Blast Vibrations

[tq3610]

In modeling seismic slope stability, it is "typical" to model the horizontal acceleration coefficient of 0.5xPGA(site modified). It is my understanding that the 0.5 factor applied to PGA is to account for short duration of loading. My question is, if a site has site-specific ground vibration data via seismographs for nearby blast vibrations, is it appropriate to apply the 0.5 factor to the recorded peak ground acceleration?

If my seismograph as a PPA of 0.18g, should my stability model use kh = 0.18 or kh = 0.09?

Your thoughts are much appreciated.

 

[BigH]

Keep to posting in only one forum - not two. Almost all geotech engrs check all geotech forums.

 

[tq3610]

Will do - deleted from the other. My apologies.

 

[GeoEnvGuy]

My thoughts are that completing a pseudo static stability assessment is a waste of time and does not provide anything truly useful beyond the Hynes Griffin purpose of estimating the crest settlement of 1 m for deep seated failures in embankments where the pseudo static factor of safety is equal to 1.

What is useful is looking at the soil profile and determining which portions may be susceptible to liquefaction and performing a triggering assessment with your PGA. Where liquefaction is inferred run a stability model with liquefied strengths. If there is no liquefaction or your slope still stands up then look at post earthquake settlements with your PGA.

 

[tq3610]

Thank you GeoEnv - your comment is understood. Your point is taken and I had already considered this - if 1m of displacement is too much, then what is the point of taking that approach...

That said, I can also perform a Newmark analysis to evaluate slope displacements. The input of the time v. acceleration plot is obviously of utmost importance to an "accurate" (loosely used term) analysis. In considering running the displacement analysis, it is my expectation that actual measured time v acceleration data should be used. I'm hoping to get some sort of confirmation of that, and if there is disagreement, some input as to why to not use those values.

Per your comment above - "look at post earthquake settlements with your PGA." That PGA needs to be translated to kh, and that translation is ultimately the question that I'm asking. Should the PGA be reduced in some form given that we have site-specific ground accelerations?

 

[GeoEnvGuy]

Regardless of the consequences if the materials are contractive and can liquefy simplified methods are not appropriate which would warrant spending more effort to estimate the deformation or accept that there will be wide spread failure and remediate the slope.

If there are no contractive materials then the level of effort put in to estimate deformations is up to the Owner which is typically related to the consequences of failure. If your project warrants the effort you can perform a nonlinear finite element deformation analysis with software packages such as Flac, but this would require significant investigation and lab testing. You can also perform a site specific numerical newmark analysis with time histories using software packages such as quake/w slope/w. Or as a screening level a simplified approach using only the parameters you have to get an order of magnitude estimate such as the Makdisi seed, or Bray and Travarasou method which can be found on Dr. Bray's website.

https://ce.berkeley.edu/people/faculty/bray/research

 

[EireChch]

Whats the source of blast loading? Is it a quarry?

Are you in a seismic area? If you are you will need to look at your seismic PGA and do an analysis based on that as well.

Is groundwater a consideration on your slope? Will liquefaction be an issue?

If it is a seismic zone - You would have two seismic (or dynamic?) slope stability analysis to do. (1) seismic (2) blast

I would suspect that the Seismic slope stability analysis would be your more conservative (ULS) design case. Is a psuedostatic analysis really useless for the seismic case. With a FoS against failure of 1 you should have 1m displacement. is that not ok for the seismic case. You wont have a dramatic landslide, you will have 1m of displacement ? Its ULS case, who cares about 1m of displacement?

Your blast loading case is likely your servicability (SLS) load case. I agree then for this SLS case that 1m of displacement isnt acceptable as you will likely have may blast events that may induce displacement. Its then hard to see that the 0.5 reduction in PGA is applicable to a non seismic case?

PS your data is likely from an accelerograph not a seismograph.

 

[tq3610]

Thanks Eire -

You are correct - both seismic and blast were considered separately. The earthquake event I would consider to be much different than the blast event (duration of vibration 10s versus blast 1s, and frequency of occurrence 500 yrs versus many times per year). So, my consideration is that 1m of displacement is acceptable for a 500 yr event, but certainly not for a "normal" occurrence. That "normal" occurrence of the blast I believe needs to keep displacement to zero, and have some degree of built in safety factor to account for uncertainty.

As far as the accelerograph v. seismograph, I'm not quite sure what the "difference" is here. There is a seismograph (my term I suppose) that we have blast vibration data from. The seismograph has a geophone with accelerometer if that is what you are referring to.

 

[EireChch]

Your approach seems reasonably to me in that the blast condition is your normal activity so definitely dont want any displacement. That to me would mean not applying the 0.5 reduction. But hey, I have never needed to consider it so could be wrong!!

Interesting project though. Good luck




comment on seismograph / accelorgraph was an unimportant side comment. I watched a lecture recently that indicated that seismograh was the old school way of measuring EQs.