Revisiting Earthquake Lessons – Wall Anchorage to Flexible Diaphragms
Thursday, August 30, 2018
This is one in a series of short articles revisiting lessons learned in past earthquakes. This series is being presented to share past lessons with newer members of our profession who were not there to experience them first hand. Contact Dave McCormick or Kelly Cobeen with your comments and suggestions.
Article By: John Lawson, SE
Structure Type: Concrete and Masonry Walls Anchored to Flexible Diaphragms
Earthquakes: 1971 San Fernando (6.7 Mw), 1984 Morgan Hill (6.3 Mw), 1989 Loma Prieta (6.9 Mw), 1994 Northridge (6.7 Mw)
Lessons: In the 1971 San Fernando earthquake, large numbers of concrete tilt-up warehouse buildings had partial roof collapses as perimeter walls separated from the roof. The culprit was a design flaw that relied upon wood ledgers in cross-grain tension/bending for anchorage of the wall to the roof diaphragm. In the 1984 Morgan Hill and 1989 Loma Prieta earthquakes, instrumented tilt-up buildings indicated that past code-level anchorage forces were too low; and in the 1994 Northridge earthquake it was observed that anchorage systems had insufficient ductility and overstrength to resist the high force demands.
The 1971 San Fernando Earthquake
Prior to the 1973 Uniform Building Code (UBC), it was standard practice by design engineers to use plywood edge nailing to the wood ledgers bolted on the inside face of walls to act as the defacto anchorage of walls to the diaphragm for out-of-plane wall forces (Figure 1). This indirect tie arrangement relied upon the wood ledger in cross-grain bending and the plywood panel in tension near its edge, both very weak material properties of wood. Wall anchorage design forces prior to the 1973 UBC were specified as 0.2Wp under allowable stress design, where Wp is the tributary wall weight being anchored. In the 1971 San Fernando earthquake, the wall anchorage in many tilt-up concrete and masonry buildings performed poorly. The common detailing practices of using an indirect wall tie led to wood ledgers failing in cross-grain bending and to plywood edge nailing tearing through panel edges from the wall anchorage tension loads. Partial roof collapses and wall collapses were common in the areas of strong ground motion (Frazier et al., 1971; NOAA, 1973; SEAONC, 2001).
Following the San Fernando Earthquake. Subsequently, the Structural Engineers Association of California recommended (SEAOC, 1973) and the 1973 UBC and 1976 UBC adopted detailing provisions intended to prevent such failures. Beginning with the 1973 UBC, the requirement for a positive and direct wall anchorage was introduced, and reliance on cross-grain bending in wood was expressly prohibited in the wall anchorage system. Furthermore, in order to transfer the heavy perimeter walls’ seismic anchorage forces effectively into the main roof diaphragm, the concept of continuous ties or crossties was explicitly required to collect the wall load and distribute the anchorage force uniformly across the diaphragm depth for further distribution to shear walls. Because of the complexity and cost of providing the necessary repetitive crosstie connections across the popular panelized wood roof structure, the concept of the subdiaphragm was introduced into the 1976 UBC as a design approach for transferring forces from the individual wall ties to the continuous crossties. Subdiaphragms are smaller portions of the main diaphragm, located adjacent to walls, and span between the continuous crossties. The 1976 UBC also increased the wall anchorage force for buildings in areas of high seismicity. Based on observations from the 1971 San Fernando earthquake, the 1976 UBC added seismic Zone 4 into the design provisions to address buildings near active fault systems, and the wall anchorage design force was increased 50% from 0.2Wp to 0.3Wp for seismic Zone 4 through the inclusion of a 1.5 default site factor, under allowable stress design procedures.
The 1984 Morgan Hill and 1989 Loma Prieta Earthquakes
In these two Northern California earthquakes, similar wall-to-roof anchorage failures were observed (EERI, 1985; Phipps and Jirsa, 1990); however, more significant was the strong motion data recorded in several undamaged instrumented tilt-up concrete buildings with flexible wood roof diaphragms. The California Strong Motion Instrumented Program (CSMIP) collected data on tilt-up buildings with flexible diaphragms in Saratoga, California (1-story gymnasium), Hollister, California (1-story warehouse), and Milpitas, California (2-story industrial).
Following the Morgan Hill and Loma Prieta Earthquakes. Research by Celebi et al. (1989) and Bouwkamp et al. (1994) determined that the roof diaphragm accelerations were amplified three to four times that measured at the ground level, suggesting that wall anchorage forces were likely underestimated by the existing UBC provisions for a design level earthquake. This created the basis for a an update in the 1991 UBC increasing wall tie forces in the center half of the diaphragm span by 50% from 0.3Wp to 0.45Wp for seismic Zone 4 (Sheedy and Sheedy, 1992; EERI, 1996) under allowable stress design procedures.
The 1994 Northridge Earthquake
This earthquake was the first test of modern, post-1976 UBC provisions for wall anchorage to flexible wood roof diaphragms under very strong shaking. Hundreds of buildings were severely damaged due to inadequate wall anchorage often resulting in partial roof collapses as seen in Figure 2 (COLA/SEAOSC, 1994; SSC, 1995; EERI, 1996). This region of the Los Angeles area contained a significant inventory of pre-1973 UBC tilt-up and masonry industrial buildings, and their poor performance was not a surprise. However, the amount of damage in buildings designed to more modern codes was unexpected. Most notably, steel anchorage straps fractured through net section in tension causing a sudden loss of anchorage strength.
