Effective Moment of Inertia for Cracked Concrete Walls

Look up the effective moment of inertia ratio \(I_e/I_g\) for a concrete shear wall and every source gives a different number. Each one also depends on the cracking state and loading condition each one assumes. CSA A23.3 Clause 10.14.1.2 sets the Canadian values. The NZ Concrete Standard commentary suggests 0.25 for walls with no axial load. ASCE 41-23 recommends 0.35 for cracked walls. CSA A23.3 uses 0.7 for uncracked walls and 0.35 for cracked walls. FEMA 356 goes up to 0.8 for previously uncracked walls.

So which one is right? It depends.

Why can’t we just agree on a number?

The short answer: the effective stiffness of a cracked concrete wall isn’t constant. It changes based on several factors, and different codes have different target applications, cracking conditions, and hazard levels in mind.

When you model a concrete wall in your analysis software, you need to input an \(EI\) value. But concrete doesn’t behave linearly. It cracks, the reinforcement yields, and the stiffness degrades as you push further into the inelastic range. A single \(EI\) value is, at best, a reasonable approximation for a specific load level and performance objective.

Here’s what different guidelines recommend:

Note that these values are not directly comparable. They correspond to different cracking states, axial load assumptions, and target performance levels. Comparing the NZ value for zero axial load (0.25) with the FEMA 356 uncracked value (0.8) as a “factor of three range” is misleading without this context. When you compare like-for-like conditions, the spread narrows considerably.

If you’re calculating periods, drifts, or force distributions, the choice still matters. Pick the value that matches your wall’s actual condition and your analysis purpose. Don’t just pick the lowest or highest number from the table.

The moment-curvature reality

To understand where these numbers come from, think about the actual moment-curvature behaviour of a concrete wall section.

Before cracking, the wall responds with its gross section stiffness (\(EI_g\)). Once the concrete cracks, the stiffness drops, sometimes significantly. Push further and you’ll see reinforcement yielding, cover spalling, and eventually a plateau or degradation in capacity.

Typical layout of \(M-\phi\) diagram. Source.

The effective moment of inertia concept tries to capture this behaviour with a single value. The typical approach is to draw a secant line from the origin to some target point on the moment-curvature curve. That point is usually the ultimate capacity, at a curvature corresponding to your design drift limit.

Where you draw that secant line determines your effective \(EI\).

If your target performance is serviceability (limited cracking, essentially elastic response), you draw the secant to a point early on the curve. That gives you a higher effective stiffness, maybe 0.7 or 0.8 of gross.

If you’re designing for life safety under a major earthquake and expect significant inelastic deformation, you draw the secant further along the curve. That gives you a lower effective stiffness, perhaps 0.25 to 0.5 of gross.

The “correct” \(I_e/I_g\) ratio depends on the performance level you’re targeting. There is no single right answer for all situations.

The axial load effect

One factor that significantly influences effective stiffness is axial compression. This often gets overlooked when engineers grab a single \(I_e/I_g\) value from the code without thinking twice.

Higher axial compression delays cracking. It pre-compresses the concrete, meaning you need more tension from bending before cracks form. The result is a stiffer response up to higher load levels.

The NZ Standard commentary acknowledges this. It recommends different values depending on whether the wall carries axial load. You use 0.25 for walls with no axial force and 0.35 for walls with axial compression. Research shows that walls with different axial load ratios (\(\frac{P}{A_g \cdot f'_c}\)) exhibit markedly different moment-curvature responses:

• At 5% axial load ratio: lower stiffness, earlier cracking
• At 10% axial load ratio: moderate stiffness improvement
• At 15-20% axial load ratio: noticeably higher effective stiffness

This is one reason the CSA A23.3 uncracked value of \(0.7I_g\) can be reasonable for many mid-rise concrete buildings. Their walls often carry moderate axial loads and may not crack much under service or moderate seismic loads. But say you’re analyzing a slender wall in a low-rise building with minimal axial load, and it’s expected to crack under design loading. Then 0.7 is unconservative for predicting drifts. Use the cracked value of \(0.35I_g\) or do a more detailed analysis.

If your wall has unusually high or low axial load, don’t blindly apply a single factor. Consider whether your situation warrants adjustment, and whether the wall will be cracked or uncracked under the relevant load level.

Boundary reinforcement matters too

The amount of reinforcement at the wall boundaries (those heavily reinforced zones at the ends of the wall) also affects effective stiffness.

More boundary reinforcement means:

• Higher ultimate moment capacity
• A different slope for the secant line to ultimate
• Generally higher effective \(EI\) for a given target curvature

Walls with minimum boundary reinforcement will show more stiffness degradation as they approach capacity compared to walls with generous boundary steel.

This is part of why performance-based seismic design guidelines often provide different stiffness modifiers depending on expected ductility demand. A wall expected to remain essentially elastic can use a higher effective stiffness than one expected to undergo significant plastic hinging.

What does CSA A23.3 actually say?

CSA A23.3, in Clause 10.14.1.2 (Table 10.14.1.2), specifies reduced stiffness values for concrete members used in elastic analysis with factored loads. For walls specifically, the code provides two values:

• Uncracked walls: \(0.7I_g\)
• Cracked walls: \(0.35I_g\)

The distinction depends on whether the wall is expected to crack under the applied loading, typically assessed by comparing the factored moment against the cracking moment. The cracking moment is based on the modulus of rupture.

