ACI 318-19 vs ACI 318-25: the new φ limit that shrinks the P-M interaction curve

ACI 318-25 quietly shrank the P-M interaction curve. The nominal strength of a column is exactly the same as under ACI 318-19, but one new clause, Section 21.2.2.3, limits the strength reduction factor φ over a band of axial load. For the column in this post, the design moment at the tension-controlled limit falls from 149.6 to 136.6 kip-ft. Here is where the change applies, why it peaks where it does, and how to check your own columns.

PM Curve · 10 min read

One new clause, one smaller curve

Put the same column through ACI 318-19 and ACI 318-25 and almost everything comes out identical. The strain compatibility rules of Chapter 22 are the same, the 0.85 f′c stress block and β1 are the same, and so are the axial strength limits. The nominal interaction curve does not move at all.

What changes is the strength reduction factor φ. ACI 318-25 adds Section 21.2.2.3, a limit on φ for members under moment and axial compression. It only acts over a band of axial load, but inside that band it can take a noticeable bite out of the design curve.

0255075100125150175200-1000100200300400500Design moment, φMn (kip-ft)Design axial load, φPn (kip)Pn = 0.1f′cAg = 89.6 kipPn = Pn,bal = 281.3 kipPoint E (P3)φ 0.90 → 0.822φMn 149.6 → 136.6ACI 318-19ACI 318-25Design strength lost
Design P-M curves of a 14 × 16 in tied column with 2-#6 top and 2-#8 bottom bars under ACI 318-19 and ACI 318-25, from PM Curve 1.2.0. Outside the band between 0.1f′cAg and Pn,bal the two curves are the same line. Inside it, the 318-25 curve is pulled in, by the most at Point E, the tension-controlled limit.

How φ worked in ACI 318-19

In ACI 318-19, φ for moment and axial load comes from Table 21.2.2 and depends on one thing only: the net tensile strain εt in the extreme layer of tension steel when the concrete reaches 0.003.

For Grade 60 steel, εty = 60 / 29,000 = 0.00207, so a section is tension-controlled once εt reaches 0.00507. The axial load plays no direct part. A column section that reaches that strain gets φ = 0.90 whether it carries no axial load or a substantial one. On a P-M diagram the point where this happens is the tension-controlled limit, called P3 in PM Curve and labelled Point E in the figure. Below it the design curve uses φ = 0.90, and above it φ falls through the transition zone to 0.65 at the balanced point.

The new limit in ACI 318-25, Section 21.2.2.3

ACI 318-25 keeps Table 21.2.2 and adds a cap. For nonprestressed members under moment and axial compression with 0.1f′cAg ≤ Pn ≤ Pn,bal, φ may not exceed a straight line from 0.90 at 0.1f′cAg down to φcc at the balanced point:

φ ≤ 0.90 − (0.90 − φcc) × (Pn − 0.1f′cAg) / (Pn,bal − 0.1f′cAg)

040801201602002402803200.600.650.700.750.800.850.900.95Nominal axial load, Pn (kip)Strength reduction factor, φPoint E: 0.90 → 0.8220.1f′cAgPn,balband where 318-25 caps φACI 318-19 φ, from εt318-25 cap, Cl 21.2.2.3ACI 318-25 φ, the lower
φ against the nominal axial load of the same column, from PM Curve. The ACI 318-19 φ (dotted) stays at 0.90 up to Point E, then follows εt down to 0.65 at the balanced point. The 318-25 cap (dashed) runs straight from 0.90 at 0.1f′cAg = 89.6 kip to 0.65 at Pn,bal = 281.3 kip. ACI 318-25 uses whichever is lower (solid).

The figure shows why the effect is concentrated. Below 0.1f′cAg the cap does not apply. At the balanced point both rules give 0.65. In between, the cap only bites where the strain-based φ sits above the straight line, and the gap is widest at Point E, where the 318-19 φ is still 0.90 but the cap has already fallen.

The example column

Section14 × 16 in, tied, bending about the 16 in depth
Materialsf′c = 4,000 psi, fy = 60 ksi, Es = 29,000 ksi
Bars2-#6 top (d′ = 2.25 in), 2-#8 bottom (dt = 13.625 in), #3 ties, 1.5 in clear cover
AreasAg = 224 in², Ast = 0.88 + 1.58 = 2.46 in²
Band limits0.1f′cAg = 0.1 × 4 × 224 = 89.6 kip; Pn,bal = 281.3 kip

Point E, step by step

Point E is the tension-controlled limit, where εt = εty + 0.003 = 0.00507. Its nominal strength is the same in both editions:

Neutral axis. c = 0.003 × 13.625 / (0.003 + 0.00507) = 5.066 in, and a = β1c = 0.85 × 5.066 = 4.306 in.

Concrete. Cc = 0.85 × 4 × 14 × 4.306 = 205.0 kip.

Top bars. ε = 0.003 × (5.066 − 2.25) / 5.066 = 0.00167, so fs = 48.4 ksi. They sit inside the stress block, so the displaced concrete is deducted: (48.4 − 3.4) × 0.88 = 39.6 kip in compression.

Bottom bars. They have yielded: 60 × 1.58 = 94.8 kip in tension.

Axial strength. Pn = 205.0 + 39.6 − 94.8 = 149.7 kip.

Moment, about the plastic centroid 8.240 in below the top: Mn = 205.0 × 6.088 + 39.6 × 5.990 + 94.8 × 5.385 = 1,995 kip-in = 166.3 kip-ft.

Then the two editions part ways:

ACI 318-19. εt = 0.00507, tension-controlled, so φ = 0.90. φMn = 0.90 × 166.3 = 149.6 kip-ft at φPn = 134.7 kip.

