20100303

UBC or IBC? Revisiting the justifications for moving from UBC to IBC as basis for new NSCP earthquake loading provisions

UPDATE: Please refer also to my newer article entitled, "Designing for near-field earthquakes."


1. Introduction

The National Structural Code of the Philippines (NSCP) has been the primary code that provides guidance to civil and structural engineers on the design and evaluation of buildings, towers, and other vertical structures around the Philippines since its first edition was published in 1972 (then as National Structural Code for Buildings or NSCB). It includes provisions for steel, concrete, timber, and masonry design as well as for estimating combined effects of dead, live, earthquake, wind, and other loads.

The NSCP has been historically based on the Uniform Building Code (UBC) from the USA, including its earthquake loading provisions (ELPs). The following table illustrates the historical code basis of the NSCP (after Abinales et al, 2009):

Table 1. Historical code basis of the NSCP


Code (Edition)Basis for general and earthquake loading provisions
NSCB 1972 (1st edition; 2nd printing in 1977)UBC 1970
NSCB 1982 (2nd edition)UBC 1978
NSCP 1987 (3rd edition)UBC 1985
NSCP 1992 (4th edition, Volume 1 – Buildings, Towers, and Other Vertical Structures; Volume 2 for Bridges published in 1997)UBC 1988
NSCP 2001 (5th edition, Volume 1 – Buildings, Towers, and Other Vertical Structures)UBC 1997
NSCP 2010 (6th edition, Volume 1 – Buildings, Towers, and Other Vertical Structures)?

The last row in the table above has been added to call attention to the forthcoming update of the NSCP scheduled this year, the work on which has been started as early as 2007 or 2008. There was early discussion on adopting the new International Building Code (IBC), which has since “replaced” the UBC in the USA. But toward the end of 2009 because of requirements that could not yet be met to be able to fully adopt the IBC, the code update discussions were starting to lean towards re-adopting the UBC 1997 particularly for the ELPs.

The UBC 1997 is in no way an inferior code basis particularly for ELPs. It was in fact a major update to earlier UBC ELPs. Among the more important features of the UBC 1997 ELPs, is the incorporation of near-source factors to account for near-field effects, conversion of earthquake loads for use in ultimate limit states, and provision for more soil types which involved classification according to other metrics such as shear wave velocities as well as flagging of liquefaction-prone sites. The UBC 1997 is actually meant only as a transition code from the old UBC ELPs (i.e. in UBC 1994) to the new IBC ELPs which are based on the 1997 NEHRP Recommendations.

However, the initial argument that became the reason why the IBC was considered during the early code update discussions is that the IBC represents the state-of-the-art in earthquake engineering. Meanwhile, the IBC which was initially published in 2000 has had 3 updates excluding addenda or corrections. The latest version is the IBC 2009.

This paper aims to further establish why the IBC should be used in lieu of the UBC for ELPs particularly in new updates to the NSCP by revisiting the justifications used in the USA for moving from the UBC to the IBC ELPs. The paper also discusses how relevant these justifications are in the Philippine setting.

2. Old or New?

To put it simply, the UBC is now an old code that has now been superseded by the newer IBC. The authors believe that updating the code means improving earthquake-resistant designs, particularly of newer structures. If the same old provisions are retained, the earthquake-resistant design of newer buildings are not improved or updated, and only the “code” is “updated.”

Furthermore, it is not a surprise that the IBC reflects the latest data and most current standards and technology, some of which are discussed in this paper. For example, the latest steel, concrete, timber, and masonry design standards in the US which are bases of similar standards here in the Philippines are now all made compatible with the IBC. The California Building Code which governs the structural designs in the State of California in the USA has been updated, with the IBC now as basis whereas previously it was the UBC. All this indicates that most engineers and engineering institutions, at least in the USA, are all now leaning towards using IBC and away from UBC.

The implication here in the Philippines is that if we would like to “update” our steel, concrete, timber, and masonry design codes (e.g. Chapters 4 to 7 of the NSCP), which presumably we all desire for reasons both of safety and economy, we would then need to update our NSCP ELPs’ code basis from the UBC to the IBC.

Lastly on the same matter, note that the UBC has less stringent quality assurance requirements as stated by Ghosh (2001) compared to the more up-to-date IBC.

Dowty and Ghosh (2002) themselves call the UBC ELPs as “outdated.” They question the use of old codes when new, updated ones are already available (and in this case, accepted already by the general public). Pong et al (2006) state that it is “imperative that the structural engineer(ing) professional” “…keep up to date with the codes.”

Of course as stated also by Pong et al (2006), it is likewise most important that engineers have “good working knowledge and understanding of the fundamentals of seismic design principles.” The authors believe that engineers would choose to use the IBC over the UBC when armed with such “good working knowledge and understanding of the fundamentals of seismic design principles.”

