beam
Shear near a support pulls along a diagonal, and the concrete cracks across it

The slanted crack is the dangerous one. The bottom bars hold a flexural crack shut and yield slowly, so the beam sags where everyone can see it. Nothing yet holds a diagonal crack shut, so once it opens, it can run through the beam at once. That's why shear, which rarely governed in Selecting Steel Beams, needs its own steel in almost every concrete beam.

What the concrete can carry

The Beams—Shear provisions in the Handbook's Design of Reinforced Concrete Components (ACI 318-14) section use the same check of demand versus capacity as flexure, ϕVn≥Vu\phi V_n \ge V_u, and split the nominal strength between the concrete and the steel, Vn=Vc+VsV_n = V_c + V_s. The concrete's part is

Vc=2λfc′ bwdV_c = 2\lambda\sqrt{f'_c}\,b_wd

Here bwb_w is the web width, which for a rectangular beam is just bb (the Handbook uses it without defining it), and dd is the effective depth from flexure. The Handbook gives λ=1.0\lambda = 1.0 for normal-weight concrete and 0.75 for lightweight concrete.

Read 2fc′2\sqrt{f'_c} as a stress over bwdb_wd: 126 psi for 4,000 psi concrete. VcV_c is roughly the shear at which diagonal cracks form, set on the low side of tests, and it goes with fc′\sqrt{f'_c} because cracking is a tension failure. The formula expects fc′f'_c in psi, with bwb_w and dd in inches, and gives pounds. Put 4 ksi under the root, and you get 2 instead of 63.2, about 32 times too small.

The Handbook's Resistance Factors list gives ϕ=0.75\phi = 0.75 for shear, against 0.9 for a tension-controlled beam in bending, because a shear failure gives no warning. For the hook's beam:

ϕVc=0.75(2)(1.0)4,000 (12)(20)=22,770 lb=22.8 kips\phi V_c = 0.75(2)(1.0)\sqrt{4{,}000}\,(12)(20) = 22{,}770\ \text{lb} = 22.8\ \text{kips}

That's the 23 kips, well short of the 54 at the support.

Checkpoint
1 of 2

A rectangular beam is 16 in. wide with d=26d = 26 in., made of normal-weight concrete with fc′=5,000f'_c = 5{,}000 psi. What is its concrete design shear strength ϕVc\phi V_c?

kips

Stirrups stitch the crack

The fix is steel across the crack. Stirrups are small bars, usually #3 or #4, bent into loops around the main bars and standing upright every ss inches along the beam. A loop runs up one side of the beam and down the other, so a crack cuts it twice, and its area AvA_v counts both legs. From the Handbook's bar table, a #3 stirrup has Av=2(0.11)=0.22A_v = 2(0.11) = 0.22 in², and a #4 has 0.40 in².

Now count. The crack climbs about dd at about 45°, so it also runs about dd along the beam, crossing about d/sd/s stirrups. Once it opens, each of them stretches until it yields, pulling the two sides together with AvfyA_vf_y:

Vs=ds⏟stirrups crossed×Avfy⏟pull in each=AvfydsV_s = \underbrace{\frac{d}{s}}_{\text{stirrups crossed}} \times \underbrace{A_vf_y}_{\text{pull in each}} = \frac{A_vf_yd}{s}
Avfy crack s stirrup d ≈ d tension steel leg leg bw d Av = area of both legs Elevation Section
A 45° crack crosses about d/s stirrups, and each stirrup has two legs

That's the Handbook's VsV_s. For design, ϕ(Vc+Vs)≥Vu\phi(V_c + V_s) \ge V_u, so the stirrups have to supply

Vs=Vuϕ−Vcs=AvfydVsV_s = \frac{V_u}{\phi} - V_c \qquad s = \frac{A_vf_yd}{V_s}

Both are in the Handbook's stirrup spacing table. Divide by ϕ\phi before you subtract: VuV_u is factored and VcV_c is nominal, so Vu−VcV_u - V_c mixes the two and shortchanges the stirrups. With AvA_v in in², the stirrups' fyf_y in ksi (usually 60), dd in inches, and VsV_s in kips, ss comes out in inches.

Strength sets one spacing, and geometry sets another. A crack that slips between two stirrups crosses neither. Cracks run about dd along the beam, so the table caps ss at the smaller of d/2d/2 and 24 in., and every crack meets a stirrup or two. When Vs>4fc′ bwdV_s > 4\sqrt{f'_c}\,b_wd, the caps halve to d/4d/4 and 12 in., so more stirrups share each crack. And VsV_s may not exceed 8fc′ bwd8\sqrt{f'_c}\,b_wd at all: past that, the concrete between the cracks crushes first, and the fix is a bigger beam.

