Test case update: Council abandons NZSEE %NBS framework

It has been a while since we have had anything to report on the test case. The reason is that the Council has wanted to do a full engineering assessment of the building and the Court has given them time to do it. We were opposed to the extension of time. The Council has had eight years to produce an assessment, and obviously should have done one in the first place. It could not have been satisfied that the building was earthquake prone on the basis of a shoddy IEP (Initial Evaluation procedure) %NBS estimate, which, on its own admission, was just a ‘coarse screening’ tool and was not to be used by anybody for any other purpose.

Trying to defend the EPB designation decision on the IEP alone will be a difficult task, so it was not a surprise that the Council would want to shore up their case with something apparently more substantive.

What was a surprise was the content of the seismic assessment report when it finally turned up after three and a half months. We were expecting a standard detailed seismic assessment that would demonstrate that building was below 34%NBS, and that the Council was right all along.

However, the Council’s engineering consultants did not use the New Zealand Society of Earthquake Engineering (NZSEE) framework because it does not consider the likelihood of collapse, which is one of the two legal tests for an earthquake prone building.

This is a big win for us on a very obvious point. We have argued this all along and the Council has now appears to have conceded. It is not that the legal test was new to the Council. They had advice from the outset. The following is an excerpt from an email in the Council’ policy file previous update :

“Advice we have been given indicates that for a building to be earthquake prone both subsection a) and b) of section 122(1) must apply and the interpretation of ‘likely to collapse must be read in the context of the definition as a whole. So the part b) is about collapse (even if only partial collapse) in a moderate earthquake (the context). This raises some queries for me including:

Is this different from your understanding? You are aware I do not like the second part of the Act clause. My own view is different from what I know has been given as legal advice. I suspect that only a Court case would establish the “correct” legal interpretation. I believe that if this part of the act is to remain it will be necessary to interpret it in the broadest manner possible, as we have indicated in the guidelines.

So the Council officer did not like the legal opinion, so developed a policy based on the NZSEE guidelines that effectively ignored it.

Three Courts, including the Supreme Court, have found that likely to collapse is one of the EPB tests. But only now, at the 11th hour, has the Council changed its stance.

In place of the discredited NZSEE approach, the consultants used the American Society of Civil Engineers (ASCE) framework, which is set out in their standard document ‘Seismic Evaluation and Retrofit of Existing Buildings (ASCE/SEI 41-13). This framework, it is suggested, allows an assessment of the likelihood of collapse of the building. The building was modelled in a moderate earthquake mostly using the model coefficients specified in ASCE 41-13.

The other point of interest was that the modelling took into account the fact that the building was structurally joined to the next building.  In their IEP the Council missed this obvious point.

And what was the outcome? The building did indeed collapse (or at least a very small part of it did) in a moderate earthquake. This of course does not get the Council off the hook, because it agreed at the hearing that the EPB evaluation was of the building as a whole not of a part, and in any event the law does not provide for parts of buildings to be earthquake prone.

More fundamentally, the modelled outcome simply doesn’t make sense. A useful piece of information from the report is that peak ground acceleration (PGA) close to the site was 0.11 g on two occasions in the 2013 Cook Strait earthquakes. A moderate earthquake is defined as having a PGA of 0.13g. There was absolutely no structural damage to the building in the earthquakes, but if the model is correct then there should have been material damage.

This was not an isolated case.  There were about allegedly 800 earthquake prone buildings in Wellington at the time. None of them incurred structural damage. Even if there was just a two percent chance of an individual building being visibly structurally damaged or collapsing in a single event, the odds that there was not a single such incident in the two earthquakes are 109,200 billion to one. Even if there were a few structural damage cases that were not recorded the odds of such an outcome are still extremely remote.

So what is wrong with the model?

First, ASCE/SEI 41-13 does not model collapse at all and is somewhat similar to the NZSEE new building model that is calibrated to ensure that there is a very low probability of collapse at the ultimate limit state. The consultants then arbitrarily apply their own factor of 2 to adjust to a collapse state. But there is no evidence to support this number, and nothing in ASCE 41-13 that would lead to this conclusion.  On the evidence from the Cook Strait Earthquakes, and the other evidence we produced in our submissions, the calibration is clearly wrong.

