When “Good Enough” Isn’t: Rethinking Lap Splices in RC Structural Walls

Reinforced concrete (RC) walls are often the quiet workhorses of seismic design—expected to flex, dissipate energy, and keep buildings standing when earthquakes strike. A seemingly small detailing choice, however, can have an outsized impact on their performance: the use and placement of lap splices.

A recent experimental study On estimating the drift capacity of reinforced concrete walls with lap splices at their bases sheds new light on this issue, focusing on staggered lap splices in RC structural walls. While current design standards permit their use—even near critical regions like the base of walls—the findings suggest that this long-standing practice may not be as safe as previously assumed.

Let’s unpack what’s going on, starting simple and gradually diving into the mechanics.

The Basics: What Are Lap Splices and Why Do They Matter?

In reinforced concrete, steel bars carry tension while concrete handles compression. When a bar needs to be extended, engineers often overlap two bars—this overlap is called a lap splice. The idea is straightforward: forces transfer from one bar to the other through the surrounding concrete via bond stress.

There are two common configurations:

Non-staggered splices: all bars splice at the same section
Staggered splices: splices are offset so not all bars overlap at once

Modern standards (like NZS 3101) encourage staggering to avoid creating a weak plane. On paper, this seems like a sensible compromise between constructability and performance.

But earthquakes don’t care about convenience—they expose the weakest link.

The Problem: Earthquake Damage Tells a Different Story

Field observations from major earthquakes (Chile 2010, Taiwan 2016, Turkey 2023) consistently show severe damage concentrated at lap splice regions in RC walls.

This isn’t just anecdotal. It reflects a deeper issue:

Lap splices may be strong enough to carry load—but not ductile enough to sustain large deformations.

In seismic engineering, ductility—the ability to deform without losing strength—is critical. Structures don’t just need to resist forces; they need to absorb and survive them.

The Experiment: Full-Scale Testing of RC Walls

To investigate this, researchers tested two full-scale RC wall specimens:

Height: ~8.1 m
Shear span: 6.45 m
Flexure-dominated behavior (typical for slender walls)
Staggered lap splices near the foundation (a critical region)

The only difference between the two walls (W1 and W2) was how the splices were arranged.

Both were subjected to cyclic lateral loading to simulate earthquake effects.

What Happened: Early Failure and Limited Drift Capacity

Here’s where things get interesting.

Both walls yielded at ~0.7% drift
Splice failure began shortly after:
- W1: ~1.0% drift
- W2: ~1.25% drift
Ultimate drift capacities:
- W1: 1.6%
- W2: 1.3%

These are low values for modern seismic design.

Why is this concerning?

Because expected earthquake demands are higher.

The Reality Check: Demand vs Capacity

Using established seismic estimation methods, a plausible drift demand for RC wall buildings is about:

~2.3% drift

Compare that to the tested capacities:

Wall	Drift Capacity
W1	1.6%
W2	1.3%

Both fall short.

This means that, under realistic earthquake conditions, these walls would likely experience significant damage or failure.

The Deeper Issue: Design Standards Overestimate Performance

Here’s where the study becomes especially critical for engineers.

Two common approaches were used to predict drift capacity:

Design standard (NZS 3101) → ~3.2%
Assessment method (C5 guidelines) → ~2.6%

Both significantly overestimated actual performance.

Why the discrepancy?

Because both methods assume:

Strains are evenly distributed along a plastic hinge region
Bond between steel and concrete degrades gradually

But the experiments showed something very different:

Deformations localized at splice regions, and bond failure occurred suddenly and early.

In other words, the real behavior is far more brittle and concentrated than models assume.

Failure Mechanism: What Actually Goes Wrong?

The sequence of failure was consistent:

Outer splices fail first (bond failure)
Load shifts inward
Inner splices fail
Rapid degradation in strength

This cascading failure is dangerous because:

It reduces redundancy
It accelerates strength loss
It limits energy dissipation

A Misleading Safety Net: “Rocking” Behavior

There’s a common argument that after splice failure, walls may still survive through rocking (rotating at the base).

The study challenges this assumption.

In one specimen:

The “rocking plane” formed 1 meter above the foundation, not at the base

This unpredictability means:

The structure may not behave as intended
Load paths become uncertain
Collapse mechanisms are harder to control

Conclusion: rocking is not a reliable fallback mechanism.

Key Insight: Location Matters More Than Configuration

The study reinforces a critical principle:

It’s not just how you splice—it’s where you splice.

Even properly detailed, code-compliant staggered splices performed poorly when placed near:

Plastic hinge regions
Zones of expected yielding

Practical Implications for Design

The findings point toward a clear shift in best practice:

1. Avoid lap splices in critical regions

Especially near:

Foundations
Plastic hinge zones

2. Prefer continuous reinforcement

This eliminates bond transfer uncertainty altogether.

