Reassessment of Earthquake Design Philosophy in Australia after the Christchurch Earthquake

Transcription

1 Reassessment of Earthquake Design Philosophy in Australia after the Christchurch Earthquake by Helen Goldsworthy, Department of Infrastructure Engineering, University of Melbourne, and Paul Somerville, Risk Frontiers, Macquarie University Introduction The ongoing Canterbury earthquake sequence has caused the largest natural hazard disaster in New Zealand's written history. Impacts are estimated at 8% of national GDP, much more than the relative impact of Hurricane Katrina on the US economy or the recent Tohoku earthquake and tsunami on the Japanese economy. The severe consequences of the Canterbury sequence have led New Zealanders to re-evaluate the performance of both communities and the built environment to find a balance between protection and costs that the society will tolerate. Although Australia is located in a region of low to moderate seismicity, it is one of the most active intraplate regions in the world due to strains created by the Indo-Australian plate colliding with the Eurasian and Pacific plates. Most Australian capital cities have known faults in their vicinity that are capable of generating damaging shallow earthquakes, and historical earthquakes of magnitude 6 or higher in Australia have been caused by ruptures on shallow reverse faults that are similar to the one experienced in Christchurch on February 22 nd, For example, the M w Newcastle earthquake occurred on a thrust fault and would cause losses of $3.2 billion if it were to recur today (Crompton, 2011). Australian earthquakes have sometimes occurred in clusters (three M w 6.25 to 6.5 earthquakes occurred in one day in the 1988 Tennant Creek sequence), and have been followed by aftershock sequences like that of the Canterbury sequence (such as the sequence of events that occurred off the east coast of Tasmania near Flinders Island from 1884 to 1886 with magnitudes as large as M w 6.9). Due to the vulnerable nature of the building stock, a M w 6.0 or greater earthquake occurring within an Australian capital city would probably cause extensive damage and a large number of fatalities. This is recognised by the insurance industry, which has determined that Sydney presents one of the highest insured earthquake risks in the world. This briefing proposes changes in building codes in Australia that, over time, would reduce the potential for catastrophic losses from earthquakes in Australia. The proposed changes would use the framework of performance-based design based on quantitative estimation of the capacity of buildings to withstand strong ground motion. Performance Based Seismic Design Performance-based building design philosophies have been adopted in regions of moderate to high seismicity throughout the developed world, including in Christchurch, over the last fifteen to twenty years. This practice was initiated partly by the severe economic disruption caused by moderate-level earthquakes in Los Angeles, California (1994 M w 6.7 Northridge earthquake) and in Kobe, Japan (1995 M w 6.9 Kobe earthquake). A performance objective matrix is a chart showing the performance level that should be achieved for several different earthquake design levels. There are objectives for structures of ordinary importance (the basic objectives) and more demanding ones for essential or safety critical structures. The performance matrix most commonly adopted is shown in Figure 1. The earthquake design levels are expressed by the return period of the event. In New Zealand in recent times the design level earthquake used in the loading codes would, for most ordinary buildings, be the Briefing Note No. 232 Page 1

2 500 year return period ("rare") event, and the performance level expected for this level of event would be "Life safe" or better. Figure 1. The current Performance Objective Matrix. Source: Buchanan et al, Building Performance in the Christchurch Earthquake The ground motions recorded in the Christchurch earthquake were locally very high in the Christchurch CBD, with shaking levels in excess of those codified for a 2500 year return period earthquake. Despite this, many of the modern buildings did not collapse. This is primarily because of detailing provisions in the various material codes that enabled the performance objective of "Near Collapse" or better under a 2500 year return period ("very rare") event to be achieved. These detailing requirements are based on capacity design principles and were largely developed at the University of Canterbury in the 1970s and 1980s. It is mostly because of those developments and the implementation of them by practicing structural engineers in New Zealand that the number of fatalities in the Christchurch earthquake was not greater. Building Codes in Australia In Australia, the current return period of the design level earthquake for ordinary buildings in our capital cities is lower than in regions with similar levels of seismicity in the U.S. and Canada. The Building Code of Australia stipulates the return period for design level earthquakes to be 500 years for ordinary buildings (importance level 2), with a higher return period for "more important" buildings, e.g years for buildings of importance 4. This differs from current U.S. practice, which attempts to create a uniform margin against collapse due to the design ground motion level for all regions across the U.S. It has been recognised for some time that there is a larger ratio between the level of ground motions experienced in a 2500 year return period event to those in a 500 year return period event in regions of low to moderate seismicity as compared with regions of high seismicity. To account for this, the ground motion hazards in the U.S. are defined in terms of maximum considered (MCE) ground motions. A lower bound estimate of the margin against collapse in structures designed to the seismic provisions in the U.S. standard has been taken as 1.5. Hence, the design earthquake ground motion was selected at a ground shaking level that is 1/1.5 (or 2/3) of the MCE ground motion, which in most regions in the U.S. has a uniform probability of exceedance of 2 per cent in 50 years (return period of about 2500 years). While it is acknowledged that stronger shaking could occur, it was judged economically impractical to design for such ground motions. Briefing Note No. 232 Page 2