Following the Northridge Earthquake
Wall anchorage damage to newer buildings was later attributed in large part to inadequate steel strap connection ductility and overstrength to accommodate the very large roof accelerations that occurred (Harris et al., 1998). With prior research indicating that roof top accelerations may be three to four times the ground acceleration, and the Northridge earthquake damage indicating insufficient overstrength or ductility was provided in past connection practices, code writers of the 1997 UBC decided that instead of relying upon connection ductility it was more appropriate to elevate the wall anchorage design forces up to expected levels. As a result, wall anchorage forces at the roof were increased to 0.80Wp using strength-level design provisions. In addition, connection material-specific load factors (1.4 for steel, 0.85 for wood, 1.0 concrete/masonry) were specified to obtain more uniform demand-to-capacity ratios within the anchorage connection considering material overstrengths (SEAOC, 1999).
Under the 2018 International Building Code (IBC), the wall anchorage provisions are found in ASCE 7-16 and have changed little since the 1997 UBC, except the material-specific load factor for wood was set to 1.0 for simplicity. These provisions have yet to be tested in the field by a strong earthquake. Figure 3 illustrates the trial-and-error nature of using lessons learned from earthquakes to influence changes to the building code’s wall anchorage provisions. Recent numerical modeling research indicates that current design force levels are generally appropriate (Lawson et al., 2018).
Bouwkamp J., Hamburger, R., and Gillengerten, J., 1994. Degradation of Plywood Roof Diaphragms under Multiple Earthquake Loading, CSMIP/94-02, California Department of Conservation, Division of Mines and Geology, Sacramento, CA
Celebi, H., Bongiovanni, G., Safak, E., and Brady, A.G., 1989. “Seismic Response of Large-Span Roof Diaphragms,” Earthquake Spectra, 5(2), Earthquake Engineering Research Institute, Emeryville, CA.
City of Los Angeles/SEAOSC (COLA/SEAOSC), 1994. Findings & Recommendations of the City of Los Angeles/SEAOSC Task Force on the Northridge Earthquake, Structural Engineers Association of Southern California, Whittier, CA.
Earthquake Engineering Research Institute (EERI), 1985. “The Morgan Hill earthquake of April 24, 1984,” Earthquake Spectra, 1(3), Emeryville, CA.
Earthquake Engineering Research Institute (EERI), 1996. “Northridge Earthquake of January 17, 1994 Reconnaissance Report,” Earthquake Spectra, 11 (Supplement C), Oakland, CA.
Frazier, G. A.; Wood, J. H.; and Housner, G. W., 1971. “Earthquake Damage to Buildings.” Engineering Features of the San Fernando Earthquake February 9, 1971, EERL 71-02; Earthquake Engineering Research Laboratory, California Institute of Technology, Pasadena, CA.
Harris, S.K.; Hamburger, R.O; Martin, S.C.; McCormick, D.L. and Somerville, P.G., 1998. “Response of Tilt-Up Buildings to Seismic Demands: Case Studies from the 1994 Northridge Earthquake,” Proceedings of the NEHRP Conference and Workshop on Research on the Northridge, California Earthquake of January 17, 1994. California Universities for Research in Earthquake Engineering (CUREE), Richmond, CA.
Lawson, J., Koliou, M., Filiatrault, A., and Kelly, D. J., 2018. “The Evaluation of Current Wall-to-Roof Anchorage Force Provisions for Single-Story Concrete and Masonry Buildings with Lightweight Flexible Diaphragms,” SEAOC 2018 Convention Proceedings (accepted manuscript), Structural Engineering Association of California, Palm Desert, CA.
National Oceanic and Atmospheric Administration (NOAA), 1973. San Fernando, California, Earthquake of February 9, 1971. Volume 1, United States Department of Commerce, Washington, DC.
Phipps, M. and Jirsa, J., 1990. “Performance of Industrial Facilities and Implications for Design of New Structures and Upgrading of Existing Hazardous Buildings,” Proceedings: Putting the Pieces Together – The Loma Prieta Earthquake One Year Later, Bay Area Regional Earthquake Preparedness Project (BAREPP), San Francisco, CA.
SEAOC Seismology Committee, 1973. Recommended Lateral Force Requirements and Commentary, Seismology Committee, Structural Engineers Association of California (SEAOC), Sacramento, CA.
SEAOC Seismology Committee, 1999. Recommended Lateral Force Requirements and Commentary, Seismology Committee, Structural Engineers Association of California (SEAOC), Sacramento, CA.
Seismic Safety Commission (SSC), 1995. Northridge Earthquake, Turning Loss to Gain, SSC Report No. 95-01, Sacramento, CA.
Sheedy P. and Sheedy C., 1992. “Concrete and Masonry Wall Anchorage,” Building Standards, July-April 1992, International Conference of Building Officials, Whittier, CA.
Structural Engineers Association of Northern California (SEAONC), 2001. Guidelines for Seismic Evaluation and Rehabilitation of Tilt-Up Buildings and Other Rigid Wall/Flexible Diaphragm Structures, San Francisco, CA.
Figure 1. Typical Pre-1973 UBC Wall-to-Roof Anchorage Detail
Figure 2. Damage to tilt-up concrete building due to loss of wall anchorage during the 1994 Northridge earthquake. Photo credit: EERI
Figure 3. Evolution of UBC/IBC provisions for wall anchorage to flexible diaphragms (Lawson et al. 2018).