Canadian Floor Vibration — CSA S16 Design Guide

Floor vibration serviceability is an increasingly important design consideration for Canadian steel-framed buildings. Longer spans, open-plan layouts, lighter floor systems, and higher-strength steels all contribute to floors that may satisfy strength and deflection limits but feel "bouncy" or uncomfortable to occupants.

CSA S16:19 Commentary K provides detailed guidance for assessing human comfort under walking-induced vibration. The commentary references the AISC Design Guide 11 methodology and adapts it for Canadian practice with metric units and CSA S16 limit states design conventions. NBCC 2020 Sentence 4.1.5.3(2) also requires that floor systems be designed to limit vibration to acceptable levels, though it does not prescribe specific numerical criteria.

CSA S16 Commentary K — Vibration Assessment Method

Commentary K adopts the AISC Design Guide 11 resonant response method. The peak acceleration from a walking excitation is evaluated and compared to an acceleration tolerance limit based on occupancy:

a_p / g = P_0 x exp(-0.35 x f_n) / (beta x W)

Where:

• a_p/g = peak floor acceleration as a fraction of gravity
• P_0 = walking force amplitude = 0.29 kN (65 lb) — represents a single footfall
• f_n = fundamental natural frequency of the floor system (Hz)
• beta = modal damping ratio (fraction of critical damping)
• W = effective weight of the floor panel participating in the vibration mode (kN)

The floor is acceptable when a_p/g <= a_o/g, where a_o/g is the acceleration tolerance limit for the occupancy.

Acceleration Tolerance Limits (a_o/g)

Commentary K recommends the following acceleration limits for human comfort:

The residential limit of 0.2% g is notably more stringent than the 0.5% g office limit. This difference is critical for residential conversions of former office buildings, where the existing floor system may not satisfy the tighter criteria.

Damping Ratios for Canadian Construction

The damping ratio beta depends on the level of fit-out and non-structural components:

Commentary K recommends beta = 0.03 for typical Canadian office fit-out with ceiling, ductwork, and sprinklers but open-plan layout (few full-height partitions). This is slightly higher than AISC DG11's default of 0.025, reflecting Canadian practice of more extensive ceiling and mechanical systems.

Natural Frequency Calculation

The fundamental natural frequency of a simply supported beam or one-way floor system is estimated from the midspan deflection under sustained load:

f_n = 0.18 x sqrt(g / delta_j)

Where:

• g = 9,810 mm/s^2 (gravitational acceleration)
• delta_j = beam midspan deflection under sustained load (mm)
• f_n = fundamental frequency (Hz)

For a two-way system (beams supported by girders), the combined frequency is:

f_n = 0.18 x sqrt(g / (delta_j + delta_g))

Where delta_j = beam midspan deflection and delta_g = girder midspan deflection at the beam connection point.

The "sustained load" for vibration assessment includes:

• Slab self-weight (full dead load)
• Superimposed dead load (full SDL)
• 10-25% of the design live load (representing "normally present" live load — not the full code value)

Using the full code live load (e.g., 4.8 kPa for office) artificially increases the mass, lowers the computed frequency, and produces an unconservative result — the floor appears more acceptable than it actually is.

Commentary K Minimum Frequency Recommendations

These are guidance values. The acceleration criterion (a_p/g <= a_o/g) is the governing check.

Worked Example — Vibration Assessment for a Composite Floor

Given: A composite steel floor in a Toronto office building. W410x46 composite beams (W16x31 equivalent) at 3.0 m spacing, span L = 10.0 m. W610x92 composite girders (W24x62 equivalent), span L_g = 9.0 m. 75 mm lightweight concrete slab (unit weight = 1,840 kg/m^3, f'_c = 25 MPa) on 38 mm metal deck (total slab depth = 113 mm). SDL = 1.0 kPa, partition load = 0.7 kPa. Office occupancy (a_o/g = 0.5%). Ceiling and ductwork below, open-plan (beta = 0.03).

Step 1 — Sustained load on beam:

Slab self-weight: 0.113 m x 18.4 kN/m^3 x 0.70 (deck rib void factor) = 1.46 kPa Total sustained load w = (1.46 + 1.0 + 0.7 + 0.10 x 4.8) x 3.0 m = (3.16 + 0.48) x 3.0 = 10.92 kN/m = 0.01092 kN/mm

(Using 10% of the 4.8 kPa office live load as "normally present")

Step 2 — Composite section properties (W410x46):

A_steel = 5,890 mm^2, I_steel = 156 x 10^6 mm^4, d = 403 mm Effective slab width = min(L/4, s) = min(2,500, 3,000) = 2,500 mm Modular ratio n = E_steel / E_concrete = 200,000 / (4,500 x sqrt(25)) = 200,000 / 22,500 = 8.9 Transformed slab width = 2,500 / 8.9 = 281 mm I_composite = I_steel + A_steel x (d/2 + t_slab/2 - y_bar)^2 + b_trans x t_slab^3 / 12 + b_trans x t_slab x (y_bar - t_slab/2)^2

For a W410x46 with 113 mm slab, I_composite ≈ 420 x 10^6 mm^4 (fully composite, transformed section).