What often gets missed: these values are intended for a specific context - force-based analysis where you’re looking at member forces under factored loads. They are not necessarily appropriate for:

• Drift calculations at serviceability limit states (where gross section properties are typically used)
• Period calculations for dynamic analysis
• Nonlinear analysis where you’re tracking stiffness degradation explicitly

For period calculations using elastic analysis, \(0.7I_g\) (uncracked) will give you a shorter period than the cracked value of \(0.35I_g\) would. To know whether this is conservative, think about where your building sits on the design spectrum. On the constant acceleration plateau, a shorter period gives you the same spectral acceleration. No penalty, but no benefit either. Most mid- to high-rise wall buildings sit on the descending branch. There, a shorter period moves you up the curve to a higher spectral acceleration and therefore a higher base shear. So assuming uncracked stiffness when the wall is actually cracked is generally conservative for force calculations. You’ll overestimate forces. For drift predictions, though, using the uncracked stiffness on a cracked wall will underestimate drift. That is unconservative.

When checking your analysis approach, ask yourself: “What am I using this stiffness for, and is my wall cracked or uncracked under this load level?” The answers should guide your choice of reduction factor.

A practical framework for choosing \(I_e/I_g\)

Rather than memorizing specific values, it helps to have a mental framework.

For force-controlled design (strength checks), higher \(I_e/I_g\) values (0.6-0.8) are often conservative. The resulting shorter period either keeps you on the constant acceleration plateau (same base shear) or moves you up the descending branch (higher base shear). Either way, you don’t underestimate forces. The CSA A23.3 uncracked value of 0.7 is reasonable for walls that aren’t expected to develop significant flexural cracking under the applied loads.

For drift-controlled design (deformation checks), lower \(I_e/I_g\) values (0.35-0.5) may be more appropriate. You want to capture the flexibility of the cracked system. Underestimating drift is unconservative. CSA’s cracked value of 0.35 - which aligns with ASCE 41’s cracked wall value - is a reasonable starting point.

For special situations: high axial load may justify increasing effective stiffness above the default cracked value. Slender walls with low axial load may need a decrease. Performance-based design should match stiffness to the target performance level - modern standards like ASCE 41-23 provide detailed guidance on this.

What you shouldn’t do:

• Ignore the cracked vs. uncracked distinction and always use 0.7
• Ignore axial load effects when they’re significant
• Apply residential building values to a high-rise core wall or vice versa

What requires judgment: running sensitivity analyses with different stiffness values and enveloping results can be a sound approach. It only works when you understand which stiffness applies to which check and why. Mindless enveloping without understanding the underlying behaviour adds complexity without insight. The same principle extends beyond stiffness. Understanding how much margin is embedded in your CSA A23.3 code equations is the foundation of confident RC design.

Why this matters for your projects

Getting the effective stiffness wrong has real consequences:

1. Force distribution: In buildings with mixed lateral systems (walls and frames), relative stiffness determines how forces split between elements. Overestimate wall stiffness and you might underdesign connecting frames.

2. Drift calculations: Underestimating flexibility means underestimating drift. You might miss a P-delta problem or undersize seismic joints.

3. Period and dynamic response: Your building period affects spectral acceleration, base shear, and expected response. It also affects torsional behaviour in irregular buildings.

4. Capacity design: If you’re doing capacity design for non-ductile elements, the stiffness assumption affects the demand you calculate.

A common mistake is using the same stiffness value for all analysis purposes without considering whether it’s appropriate for what you’re checking and whether your wall is cracked under the relevant load level.

Putting it into practice

Here’s a reasonable approach when starting a concrete building analysis:

1. Start with the code provisions. For CSA A23.3 applications, determine whether your walls are expected to be cracked or uncracked under the applied factored loads. Use \(0.7I_g\) for uncracked walls and \(0.35I_g\) for cracked walls per Table 10.14.1.2. This is a defensible starting point.

2. Check your axial loads. Are your walls carrying significantly more or less axial load than typical? Walls with high axial load ratios may remain uncracked at higher load levels, supporting the use of the higher stiffness value. Walls with low axial load will crack earlier.

3. Consider your analysis purpose. If you’re primarily checking forces, using higher stiffness (uncracked values) is generally conservative. It shortens the period and either maintains or increases base shear, depending on where you sit on the spectrum. If you’re checking drifts, use the cracked stiffness or run a sensitivity check to understand the impact.

4. Document your assumptions. Whatever value you use, write down why. “Used \(0.7I_g\) per CSA A23.3 Table 10.14.1.2 for uncracked walls. Wall axial load ratio approximately 10%. Cracking check confirms wall remains uncracked under factored loads. Primary concern is force distribution to frames.” That is much better than just “used \(0.7I_g\).”

5. Be consistent within each analysis. Don’t mix-and-match different stiffness assumptions for different walls without good reason (e.g., some walls genuinely cracked and others not).

Next steps

• Review your current project’s stiffness assumptions. Do they match the analysis purpose? Have you correctly assessed which walls are cracked vs. uncracked?
• Try running sensitivity analyses with \(I_e/I_g\) values at the high and low end of the typical range to see how your results change
• If you’re modelling walls as shell elements, make sure you’re also choosing the right element type for your analysis objectives

The range of code values for effective moment of inertia isn’t a sign of confusion in the profession. It’s a recognition that the “right” value depends on context - cracking state, axial load, performance objective, and analysis purpose. Understanding that context makes you a better engineer.

Before the next run, write down the cracking state and analysis purpose you assumed, and check that the stiffness value you picked matches both.