ACI 318-25. Pn = 149.7 kip is between 89.6 and 281.3 kip, so the cap applies: φ ≤ 0.90 − 0.25 × (149.7 − 89.6) / (281.3 − 89.6) = 0.90 − 0.25 × 0.313 = 0.822. That is below 0.90, so φ = 0.822 and φMn = 0.822 × 166.3 = 136.6 kip-ft at φPn = 123.0 kip.

The design strength at Point E drops by 13.0 kip-ft, 8.7%. Because φ multiplies both Pn and Mn, the point moves straight in toward the origin, which is why the lost area in the first figure is a wedge pointing at Point E.

Through the band, point by point

The same comparison at other axial loads on this column, at the same nominal point in both editions:

Pn (kip)εtφ, 318-19φ, 318-25φMn, 318-19 (kip-ft)φMn, 318-25 (kip-ft)Change
60 (below the band)0.009390.9000.900116.3116.30
89.6 (0.1f′cAg)0.007630.9000.900128.3128.30
1200.006190.9000.860139.6133.5−4.4%
149.7 (Point E)0.005070.9000.822149.6136.6−8.7%
1800.004150.8230.782145.1137.8−5.0%
2200.003190.7440.730139.2136.6−1.8%
2600.002430.6800.678132.8132.4−0.3%
281.3 (Pn,bal)0.002070.6500.650128.9128.90

The loss grows from zero at 0.1f′cAg to its peak at Point E, then shrinks back to zero at the balanced point, where the transition φ of ACI 318-19 has caught up with the cap. Everything above the balanced point is compression-controlled in both editions and does not change.

Worked check: a column that no longer passes

Take a factored load on this column of Pu = 130 kip with Mu = 140 kip-ft, and read the design moment capacity at φPn = 130 kip from each curve:

ACI 318-19. φMn = 147.8 kip-ft. 140 / 147.8 = 0.95. Passes.

ACI 318-25. φMn = 137.3 kip-ft. 140 / 137.3 = 1.02. Does not pass.

Same column, same load, same nominal strength. Under the new edition this column needs more steel or a larger section. A design that worked with a few percent to spare under 318-19 in this range of axial load is worth checking again.

Does every column lose strength?

No. The loss depends on where Point E falls relative to the band. Three columns, checked in PM Curve across the whole curve:

ColumnBand (kip)Point E, Pn (kip)φ at Point ELargest loss
14 × 16 in tied, 2-#6 top, 2-#8 bottom89.6 to 281.3149.70.900 → 0.8228.7%
16 × 16 in tied, 8-#8 (3 top, 2 middle, 3 bottom)102.4 to 365.9113.90.900 → 0.8891.2%
18 in round spiral, 8-#8101.8 to 366.042.90.900, unchangednone

Point E tends to sit higher when the compression side of the section can carry a lot before the tension steel reaches 0.00507: lighter tension steel, deeper sections and higher f′c all push it up. The quickest way to know for your column is to run both editions and compare.

What does not change

What it means in practice

Comparing the two editions in PM Curve

PM Curve 1.2.0 has a Design standard input at the top: ACI 318-19 (the default) or ACI 318-25. Choose 318-25 and the design curve, the seven key points and the results use the limited φ, with extra points added at 0.1f′cAg and Pn,bal so the bends in the curve are sharp. The report names the standard, and for ACI 318-25 it shows the limit, its 0.1f′cAg, Pn,bal and φcc, and the φ used at each point. With Compare sections you can put the same column on the chart twice, once under each edition, and see the difference directly. Models saved before 1.2.0 open as ACI 318-19, so their results do not change.

As always, this is the strength of the cross-section. Slenderness effects have to be added to the moments first, and the column checked for every load combination.

FAQ

Frequently asked questions

What changed for columns between ACI 318-19 and ACI 318-25?

ACI 318-25 adds Section 21.2.2.3. For nonprestressed members under moment and axial compression with 0.1 f'c Ag <= Pn <= Pn,bal, phi may not exceed a straight line from 0.90 at 0.1 f'c Ag to 0.65 (tied) or 0.75 (spiral) at the balanced point. The nominal P-M curve is unchanged; only the design curve can get smaller.

Why does the design strength drop the most at the tension-controlled limit?

Under ACI 318-19, phi is still 0.90 at the tension-controlled limit (Point E), while the new cap has already fallen below 0.90 if Point E is inside the band. Above Point E the strain-based phi itself drops, so the gap closes again, reaching zero at the balanced point.

Does the ACI 318-25 phi limit affect beams?

Not beams without axial compression. The limit applies only when the nominal axial load is at least 0.1 f'c Ag, and at Pn = 0 (pure bending) it does not apply.

Does every column lose capacity under ACI 318-25?

No. If the tension-controlled limit of the section falls below 0.1 f'c Ag, as in many spiral columns, the strain-based phi is already below the cap and nothing changes. The loss is largest when the tension-controlled limit sits high in the band between 0.1 f'c Ag and the balanced load.

How do I apply the ACI 318-25 phi limit by hand?

Compute 0.1 f'c Ag and the balanced axial strength Pn,bal. If Pn is between them, compute the cap 0.90 - (0.90 - phi_cc)(Pn - 0.1 f'c Ag)/(Pn,bal - 0.1 f'c Ag) and use the lower of it and the phi from Table 21.2.2. If Pn,bal is below 0.1 f'c Ag the limit does not apply.

Keep reading

More from the blog

For use by qualified engineers. All results must be checked independently before use in any design. See the Terms of Use.