3. More Arbitrary or More Accuracy?

3.1. 3 or 6 parameters?

The IBC effectively uses 6 parameters to define site seismicity and required level of seismic detailing, compared to 3 in the UBC. In terms of seismic parameters alone, the IBC uses 2 spectral parameters (Ss and S1) while the UBC only one (zone factor Z). For specific seismic detailing requirements, the IBC effectively uses building use or occupancy, source type, soil type, the two spectral parameters, and distance as parameters (to come up with Seismic Design Categories) whereas the UBC only uses the seismic zone parameter.

Dowty and Ghosh (2002) state that this is “one of the most significant improvements in the 2000 IBC over the 1997 UBC... Soil-modified ground motion is what a structure experiences during an earthquake, so it is only proper that it should be a significant factor in determining relevant restrictions applied to a structure… It is like going from one-size-fits-all to custom-made.” They continue that this is “a major departure from prior practice and is reflective of real conditions.”

This improvement in accuracy in defining the seismicity started from as way back in 1978 in ATC 3 and reflected in subsequent NEHRP recommendations, and now found its way to the IBC. The use of seismic design categories (SDCs) is now also reflected in material-specific design codes such as the ACI 318, among others.

Related to this, we can see how the IBC puts more emphasis on the building use or occupancy than does the UBC. This is discussed further in Section 4.

Also because of this, seismic detailing would be required in more locations in the IBC. Based on case studies by Pong et al (2006), 2 out of 3 Zone 0 or Zone 1 locations, the IBC requires the equivalent of Zone 2 detailing if in Site Class E; whereas the UBC would have not required such Zone 2 detailing because the sites are in Zones 0 or 1 (i.e. no seismic detailing is required). The consequences of this are obvious, and this is further discussed also in Section 4.

3.2. Zone map or contour map?

Dowty and Ghosh (2002) say that “to begin with, the maps in the UBC have always been based on seismic zones. Spectral acceleration maps used together with the IBC are more sophisticated in that they are not zone maps at all, but rather contour maps giving spectral response quantities.”

Dowty and Ghosh (2002) further state that use of spectral acceleration maps are able to remove influences from “political processes,” or in the words of Pong et al (2006), the seismicity at the sites that were “arbitrarily” assigned a higher zone in the UBC 1997 (or say in the NSCP 2001) are now better defined in the IBC.

In the case of the NSCP, we all know that the Philippines has been historically divided into two zones, Zones 2 and 4 since the NSCP 1st edition. Our experts would have justifications for this, but meanwhile the UBC 1997 itself assigns certain locations in the Philippines to Zone 3 or 4. Recently, there are parties that still suggest that some locations in the Philippines should be in Zone 3.

In the Philippines, perhaps of most concern is Metro Manila which is dissected by the Valley Fault System (VFS) and where most economic investments particularly in infrastructure including the most concentration of tall buildings and urban residences are located. The VFS is categorized by the NSCP 2001 as Source Type A, but many are stating that it could be Source Type B if the UBC 1997 itself was the basis.

In any case, when spectral acceleration contour maps are used, such “arbitrary” assignments influenced by “political processes” would be removed. The contour maps present a more “realistic” and more “refined” approach in determining earthquake loads.

Dowty and Ghosh (2002) add that the UBC by nature of its zoning, assumes that all areas within the same zone will have the same peak ground acceleration. This would not be the case and certainly not in the Philippines. Dowty and Ghosh consider this as a fundamental shortcoming of the UBC.

The issue of course is that how and when can we have our very own spectral acceleration map for the whole of the Philippines? Who can perform the study? Who can provide the funding for such study? Some colleagues suggest looking to insurance companies for funding because they themselves would be able to use such a “hazard” map when they perform risk assessments.

4. Over-Design or Cost-Efficiency?

Dowty and Ghosh (2002) state that the UBC provides overly conservative earthquake loads for sites at near-source locations, owing to the UBC method of combining seismic zoning and near-source factors. This means that earthquake loads for buildings at some sites within 15 km from known active faults could be too much.

Practically all of Metro Manila is within 15 km from the VFS. Thankfully, the NSCP 2001 proponents did not include the UBC provision for sites between 2 and 5 km from the nearest seismic source which reduces the impact of such potential over-design in terms of earthquake loads. But there is still some over-design. The following plot of bedrock spectra in Metro Manila would illustrate this “over-design” when using the NSCP 2001 and more particularly so when the near-source factors are used (e.g. in the case of Metro Manila). The figure below suggests that the NSCP 2001 earthquake loads for sites with considerable near-source effects such as Metro Manila, are closer to a 2475-year event than the intended 475-year event. This was also illustrated by a case study by Ghosh and Khuntia (1999).