3 Implications of the Christchurch Earthquake for Building Codes The observed damage to buildings in Christchurch was sometimes associated with known inappropriate design and construction practices. Unreinforced masonry buildings, including many of considerable heritage value, performed poorly unless they had been comprehensively strengthened, as described in Briefing 222. Older reinforced concrete buildings (pre-1980s design) were known to be vulnerable even though they had performed reasonably well in the September 2010 Darfield earthquake. These buildings exhibited reasonable behaviour until their ductility was exceeded and then progressed to dramatic and sometimes catastrophic failure. More recent (from the 1980s onwards) medium-rise and tall buildings benefited from the use of capacity design principles in their design. However, consistent with the extreme nature of the ground motions, many of these buildings suffered considerable damage, usually in the regions of the building that had deliberately been designed to be weak and to dissipate energy in a controlled cyclic manner. Proposed Changes in NZ Building Code Philosophy The design process in New Zealand is currently under close scrutiny. Many people have been disappointed by the end results of the Canterbury earthquake sequence. The extent of the damage, the cordoning off of the CBD, the societal and economic effects of the disruption to services and business activities, and most importantly, and the deaths of 181 people were worse outcomes than the general populace had anticipated. The poor performance of some of the older buildings was expected and recommendations have been made to increase the required strengthening from 33% to 67% of the code requirement for new structures. Some engineers are even dissatisfied with the philosophy of capacity design and the extent of the damage suffered by some modern buildings in the Christchurch earthquake (Buchanan et al., 2011). They would like to implement new design technologies to improve the seismic performance of buildings even under extreme events. Their proposed Performance Objective Matrix is given in Figure 2 below, and is intended to replace the current performance matrix given in Figure 1. They propose a combination of two different tactics: that of increasing the level of seismic design loading, and that of switching to higher performance building technologies such as base isolation and damage-resistant technologies. These damage resistant design technologies are on the cusp of earthquake resistant design inventiveness and include rocking walls and frames, with and without post-tensioning, and a variety of different energy dissipating devices attached to the building in different ways. Figure 2. The proposed Performance Objective Matrix. Source: Buchanan et al., Briefing Note No. 232 Page 3

4 Implications for Building Code Changes in Australia The need to reassess the level of seismic hazard used to determine the design level earthquake in Australia has already been discussed earlier in the section on "Building Codes in Australia." If based on a return period of 500 years, the spectral acceleration due to the design level earthquake in Australia will be lower than that in Christchurch (about one third in accordance with current codes). The current performance objective in Australia is to achieve life safe performance or better in a rare event (currently defined as a 500 year return period event). However, for most buildings, there is no provision made for a higher level event. As happened in Christchurch, it is the very rare event that could cause major damage, potentially rendering the CBD unusable for a long period. This is exacerbated in Australia by the fact that material codes such as the Steel Structures code (Standards Australia, 1998) and the Concrete Structures code (Standards Australia, 2009) do not require designers to use capacity design principles in their design. The implementation of these design principles in New Zealand (since the 1980s), in line with the performance requirement for "near collapse" or better under a 2500 year return period event, is what is thought to have saved many lives in the Christchurch earthquake. In Australia, due to the lack of attention given to seismic design, the performance of some buildings is likely to be poor in a very rare earthquake event. The Christchurch experience indicates the need to consider additional aspects of the risks associated with earthquakes in Australia. For example, the current philosophy looks at single buildings without regard to whether they are in a built-up area or in a remote community. In the Christchurch earthquake, a building in the CBD that suffered no or only minor damage would have been out of operation for many months after the earthquake. This was because of the extreme damage suffered by many neighbouring buildings and doubts about their safety, which has led to the CBD being cordoned off. In New Zealand, the move to a different performance matrix such as the one in Figure 2 and the associated move to develop damage-resistant design technologies that can be widely employed, would lead to expectations of quite a different post-earthquake scenario. In particular, the minimum performance objectives for an ordinary building at importance level 2 would be very similar to those of the importance level 3 and 4 buildings, even for very rare earthquakes. Hence, in all cases the buildings would be expected to be repairable, and widespread closures of entire city areas rendered unnecessary. Structural engineers in New Zealand are currently trained to incorporate reliable ductility in buildings so as to enhance the building's ability to withstand a very rare earthquake event without collapse. Australian engineers need to be similarly trained (Goldsworthy, 2011). Nordensen and Bell (2000) have estimated that good seismic resistance for new construction in regions of low to moderate seismicity in the U.S. could be obtained for small incremental costs in construction. Since the wind loading requirements for many structures in these regions frequently exceed the seismic base shear, there would simply be a requirement for the addition of details that provide ductility and toughness. References Buchanan, A.H., D. Bull, R. Dhakal, G. MacRae, A.Palermo, and S. Pampanin (2011). Base Isolation and Damage-Resistant Technologies for Improved Seismic Performance of Buildings. Report to the Royal Commission of Inquiry into Building Failure Caused by the Canterbury Earthquakes. Crompton, R. (2011). Normalising the Insurance Council of Australia Natural Disaster Event List: council%20of%20australia%20natural%20disaster%20event%20list.pdf Briefing Note No. 232 Page 4

5 Goldsworthy, H. (2011), "Lessons from the February 22nd Christchurch Earthquake", Proceedings of the Australian Earthquake Engineering Society Conference, Nov, Novotel Barossa Valley Resort, South Australia, Paper No. 25, 19 pages, Keynote paper. Nordensen, G.J.P. and Bell, G. R., (2000), Seismic Design Requirements for Regions of Moderate Seismicity, Proceedings of the 12th World Conference in Earthquake Engineering, Auckland, New Zealand, Paper No. 825.Paulay, T. (1988) Seismic Design in Reinforced Concrete: The State of the Art in New Zealand, Bulletin of the NZ National Society for Earthquake Engineering, Vol. 21, No. 3 Standards Australia (1998), AS : Steel Structures. Standards Australia (2009), AS : Concrete Structures. Briefing Note No. 232 Page 5