3. Use alternative connection methods

If splicing is unavoidable:

Mechanical couplers
Welded connections

4. Rethink assessment models

Existing buildings with lap splices may be less ductile than assumed, requiring:

More conservative evaluations
Possible retrofitting

How Do These Findings Stack Up Against ACI and Eurocode?

The experimental results expose a core issue: modern design standards may overestimate deformation capacity when lap splices are present in critical regions. To understand the implications, we compare how different codes treat lap splices in seismic RC walls—from philosophy to mechanics.

1) Big Picture: Philosophical Differences

NZS 3101 (Context of the Study)

Allows staggered lap splices in plastic hinge regions
Focuses on strength development (yielding)
Limited explicit treatment of post-yield deformability

👉 Result: Code-compliant walls failed below expected drift capacity

American Concrete Institute — ACI 318

Restricts lap splices in seismic hinge zones
Requires:
- No splices in boundary elements where yielding is expected
- Use of mechanical couplers or continuous bars
Emphasizes:
- Ductility
- Confinement
- Controlled plastic hinging

👉 Philosophy: Avoid the problem entirely

Eurocode 8 (EN 1998-1)

Discourages or prohibits splices in critical regions
Requires:
- Splices placed away from plastic hinges
- Enhanced detailing if unavoidable
Emphasizes:
- Capacity design
- Ductile failure modes

👉 Philosophy: Preserve ductility through detailing control

2) Critical Comparison: Splice Location Rules

Code	Splices in Plastic Hinge Region?	Approach
NZS 3101	Allowed (if staggered)	Strength-based
ACI 318	Prohibited	Ductility-first
Eurocode 8	Generally prohibited	Capacity design

Insight: Experimental results validate the stricter ACI and Eurocode stance.

3) Drift Capacity: Prediction vs Reality

From experiments:

Measured: 1.3%–1.6%
NZS prediction: ~3.2%
C5 assessment: ~2.6%

ACI / Eurocode approach:

Rather than modeling splice behavior in hinge zones, they:

Eliminate splices from those regions
Assume:
- Continuous reinforcement
- Stable plastic hinging
- Distributed curvature

This avoids the strain localization error seen in NZ predictions.

4) Core Technical Issue: Bond vs Ductility

NZ assumption:

Splices develop yield strength
Bond degrades gradually
Strains are distributed

Experimental reality:

Bond failure at ~1% drift
Strains localize
Rapid strength degradation

ACI / Eurocode response:

Treat lap splices as potentially brittle
Avoid relying on:
- Bond under cyclic loading
- Post-yield splice behavior

Recognizes bond-slip as non-ductile and unreliable

5) Plastic Hinge Modeling Assumptions

NZ / C5:

Drift ∼ θₚ · Lₚ
Assumes uniform strain distribution

Observed behavior:

Deformation concentrates at splice zones
Effective hinge length shrinks
Leads to:
- Higher local strains
- Early failure

ACI / Eurocode implication:

By avoiding splices in hinge zones:
- Plastic hinges form in continuous steel
- Modeling assumptions remain valid

6) Failure Modes: Controlled vs Uncontrolled

NZ-compliant walls:

Governed by:
- Bond failure
- Bar slip
Progressive collapse:
- Outer → inner splice failure

Brittle and uncontrolled

ACI / Eurocode design:

Governed by:
- Steel yielding
- Confined concrete crushing
Behavior:
- Gradual
- Energy dissipating

Ductile and controlled

7) The “Rocking” Misconception

Assumption:

Walls may survive via rocking after splice failure

Experimental finding:

Rocking plane formed above foundation, not at base

ACI / Eurocode philosophy:

Do not rely on secondary mechanisms
Ensure behavior is:
- Predictable
- Designed

Aligns with capacity design principles

8) Bottom Line: Who Got It Right?

The experimental evidence strongly supports:

ACI 318
Eurocode 8

Why?

They:

Anticipate bond-related brittleness
Eliminate reliance on splices in critical zones
Preserve ductile mechanisms

Final Engineering Takeaway

This comparison highlights a gap between code assumptions and physical behavior:

NZS approach: permits a detail that behaves non-ductile under cyclic loading
ACI / Eurocode: eliminate that uncertainty

Bond-controlled mechanisms cannot be relied upon in plastic hinge regions—even if they satisfy strength requirements.

Final Takeaway

Lap splices have been used for over a century, and in many contexts, they work just fine. But in seismic applications—particularly in RC structural walls—their behavior is more nuanced and potentially hazardous than design codes currently reflect.

This study makes one thing clear:

Code compliance does not guarantee seismic resilience.

As performance-based design continues to evolve, details like lap splice placement—once considered routine—deserve much closer scrutiny.

When “Good Enough” Isn’t: Rethinking Lap Splices in RC Structural Walls

Executive Summary