Figure 1. Comparison of response spectra between probabilistic hazard assessment and NSCP 2001 with near-source factors (after Koo et al, 2009)

Lastly, the UBC uses R-values that are around 7% less than the IBC values which mean that if all other parameters are correct, UBC earthquake loads would be 7% more than IBC values.

5. Cost-Savings or Uniform Safety?

Dowty and Ghosh (2002) state that the IBC provides a uniform margin of safety for all structures in all locations, whether high or low seismic regions, particularly against collapse. Design according to the IBC effectively targets a Collapse Prevention objective at a 2475-year event (2% probability in 50 years used for low seismic regions, or approximately equivalent to the maximum credible earthquake or MCE at high seismic regions), which governs most designs whereas the UBC only designs to meet a Life Safety objective at a 475-year event (10% probability in 50 years, or the design basis earthquake).

Additionally they add that the UBC may not be safe in terms of collapse prevention at the MCE or infrequent but very strong events particularly at lower seismicity regions (e.g. this would include Zone 2 locations in the Philippines), and in terms of designing for less stronger but more frequent earthquakes whether in low or high seismic regions (i.e. all around the Philippines, whether Zone 2 or Zone 4).

Other limitations of the UBC are as follows (after Ghosh and Khuntia, 1999; Pong et al, 2006; and Ghosh, 2001):
  • The UBC puts less importance on building use or occupancy; the IBC uses importance factors as high as 1.5 for essential facilities compared with only around 1.25 in the UBC. This together with the lower R-values in the UBC, the UBC earthquake loads become more than 20% lower than the IBC earthquake loads. This suggests that important buildings might be more at risk when designed to the UBC.
  • At some locations that have been arbitrarily assigned to lower seismic zones, the UBC earthquake loads could be further underestimated compared with those using IBC spectral maps. In the US for example, some Zone 3 locations could have as much as 25% lower seismic loads when the UBC is used than when the IBC is used.
  • The level of seismic detailing, restrictions on building heights, structural system, and structural irregularity, and choice of analysis procedure are all dependent only upon the seismic zone in the UBC. This is dependent on 6 parameters in the IBC. As illustrated earlier, some locations are found to be required by the IBC to have seismic detailing, but by the UBC. Also, some locations are found to be required to have detailed site-specific geotechnical investigation according to the IBC but not to the UBC. Conversely, the UBC does not have the equivalent of siting restrictions for structures assigned to SDC E and F, making UBC less stringent. Also more specifically, the UBC does not have height and structural system limitations below Zone 3 while the IBC has such limitations for buildings under SDC B and C. This suggests that under certain soil conditions, some buildings including “important” ones, although located in Zone 2 might still actually require detailed site-specific geotechnical investigation and/or seismic detailing, or because of height, structural system, or irregularity might not be permitted.
  • Calculated drifts could be “passing” under UBC criteria but not under IBC criteria.
  • For some cases, UBC is less stringent in its treatment of certain irregular structures.
  • UBC ELPs have less stringent requirements with respect to redundancy. This is the case for an example building studied by Pong et al (2006) and located in another site where near-source effects are significant. This may also be attributed to a difference in the vertical distribution of the design base shear, especially in the upper levels. Ghosh (2002) adds that this is also “because: (a) it [the UBC] does not use variable floor areas along the height of the building, (b) it takes r as the highest floor-level r-value over the lower two-thirds of the building height, while NEHRP [the IBC] considers the entire building height, and (c) for special moment frames in SDC E or F, r is restricted to 1.1 by NEHRP [IBC], but only to 1.25 by the UBC.”
  • As mentioned earlier, the UBC is less stringent in terms of Quality Assurance requirements.
  • The UBC provisions on additions and alterations to existing buildings “are somewhat less restrictive.” (Ghosh, 2001)
  • According to Ghosh (2001), the “UBC is less stringent with respect to the consideration of vertical earthquake ground motion in special seismic load combinations.”
  • According to Ghosh (2001), the UBC is somewhat less stringent in high seismic areas because the IBC “allows T (rationally computed elastic fundamental period) to be a larger multiple of Ta in lower seismic areas, but restricts it to a lower multiple of Ta (approximate elastic fundamental period) in the highest seismic areas.”
  • According to Ghosh (2001), the “UBC is less stringent with respect to the dynamic amplification of torsion in structures of SDC C, D, E, and F, where Type 1 torsional irregularity exists.”
  • According to Ghosh (2001), the UBC is less stringent with respect to drift control for tall buildings because of the “disregard of minimum base shear given by Eq. (30-6) in drift computations.”
  • According to Ghosh (2001), the UBC is less stringent with respect to geotechnical investigation in Seismic Zone 2, in terms of pile to pile cap connections in buildings assigned to SDC C and higher, and with respect to ties between spread footings founded on Site Class E and F soils.
  • According to Ghosh (2001), the UBC is less stringent for a number of material-specific (i.e. steel, concrete, etc.) design requirements.
  • According to Ghosh (2001), it was a policy in the development of the IBC that “each basic seismic-force-resisting system …would have its own unique set of detailing requirements and commensurate R and Cd - values assigned to it. This eliminated having the same name, R, and Cd values assigned to two completely different seismic-force-resisting systems, as is the case with the 1997 UBC… R- and Wo- values for the same structural system are often different between” the IBC and the UBC.

