Designing Reinforced Concrete Beams

Run the stress block backward to find the steel a beam needs.

Take the worked example's beam from Reinforced Concrete Beam Strength, 12 in. wide with its steel 20 in. down, and keep adding bars. Each one buys a little less moment than the last. Past about 4.3 in², it gets stranger: add 60% more steel, and the design strength rises about 2%, while the beam loses its warning before failure.

Designing a beam means choosing that steel: enough for the moment, not too much, and laid out so it fits.

From the moment to the steel

Analysis hands you the bars and asks for the strength. Design hands you the moment, usually MuM_u from the factored loads, and a width and depth, and asks for the bars:

ϕAsfy(d−a2)=Mu\phi A_sf_y\left(d - \frac{a}{2}\right) = M_u

If a question gives the required nominal moment MnM_n instead, leave ϕ\phi out.

The trouble is that AsA_s shows up twice: once in the pull, and again inside aa. The Handbook gives no formula for the steel area, so you solve its two equations together. Substitute for aa, and you get a quadratic in AsA_s, which your calculator's equation solver can handle (take the smaller root):

Mn=Asfy d−(Asfy)21.7fc′ bM_n = A_sf_y\,d - \frac{(A_sf_y)^2}{1.7f'_c\,b}

By hand, there's a quicker way. The lever arm d−a/2d - a/2 is mostly dd, with a/2a/2 trimming a couple of inches off. So guess the arm at 0.9d0.9d, find the steel that arm needs, find the aa that steel makes, and correct the arm. A second round is usually enough.

Notice the minus sign. More steel means a deeper block and a shorter arm, so each added square inch buys less moment than the last.

Bar sizes go by number. Up to #8, the number is the diameter in eighths of an inch, so a #4 bar is 1/2 in. across. From #9 up, the bars run a little larger, so read them from the Handbook's bar table:

Diameters and areas of standard bars
Bar sizeDiameter (in.)Area (in²)
#30.3750.11
#40.5000.20
#50.6250.31
#60.7500.44
#70.8750.60
#81.0000.79
#91.1281.00
#101.2701.27
#111.4101.56

Source: Adapted from the FE Reference Handbook, Design of Reinforced Concrete Components (ACI 318-14), ASTM Standard Reinforcement Bars

To choose bars, divide the required area by one bar's area and round the count up, never down. The most efficient arrangement is the one with the least area that's still enough, as long as the bars fit across the beam. Then confirm εt\varepsilon_t, because the whole calculation assumed ϕ=0.9\phi = 0.9.

Worked Example Steel for a factored moment
1 of 3

A beam 12 in. wide with d=20d = 20 in. must carry Mu=185M_u = 185 ft-kip. It has fc′=4,000f'_c = 4{,}000 psi and fy=60,000f_y = 60{,}000 psi. Find the steel area it needs, and choose bars.

When φ = 0.9 no longer applies

Every design so far assumed ϕ=0.9\phi = 0.9 and then checked εt\varepsilon_t. A beam that needs a lot of steel can fail that check, and the Handbook sorts sections into three bands by their net tensile strain. That strain is measured at dtd_t, the depth to the lowest layer of bars. With one layer, dt=dd_t = d.

As dt Section c 0.003 εt Tension-controlled εt ≥ 0.005 c ≤ 0.375 dt φ = 0.90 c 0.003 εt Transition 0.002 < εt < 0.005 φ = 0.48 + 83εt c 0.003 εt Compression-controlled εt ≤ 0.002 c ≥ 0.6 dt φ = 0.65
Three strain conditions at the moment the concrete crushes

Source: Adapted from the FE Reference Handbook, Design of Reinforced Concrete Components (ACI 318-14), Strain Conditions. Values of φ added.

All three start from the same 0.003 at the top, where the concrete crushes. What differs is how deep the neutral axis sits, and so how far the steel has stretched by then.

  • Tension-controlled, εt≥0.005\varepsilon_t \ge 0.005: the steel has stretched well past yield, the beam has cracked and sagged in plain sight, and ϕ=0.90\phi = 0.90.
  • Compression-controlled, εt≤0.002\varepsilon_t \le 0.002: that's about the yield strain of grade 60 steel, 60/29,000=0.002160/29{,}000 = 0.0021, so the steel has barely yielded, if at all, when the concrete crushes. The beam fails the way a short column does, by crushing, suddenly, and it gets a column's ϕ\phi. The Handbook's Resistance Factors list gives 0.65 for members with tied reinforcement, and beams use the same values.
  • Transition, in between: ϕ\phi slides along a straight line, which the Handbook writes as ϕ=0.48+83εt\phi = 0.48 + 83\varepsilon_t. That's about 0.65 at 0.002 and 0.90 at 0.005, with the coefficients rounded.