Step 3 — Beam midspan deflection under sustained load:

delta_j = 5 x w x L^4 / (384 x E x I) delta_j = 5 x 0.01092 x (10,000)^4 / (384 x 200,000 x 420 x 10^6) delta_j = 5 x 0.01092 x 1.0 x 10^16 / (3.226 x 10^13) delta_j = 5.46 x 10^14 / 3.226 x 10^13 delta_j = 16.9 mm

Step 4 — Beam natural frequency:

f_n,beam = 0.18 x sqrt(9,810 / 16.9) = 0.18 x sqrt(580.5) = 0.18 x 24.1 = 4.34 Hz

Step 5 — Girder frequency:

Girder sustained load from beam reactions at third-points. Each beam reaction = 10.92 x 10.0 / 2 = 54.6 kN. At two third-points on a 9.0 m girder span, the equivalent uniform load for deflection = 2 x 54.6 x 3.0 x (9,000^2 - 4 x 3,000^2) / (9,000^3) x ...

Per standard beam formula: delta at midspan for two equal loads at third-points = 23 x P x L^3 / (648 x E x I)

Girder I_composite ≈ 1,200 x 10^6 mm^4 delta_g = 23 x 54,600 x (9,000)^3 / (648 x 200,000 x 1,200 x 10^6) delta_g = 23 x 54,600 x 7.29 x 10^11 / (1.555 x 10^14) delta_g = 9.15 x 10^17 / 1.555 x 10^14 delta_g = 5.9 mm

f_n,girder = 0.18 x sqrt(9,810 / 5.9) = 0.18 x sqrt(1,663) = 0.18 x 40.8 = 7.34 Hz

Step 6 — Combined frequency:

f_n = 0.18 x sqrt(9,810 / (16.9 + 5.9)) = 0.18 x sqrt(430.3) = 0.18 x 20.7 = 3.73 Hz

The combined frequency (3.73 Hz) is significantly lower than the beam-alone frequency (4.34 Hz). This highlights the importance of checking the combined system — checking only the beam would overestimate the frequency by 16%.

Step 7 — Effective panel weight (simplified method):

Effective panel width B = 0.6 x L = 0.6 x 10.0 = 6.0 m (typical for 3.0 m beam spacing with LW concrete slab)

W = sustained load per unit area x B x L = (3.16 + 0.48) kPa x 6.0 x 10.0 = 218.4 kN

Step 8 — Peak acceleration check:

beta = 0.03 (office with ceiling and ductwork)

a_p/g = 0.29 x exp(-0.35 x 3.73) / (0.03 x 218.4) a_p/g = 0.29 x exp(-1.306) / 6.55 a_p/g = 0.29 x 0.271 / 6.55 a_p/g = 0.0786 / 6.55 a_p/g = 0.0120 = 1.20% g

Result: a_p/g = 1.20% g > a_o/g = 0.50% g — FAILS.

Step 9 — Remedial options:

Option A — Increase beam depth to W530x74 (W21x50 equivalent, I_composite ≈ 720 x 10^6 mm^4): delta_j = 9.9 mm, f_n,combined = 4.53 Hz a_p/g = 0.29 x exp(-0.35 x 4.53) / (0.03 x 210) = 0.66% g — still fails.

Option B — Reduced beam spacing from 3.0 m to 2.4 m with lighter beams W360x39: Total sustained load per beam reduces. Effective panel width increases. a_p/g = approximately 0.55% g — marginal.

Option C — Add full-height partitions (beta = 0.05): Repeating with beta = 0.05: a_p/g = 1.20% g x (0.03/0.05) = 0.72% g — still above 0.50%.

Option D — Upsize to W530x82 (W21x55 equivalent) with beta = 0.03: f_n = 4.8 Hz, a_p/g ≈ 0.48% g — PASS marginally.

This example illustrates that 10 m spans in office construction with lightweight concrete are at the practical limit for vibration serviceability. Deeper beams, tighter spacing, and fit-out damping all contribute incrementally. A combined approach of W530x74 at 2.4 m spacing with full partitions (beta = 0.05) would produce a robust pass.

Commentary K Comparison with Other Standards

Commentary K is closely aligned with AISC DG11. The primary differences are the use of SI units and the marginally higher recommended damping for Canadian office construction.

Design Strategies for Vibration Control

When a floor fails the vibration check, the following strategies are available, ranked by cost-effectiveness:

1. Increase damping (beta): The most cost-effective measure. Specifying full-height partitions increases beta from 0.03 to 0.04-0.05, reducing a_p/g by 25-40%. This requires architectural coordination.