It would also be interesting to see how the ASCE based model performs when applied to school buildings. A couple of years ago the Education Ministry tested the actual strength of standard school buildings and found that the NZSEE model was wrong by factor of 5 to 10.

There is a maxim in science: if your model doesn’t explain reality, change your model.

Second, there is a serious question as to whether the ASCE 41-13 model is fit for the purpose for assessing building behavior in moderate earthquakes. A moderate earthquake (as defined in New Zealand regulations) has a return time of about 60-70 years. This corresponds to the seismic hazard level of a 50% chance of a return in 50 years (a return time of 72 years) in table C2-1 on page 32 of the ASCE standard. The standard (table C2-2) does not set performance levels for this level of seismic hazard.

The performance standards (Basic performance objectives for existing buildings – BPOE) in the standard are set to much higher seismic hazard levels.

For the BPOE, this standard sets the Seismic Hazard Levels based on 5% in 50-year and 20% in 50-year probabilities of exceedance.  225 and 975 yr. return times”

The reason that performance objectives are not set for low seismic hazards is possibly that there is no point, as there is little risk of material damage at those hazard levels.

Applying performance standards calibrated to much stronger earthquakes, to moderate earthquake outcomes, does not appear to be appropriate.

In our view the use of ASCE 41-13 is a diversion. It appears to be an attempt to cloak a collapse modelling framework, which at the critical point relies on pure guesswork, with apparent scientific respectability. ASCE 41-13 might be world famous, but the collapse adjustment factor is, apparently, only world famous in the Council’s seismic strengthening unit.

The basic idea that an ultimate limit state calculation can be uprated by a factor to a collapse standard, is, however, sound. It is what we have been arguing all along. But the obvious starting point is the familiar NZSEE %NBS framework. If 2 is a good adjustment factor (uprating a 20%NBS building to 40%NBS and so getting it over the 34 % NBS line) then we would accept that. But, ideally, the adjustment factor should be based on good science and evidence.

The other obvious problem with the Council using an ASCE 41/13 based assessment for an EPB designation is that it has no basis in Council policy. NZSEE 2006 is referenced but ASCE 41/13 is not.

More generally if ASCE 41/13 is to be used then Wellington would be hostage to US engineering ‘politics’. Standards do not capture just the science but also embed the philosophies and economic interests of the promoters.

Using ASCE 41/13 would also pose a challenge for engineers and lawyers. They would have to become familiar with a whole new standard (it runs to 517 pages plus supporting documents). Our lawyer would be aghast.  He has only recently (just) mastered the Alice in Wonderland mysteries of the NZSEE framework and their legal implications.

Another issue that has arisen is that MBIE wants to be represented at the hearing. We are opposing that. MBIE has had their chance in the Determination, and on that performance has nothing useful to say (read the update ‘I see nothiiing’). MBIE will be making a submission to support their application to be heard.

The Court has also asked the parties to get together to chart a way forward. We have not made any final decisions in this respect, but a productive process could be as follows.

First the numbers.

The appellants have already submitted a detailed seismic assessment, done on a conventional NZSEE basis. The %NBS was 15-25 percent. It is appropriate to take the mid-point of 20 percent as a starting point.

This number needs to be adjusted for the following:

  • A positive adjustment because the building is structurally connected to the adjoining building. The DEA did not take account of this. We could accept a conservative uplift of 5 percentage points in the %NBS.
  • Using the conservative ductility factor of 2 in the Council’s IEP. A ductility factor of 1.25 was used in the DEA, but there was no proper justification for it, and is not accepted as robust by the appellants. Because the ductility factor has a near linear impact on %NBS, the factor of 2 will increase the %NBS by a factor of 1.6.

This gets us to a baseline % NBS of 40.

We then apply the Council’s collapse state conversion factor of 2.  This may well be a low water mark, but we would be prepared to accept it to resolve this case. This gets us to 80 percent NBS.

The 80 percent result is compared with the 34 percent trigger point.

There is such a divergence that the Council and appellants agree that the building could not possibly be earthquake prone.

The EPB designation is removed and there is no need for a court case.