Since c=a/β1c = a/\beta_1, and β1\beta_1 shrinks above 4,000 psi, the same block depth means a deeper neutral axis in stronger concrete. Leaving β1\beta_1 at 0.85 makes cc too small and εt\varepsilon_t look better than it is.

Worked Example A section in the transition zone
1 of 3

A beam 12 in. wide has As=6.00A_s = 6.00 in² at d=dt=20d = d_t = 20 in., with fc′=5,000f'_c = 5{,}000 psi and fy=60,000f_y = 60{,}000 psi. Classify the section and find ϕMn\phi M_n.

The most steel worth using

The bands have a simple geometry. At the edge of tension-controlled, the strain line runs from 0.003 at the top to 0.005 at the steel, so it covers 0.008 over the depth dtd_t. Similar triangles put the neutral axis at

c0.003=dt0.008⇒c=0.375 dt\frac{c}{0.003} = \frac{d_t}{0.008} \quad\Rightarrow\quad c = 0.375\,d_t

which is the Handbook's label on the tension-controlled profile, c≤0.375 dtc \le 0.375\,d_t. The same reasoning at 0.002 gives 0.6 dt0.6\,d_t. That cc fixes a=β1ca = \beta_1 c, and C=TC = T turns it into the largest steel area that keeps the section tension-controlled:

As,max⁡=0.85fc′ b β1(0.375 dt)fyA_{s,\max} = \frac{0.85f'_c\,b\,\beta_1(0.375\,d_t)}{f_y}

For the beam in the Hook, 12 in. wide with d=20d = 20 in. and 4,000 psi concrete, c=7.5c = 7.5 in., a=6.375a = 6.375 in., and As,max⁡=0.85(4)(12)(6.375)/60=4.34A_{s,\max} = 0.85(4)(12)(6.375)/60 = 4.34 in². Here's what its strength does as steel goes in:

b = 12 in., d = 20 in., f′c = 4,000 psi, fy = 60,000 psi 0 1 2 3 4 5 6 7 0 100 200 300 400 500 As (in²) Moment strength (ft-kip) tension-controlled transition 4.34 in² 6.94 in² Mn φMn 328 336
Moment strength of the Hook's beam as its steel grows

Up to 4.34 in², ϕMn\phi M_n climbs, a little less with each square inch, as the arm shrinks. Past it, MnM_n keeps climbing, but each added bar pushes the neutral axis deeper and leaves the steel less strain, so ϕ\phi falls almost as fast. ϕMn\phi M_n crawls from 328 to 336 ft-kip while the steel grows 60%, to 6.94 in² at the compression-controlled edge. The extra steel buys almost nothing, and it costs the beam its warning. So As,max⁡A_{s,\max} works as a ceiling: a beam that needs more strength than it gives needs more depth.

Checkpoint

A beam is 14 in. wide with dt=22d_t = 22 in., fc′=5,000f'_c = 5{,}000 psi, and fy=60,000f_y = 60{,}000 psi. What is the largest steel area that keeps it tension-controlled?

in²

Every change to a section moves the neutral axis through one expression. Put a=β1ca = \beta_1 c into C=TC = T:

c=Asfy0.85fc′ β1bc = \frac{A_sf_y}{0.85f'_c\,\beta_1 b}

A larger cc means a deeper neutral axis, less strain left for the steel, and less warning. A smaller cc buys strain back.

Checkpoint

A tension-controlled beam with fc′=4,000f'_c = 4{,}000 psi is redesigned, one change at a time. Which statements are true?

A.
Adding tension steel makes the beam more ductile, because more steel is stretching
B.
Adding tension steel moves the neutral axis down and lowers εt\varepsilon_t, so the beam is less ductile
C.
Raising fc′f'_c to 6,000 psi lowers β1\beta_1, so the neutral axis moves down and εt\varepsilon_t falls
D.
Raising fc′f'_c to 6,000 psi moves the neutral axis up and raises εt\varepsilon_t, even though β1\beta_1 falls

How deep must the beam be?

Sometimes the bars are chosen and the depth is the unknown. That's easier, because nothing about aa is unknown anymore: the bars fix T=AsfyT = A_sf_y, and C=TC = T fixes aa. All that's left is the lever arm the moment needs:

d−a2=MuϕAsfy⇒d=MuϕAsfy+a2d - \frac{a}{2} = \frac{M_u}{\phi A_sf_y} \quad\Rightarrow\quad d = \frac{M_u}{\phi A_sf_y} + \frac{a}{2}

For the 185 ft-kip beam with three #8 bars, T=142.2T = 142.2 kips and a=3.49a = 3.49 in. The arm must be 2,467/142.2=17.352{,}467/142.2 = 17.35 in., so d=17.35+1.74=19.09d = 17.35 + 1.74 = 19.09 in.