2. Increase beam stiffness (higher I): Going from W410 to W530 typically increases I by 60-80% and raises f_n by 15-25%. However, the exponential term in the acceleration equation means the a_p/g benefit is 30-50%.

3. Reduce beam spacing: Closer spacing distributes the load over more beams, increasing effective weight W. Changing from 3.0 m to 2.4 m spacing reduces load per beam by 20% and increases effective panel width.

4. Increase girder stiffness: If the girder frequency controls, upsizing a single girder can fix multiple bays simultaneously. This is often more economical than upsizing every beam.

5. Use normal-weight concrete: NWC (2,400 kg/m^3) vs. LWC (1,840 kg/m^3) increases slab mass by 30%, increasing W and damping. The trade-off is higher dead load on columns and foundations.

6. Tuned mass dampers (TMDs): Effective for severe cases. A TMD adds a counter-oscillating mass (typically 1-2% of the modal mass) tuned to the floor frequency. Costs $5,000-$15,000 per bay but avoids structural modifications.

7. Stiffening beams: Adding cover plates, increasing flange thickness, or replacing with deeper sections all provide incremental gains but at increasing cost per unit of improvement.

Common Vibration Problems in Canadian Construction

Design Resources

• Canadian Steel Grades
• Canadian Steel Properties
• CSA S16 Code Overview
• CSA S16 Beam Design
• Beam Deflection Calculator
• Beam Capacity Calculator
• Canadian CSA Bolt Capacity
• All Canadian References

Frequently Asked Questions

How does CSA S16 address floor vibration? CSA S16:19 Commentary K provides guidance on floor vibration from human activity. The commentary adopts the AISC Design Guide 11 resonant response method, where the peak acceleration a_p/g from walking is compared to an occupancy-dependent tolerance limit a_o/g. The key parameters are the natural frequency f_n (calculated from the midspan deflection under sustained load), the effective modal weight W, and the damping ratio beta. Commentary K is an informative annex, not a mandatory code clause — but NBCC 2020 Sentence 4.1.5.3(2) requires that floors limit vibration to acceptable levels, making the commentary the de facto standard for Canadian vibration assessment.

What vibration criteria apply to steel-framed floors in Canada? For walking excitation, the peak acceleration tolerance limits are: 0.5% g for offices, 0.2% g for residential, 1.5% g for shopping malls and retail, and 5.0% g for footbridges per CSA S16 Commentary K. These match AISC Design Guide 11 values. The natural frequency of the floor system should ideally be above 4 Hz for offices and above 6 Hz for residential to avoid resonance with walking excitation harmonics (walking pace is 1.6-2.2 Hz, and the first four harmonics reach 8 Hz). The minimum recommended frequency by Commentary K is 4 Hz for open-plan offices and 6 Hz for residential occupancies.

What is the difference between CSA S16 Commentary K and AISC Design Guide 11? They are fundamentally the same assessment method. Both use the resonant response equation a_p/g = P_0 x exp(-0.35 x f_n) / (beta x W) with the same walking force P_0 = 0.29 kN (65 lb). The key differences are: (a) Commentary K presents the method in metric units (kN, mm, MPa) while DG11 uses imperial (lb, in); (b) Commentary K recommends beta = 0.03 for typical Canadian office fit-out while DG11 uses 0.025 — reflecting Canadian practice of more extensive ceiling and mechanical systems; and (c) Commentary K does not include the additional walking force models (rhythmic activity, sensitive equipment) covered in DG11 Chapter 5. For most Canadian office buildings, either standard produces comparable results.

What is the effective panel width and why does it matter for vibration? The effective panel width B determines how much of the floor slab participates in the vibration mode with the beam. Per Commentary K and DG11, B = C x (D_s / D_j)^0.25 x L, where C = 1.0 (simple span), D_s = slab flexural stiffness per unit width, D_j = beam flexural stiffness per unit width, and L = beam span. For typical composite floors, B ranges from 0.4 x L to 0.8 x L. A wider effective panel increases the effective weight W, which directly reduces a_p/g. Beams spaced closer together produce a wider effective panel (more slab mass participates), which is why reducing beam spacing is an effective vibration mitigation strategy.

How should sustained load be determined for frequency calculation? This is the most common source of error in vibration assessment. The sustained load used to compute f_n must represent the mass actually present during normal use, not the full code-specified live load. Commentary K recommends using the slab self-weight, full superimposed dead load, and 10-25% of the design live load (representing the furniture and occupants typically present). Using the full code live load (e.g., 4.8 kPa for office) artificially increases the mass, lowers computed f_n, and produces an unconservative (lower) a_p/g estimate — potentially leading to a floor that passes the check but feels unacceptable to occupants. For a typical office floor, the sustained load is approximately 3.0-3.5 kPa compared to a total specified load of 5.5-6.0 kPa.

Educational reference only. Verify all values against the current edition of CSA S16:19 Commentary K and NBCC 2020 Part 4. This information does not constitute professional engineering advice. Always consult a qualified structural engineer.