Seismic Anchorage Failure: Insufficient Anchor Capacity

Case classification

Illustrative case study. This scenario is constructed from common patterns observed in post-earthquake damage assessments and anchorage failure investigations. It represents a typical failure mode rather than a specific documented event.

Scenario description

An electrical cabinet weighing 800 kg was mounted on a reinforced concrete floor using four post-installed expansion anchors. The cabinet had a fundamental natural frequency of 12 Hz, placing it in the amplified region of the floor response spectrum. During a seismic event with a peak ground acceleration of 0.25g, two of the four anchors failed in tension, causing the cabinet to tilt and separate from the floor on one side. The cabinet remained attached by the remaining two anchors but was non-functional due to internal component damage from the tilt.

Analysis of the failure

The seismic analysis of the cabinet had been performed using a static equivalent force method with a seismic coefficient of 0.25 (corresponding to the ZPA). However, the cabinet natural frequency of 12 Hz was in the amplified region of the floor response spectrum, where the spectral acceleration was 0.45g — nearly twice the ZPA used in the static analysis. The anchor tension from the overturning moment at 0.45g exceeded the tension capacity of the expansion anchors. Additionally, the anchors were installed at an edge distance that reduced their concrete breakout capacity below the catalogue value used in design.

Root causes

• The cabinet was analysed as rigid (using the ZPA) when it was flexible (natural frequency in the amplified region)
• The spectral acceleration at the cabinet frequency was 80% higher than the ZPA used in the static analysis
• The anchor capacity was taken from the catalogue value without reduction for edge distance
• The anchor edge distance was below the minimum specified by the anchor manufacturer
• No consideration was given to the interaction between tension and shear on the anchors

Lessons learned

• Equipment with a natural frequency below the rigid boundary must be analysed using the response spectrum at its frequency, not the ZPA
• Anchor capacity must be reduced for edge distance, spacing, and concrete strength per the anchor standard
• The interaction between tension and shear must be checked — anchors that are adequate in tension alone or shear alone may fail under combined loading
• Post-installed expansion anchors are sensitive to installation quality; the installation should be inspected and tested
• The rigid classification threshold should be confirmed against the actual floor response spectrum, not assumed

Remediation

The failed anchors were replaced with larger-diameter adhesive anchors with greater embedment and adequate edge distance. The cabinet was re-analysed using the floor response spectrum at its natural frequency, confirming that the new anchors had adequate capacity with margin. A sensitivity study on the cabinet frequency was performed to confirm that small changes in frequency (from mass changes) would not move the cabinet into a more severe part of the spectrum.