But dd runs to the bar centers, and the beam's outline runs past them. Below the centers sit half a bar, the stirrup that wraps the bars, and the concrete cover outside the stirrup:

b h d db/2 dstirrup cover cover dstirrup db sclear
What lies between the bars and the beam's faces
h=d+db2+dstirrup+coverh = d + \frac{d_b}{2} + d_{\text{stirrup}} + \text{cover}

Here dbd_b is the bar diameter. The Handbook gives no cover, stirrup size, or clear spacing between bars, so a question states them, with the cover usually measured to the stirrup. With 1.5 in. of cover and #4 stirrups, the beam above needs h=19.09+0.50+0.50+1.5=21.59h = 19.09 + 0.50 + 0.50 + 1.5 = 21.59 in., which a designer rounds up to 22 in.

The same stack runs across the beam. Inside the cover and the two stirrup legs, nn bars in a row need nn diameters and n−1n - 1 clear spaces between them:

bmin⁡=2(cover+dstirrup)+n db+(n−1) sclearb_{\min} = 2(\text{cover} + d_{\text{stirrup}}) + n\,d_b + (n - 1)\,s_{\text{clear}}
Checkpoint

A beam 14 in. wide must carry Mu=300M_u = 300 ft-kip with four #9 bars in one row, fc′=4,000f'_c = 4{,}000 psi, and fy=60,000f_y = 60{,}000 psi. It has #4 stirrups and 1.5 in. of cover to the stirrups. What is the smallest overall height hh that works?

in.

Two rows of bars

When the bars won't fit across the beam, some go in a second row above the first, with a clear space between the rows. Now there are two depths, and they do different jobs.

The pull TT is the sum of every bar's pull, and it acts at their centroid. Effective depth is the far end of a lever arm, so dd runs to that centroid, the area-weighted average depth of all the tension steel:

d=∑Ai di∑Aid = \frac{\sum A_i\,d_i}{\sum A_i}

Weight by area, not by bar count: a row of big bars pulls harder than a row of small ones. The second row sits higher, so dd comes out smaller than it would with every bar at the bottom. That lost lever arm is the price of the second row. The strain check still uses dtd_t, the depth to the bottom row, because that's where the steel stretches most.

dt d centroid of the tension steel
Two rows of tension steel: the lever arm ends at their centroid, the strain check at the bottom row
Worked Example A beam with two rows of bars
1 of 3

A beam is 14 in. wide and 28 in. tall, with #4 stirrups and 1.5 in. of cover. It needs six #9 bars, with 1.5 in. clear between bars in a row and 1 in. clear between rows. It has fc′=4,000f'_c = 4{,}000 psi and fy=60,000f_y = 60{,}000 psi. Lay out the bars, then find dd and ϕMn\phi M_n.

One-way slabs

A floor slab that spans one way, between parallel beams, is a very wide, shallow beam. Cut out a strip 12 in. wide, design it as a beam with b=12b = 12 in., and every answer is per foot of slab.

h d cover 12 in. s
A 12 in. strip cut from a one-way slab
Where the strip is cut is arbitrary, so count 12/s bars per strip, not the ones that happen to fall inside it.

Slab bars are given by size and center-to-center spacing ss, in inches. A 12 in. strip holds 12/s12/s bars, so the steel per foot is

As=Ab 12sA_s = A_b\,\frac{12}{s}

where AbA_b is one bar's area. #4 bars at 8 in. give 0.20(12/8)=0.300.20(12/8) = 0.30 in² per foot. A slab has no stirrups, so d=h−cover−db/2d = h - \text{cover} - d_b/2. Slab steel is light, so aa is small, often under an inch, and εt\varepsilon_t lands far above 0.005.

Checkpoint

A one-way slab is 7 in. thick, with #5 bars at 10 in. on center, 0.75 in. of cover, fc′=4,000f'_c = 4{,}000 psi, and fy=60,000f_y = 60{,}000 psi. What is ϕMn\phi M_n per foot of width?

ft-kip/ft

Recap

  • For the steel, guess the arm at 0.9d0.9d, correct it once, or solve the quadratic. Round the bar count up, and take the least area that fits.
  • With the bars chosen, d=Mu/(ϕAsfy)+a/2d = M_u/(\phi A_sf_y) + a/2, and hh adds half a bar, the stirrup, and the cover.
  • dd runs to the area-weighted centroid of the tension steel, and εt\varepsilon_t uses dtd_t, the bottom row. A slab is a 12 in. strip with As=Ab(12/s)A_s = A_b(12/s) and no stirrup.
  • ϕ\phi is 0.90 for εt≥0.005\varepsilon_t \ge 0.005, 0.65 for εt≤0.002\varepsilon_t \le 0.002, and 0.48+83εt0.48 + 83\varepsilon_t between, with β1\beta_1 reduced above 4,000 psi.
  • Tension-controlled means c≤0.375 dtc \le 0.375\,d_t, which caps the steel worth using. More steel lowers the neutral axis and costs ductility, and stronger concrete raises it.