Worked Example Stirrups at the hook's support
1 of 3

The hook's beam is 12 in. wide, with d=20d = 20 in. and normal-weight concrete with fc′=4,000f'_c = 4{,}000 psi. At the support, Vu=54V_u = 54 kips. Find the spacing of #3 stirrups with fy=60f_y = 60 ksi there.

Where the beam needs stirrups

A beam with no stirrups has nothing to catch a diagonal crack, and VcV_c is only an estimate of when one forms; real beams scatter around it. So the code doesn't let the concrete stand alone even when ϕVc\phi V_c covers VuV_u on paper. A few stirrups let a crack open slowly, in plain view. Only when VuV_u is at most half of ϕVc\phi V_c can the beam go without. The Handbook's table sorts it out:

Required and maximum stirrup spacing
Factored shearStirrupsRequired spacingMaximum spacing
Vu≤ϕVc/2V_u \le \phi V_c/2None——
ϕVc/2<Vu≤ϕVc\phi V_c/2 < V_u \le \phi V_cMinimumAvfy50bw or Avfy0.75bwfc′\dfrac{A_vf_y}{50b_w}\ \text{or}\ \dfrac{A_vf_y}{0.75b_w\sqrt{f'_c}}d/2 or 24 in.d/2\ \text{or}\ 24\ \text{in.}
Vu>ϕVc, Vs≤4fc′ bwdV_u > \phi V_c,\ V_s \le 4\sqrt{f'_c}\,b_wdDesignedAvfyd/VsA_vf_yd/V_sd/2 or 24 in.d/2\ \text{or}\ 24\ \text{in.}
Vu>ϕVc, Vs>4fc′ bwdV_u > \phi V_c,\ V_s > 4\sqrt{f'_c}\,b_wdDesignedAvfyd/VsA_vf_yd/V_sd/4 or 12 in.d/4\ \text{or}\ 12\ \text{in.}

Source: Adapted from the FE Reference Handbook, Design of Reinforced Concrete Components (ACI 318-14), Beams—Shear

Each row's spacing is the smallest value it lists. The two minimum-stirrup spacings are the code's minimum stirrup area, turned into a spacing, and they take fyf_y in psi, like fc′f'_c. For #3 stirrups in the hook's beam, with fy=60,000f_y = 60{,}000 psi, they're 22 and 23.2 in., so d/2=10d/2 = 10 in. governs, as it usually does.

Along a uniformly loaded beam, VuV_u falls in a straight line from the support to zero at midspan, so the cases become zones. For the hook's beam under its full 6.0 kip/ft, Vu=54−6.0xV_u = 54 - 6.0x kips at xx ft from a support:

0 1 2 3 4 5 6 7 8 9 0 10 20 30 40 50 60 midspan distance from support (ft) Vu (kips) φVc = 22.8 φVc/2 = 11.4 5.2 7.1 54 designed stirrups minimum none
The hook's beam: factored shear over half its span, and the stirrups each part needs

Designed stirrups run 5.2 ft in from each support, minimum stirrups the next 1.9 ft, and the middle 3.8 ft needs none.

Checkpoint
1 of 2

A beam 10 in. wide with d=16d = 16 in., made of normal-weight concrete with fc′=3,000f'_c = 3{,}000 psi, carries Vu=10V_u = 10 kips at a section. What does it need there, with #3 stirrups and fy=60f_y = 60 ksi?

A.
No stirrups, because VuV_u is less than ϕVc\phi V_c
B.
Designed stirrups, from Vs=Vu/ϕ−VcV_s = V_u/\phi - V_c, no more than 8 in. apart
C.
Minimum stirrups, no more than 8 in. apart
D.
Minimum stirrups, no more than 4 in. apart

Recap

  • Near a support, shear stretches the concrete along a diagonal, so it cracks at about 45°, and with nothing across the crack, the beam fails suddenly.
  • Vc=2λfc′ bwdV_c = 2\lambda\sqrt{f'_c}\,b_wd takes fc′f'_c in psi and gives pounds, with λ=0.75\lambda = 0.75 for lightweight concrete. For shear, ϕ=0.75\phi = 0.75.
  • A crack runs about dd along the beam, crossing d/sd/s stirrups that each pull AvfyA_vf_y, with both legs in AvA_v: Vs=Avfyd/sV_s = A_vf_yd/s.
  • Design with Vs=Vu/ϕ−VcV_s = V_u/\phi - V_c and s=Avfyd/Vss = A_vf_yd/V_s, capped at d/2d/2 and 24 in., or d/4d/4 and 12 in. once Vs>4fc′ bwdV_s > 4\sqrt{f'_c}\,b_wd.
  • Vu≤ϕVc/2V_u \le \phi V_c/2 needs no stirrups, up to ϕVc\phi V_c needs minimum stirrups, and above ϕVc\phi V_c needs designed ones.