6. Conclusion

By adopting the IBC, the new NSCP ELPs would:
  • reflect the most current data, standards, and technology,
  • remove “arbitrary” assignments because of any “political” influences,
  • provide a more accurate definition of earthquake loads,
  • removes all conservatism associated with near-source factors, and
  • provides a better and more uniform margin of safety for more buildings in more locations.

The authors hope that a case has been made for using the IBC rather than the UBC. It is not discounted that the NSCP can still choose to adopt the UBC but allow structural designers to use the IBC where necessary economically or technically, and where possible (e.g. when spectral acceleration data are available). This is of course a feature of the UBC itself. Lastly, it is re-emphasized that the UBC 1997 particularly was meant as a transition code from the old UBC ELPs to the new IBC ELPs. The new IBC ELPs are now out and thus it is now time to let go of the UBC 1997 ELPs; their purpose has been served.

7. References

A. Abinales, E. Morales, and R. Aquino, “ASEP code development program 2007-2010: update of the NSCP,” Proc. 14th ASEP International Convention, Quezon City, Philippines, 2009.

Association of Structural Engineers of the Philippines, National Structural Code of the Philippines, 5th edition, NSCP-2001.

International Conference of Building Officials, Uniform Building Code, UBC 1997.

International Code Council, International Building Code, IBC 2009.

R Koo, T Mote, R Manlapig, and C Zamora, “Probabilistic Seismic Hazard Assessment for Central Manila in Philippines,” Proc. 2009 Australasian Earthquake Engineering Society Conference.

S Dowty and SK Ghosh, “IBC Structural Provisions: A Better Alternative,” Building Standards, May-June 2002, pp. 12-16.

SK Ghosh, “Comparison of the Seismic Provisions of the 1997 Uniform Building Code to the 1997 NEHRP Recommended Provisions,” A report to: Building and Fire Research Laboratory, National Institute of Standards and Technology, Gaithersburg, MD, February 2001.

SK Ghosh and M Khuntia, “Impact of Seismic Design Provisions of 2000 IBC: Comparison with UBC,” Proc. SEAOC 1999 Convention, pp. 229-254.

WS Pong, A Lee, and ZH Lee, “The International Building Code and Its Implications On Seismic Design,” Proc. 4th International Conference on Earthquake Engineering, Taipei, Taiwan, October 12-13, 2006.
 
8. Further Reading

B Nagel and J Russell, “2001 California Building Code (1997 UBC) -vs- 2007 California Building Code (2006 IBC): Conventional Framing Provisions Comparison,” September 7, 2007, reviewed by the conventional framing subcommittee of the CALBO (California Building Officials) Seismic Safety Committee.

International Code Council, UBC-IBC Structural (1997-2000): Comparison and Cross Reference, 2000.

International Code Council, 1997 UBC/2006 IBC Structural Comparison and Cross Reference, 2006.

JR Harris and RE Bachman, “Seismic Design Provisions of ASCE/SEI 7-2005,” Proc. 2007 ASCE/SEI Structures Congress.

3 comments:

jojitsan said...

a querry, not for IBC seicmic but for over-over-sized axial concrete, ACI says p min of 0.01(Ag/2), IBC says 0.0025, is it ok then to use 0.0025Ag min ?

ronjiedotcom said...

@jojitsan: can you be more specific where IBC specifies such? And which IBC specifically. The IBC of today (2006 or 2009) that I know refers to ACI (318-05 or 318-08), and therefore there should be no conflict. In some older codes, the "General Requirements" or earlier sections of the code could have some conflicting information with later sections (such as the Concrete Design section). The more onerous provision should govern. Anyway in this case you mentioned, I remember from 8 years ago when I took up advanced concrete design in grad school that there is a good, rational basis for the 1% minimum set by ACI. I would think the 0.0025 you are talking about could be for something else. That minimum value appears very close to that use for slabs - which are over-over-sized for axial loads (because flexure governs). Please read the codes carefully and look at the bigger context of such provisions. But as mentioned, I am not 100% familiar with all "new" provisions in the IBC.

ronjiedotcom said...

@jojitsan: I posted your question and my reply as a new article. Please visit http://engg.ronjie.com/2010/07/reply-to-reader-question-on-ibc-vs-aci.html.