Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Dam Failure Vale Brazil Jan 25, 2019 6

Status
Not open for further replies.
GeoEnvGuy

GeoEnvGuy

Geotechnical
Nov 27, 2018
582
Breaking news, 11 dead hundreds missing after a release of a tailings dam.



Slideshow pictures



Videos of rescue




Vale website

Clarifications on the Dam I of the Mine Bean Stream

Vale is investigating the causes to ascertain why this happened
Vale is investigating the causes to ascertain why this happened. The following information is in regard to the structure and management of the project.

Dam I of Bean Creek Mine is used to dispose tailings from ore production and was located in Brumadinho, Minas Gerais. It was inactive (no tailings were being added), there was no pond and there was no other type of operational activity in progress. The decommissioning project was also under development.

The dam was built in 1976 by Ferteco Mineração (acquired by Vale on April 27, 2001), using the upstream method. The height of the dam was 86 meters and had a crest length of 720 meters. The waste disposal area was 249.5 thousand m2 and the volume disposed was 11.7 million cubic meters.

Dam I had 'Stability Condition Statements' issued by TUV SUD of Brazil, an international company specializing in Geotechnics. The 'Stability Condition Statements' were issued on 06/13/18 and 26/09/18, related to the Periodic Safety Review of Dams and Regular Dam Safety Inspection procedures, respectively, as determined by DNPM Decree 70.389 / 2017. The dam had the Safety Factor in accordance with world best practices and above the Brazilian Standard reference point. Both of these stability declarations attest to the physical and hydraulic safety of the dam.

The Dam went through biweekly field inspections, all of which reported to ANM (National Mining Agency) through the SIGBM (Integrated System for Safety Management of Mining Dams). The last inspection registered on the ANM system was executed on 12/21/18. In addition, it underwent inspections on 1/8/19 and 01/22/19, and was registered on Vale's own monitoring system. The registration of each inspection on the ANM, according to legislation, must be executed by the end of the following fortnightly period. All these inspections did not detect any change in the state of the structure.

The dam had 94 piezometers (an instrument for measuring the pressure of a liquid) and 41 Water Level Indicators to monitor its integrity. The information from the instruments was collected periodically and all their data analyzed by the geotechnitians responsible for the dam. Of the 94 piezometers, 46 were automated.

The dam had a PAEBM (Mining Dam Emergency Action Plan), as established by DNPM regulation 70.389 / 2017. The same was filed in Federal, State and Municipal Civil Defenses between June and September of 2018. The PAEBM was constructed based on a hypothetical breach study that defined the flooding area. In addition, the dam had a video monitoring system, siren alert system (all tested) and downstream population registration. The external emergency simulation was also carried out on 06/16/2018, under the coordination of Civil Defenses, with Vale's full support, and internal training with employees on 10/23/18.

In spite of all the points described above, we are still seeking answers to find out exactly what happened.
 
Replies continue below

Recommended for you

Sort by date Sort by votes
My heart sank as I saw all the relatively small individuals running, or driving, from the oncoming water and debris.

Check out Eng-Tips Forum's Policies here:
faq731-376
 
Upvote 0 Downvote
"The order from Brazil’s mining agency was a preventative measure, Vale said, after engineering company Walm refused to give the mine’s Sul Superior tailings dam a declaration of stability."

It looked a bit like Vale tried to throw TUV SUD under the bus for the Feijao failure (Edit: and there were also the arrests). I'm not surprised the geotechs don't want to sign off now.
 
Upvote 0 Downvote
Bet this was popular with the locals for something that was not an immediate emergency:

Warning sirens sounded and evacuations began at about 1am local time, TV Globo reported. “Attention, attention. This is an emergency … Abandon your homes immediately and follow the escape route,” the announcements said.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.
 
Upvote 0 Downvote
Interesting 1-hr press conference on the Vale link. I'm sure I couldn't keep my composure as well as they did. Aside from that it reads a bit like damage control.

Vale's message seems to be that they did everything they could and that all the proper monitoring and inspections and experts and reports and technologies were in place, and that there were no advance indicators of failure. If that is their position then it begs the question as to how they can simultaneously express confidence in the condition of the other similar dams. I'm not sure it helps their case that they double down on the impeccability of their procedures in the face of an actual failure.

I didn't see any reference to slime encroachment in the press conference but from GeoEnvGuys link regarding Fundao it seems like a possible mechanism, and one that someone would have known about well in advance, notwithstanding any monitoring reports. I think they do refer to a possible seismic event in the press conference but I didn't read/watch the whole thing closely.

 
Upvote 0 Downvote
Interesting that the dam looked like nothing more than a huge pile of dirt with plants growing on it and poof! it all goes in liquid motion.

Seems to me that if that whole thing is half water half solids, in sort of an uneasy truce, that eventually you can have a wetter area migrating to the lower areas which tend to be the face of the damn. If there's a bunch of water there 'somewhere' but you can't actually see it anywhere you will have a very hard time knowing where it is. Clearly, going there and drilling or using seismic mapping are out of the question so you end up blind to what exactly is going on.

From my ignorant position I predict the face reached a prestaged liquefaction state ready for the slightest trigger which the passing train provided that day.

Keith Cress
kcress -
 
Upvote 0 Downvote
I managed to read through some of the liquefaction assessment completed for the site. From what I gather it was done by graduate students using data from a consultant in 2006, made into a masters thesis in 2010 and published in 2013.

They used the methods primarily proposed by Olson in 2001 and 2003. They used SPT and CPT data to determine which layers would be liquefiable. They also correlated a shear strength ratio with the data.

They then did 2d limit equilibrium analysis and reduced the shear strength of the liquefiable layers until they got a factor of safety of 1. They compared the shear strengths from the correlation to the stability results and got a factor of safety between 1.1 and 1.3.

My thoughts on this start with graduate students doing work, I hope actual engineers also did the job and Vale wasn't relying dam safety to students. Looking at the shear strengths determined they are on the upper end of what I would expect as peak or yield shear strength of loose tailings. Those shear strengths also represent the steady state condition so a factor of safety of 1.1 to 1.3 is low, if they would have used liquefied shear strengths, even the correlation by Olson in 2003 the factor of safety would be less than 1 which would be the post earthquake condition which anything above 1 is good. Also if you read Olson's work he provides a plus minus on his shear strength correlation. If you use the minus the factor of safety by the authors may be closer to 1.


Looking at the geometry in the stability analysis it's odd that dilative (strong) tailings are in the middle area which is traditionally slimes as it's full of uncompacted tailings settling in the pond. The failure surface is rounded where it should be a sliding block along the weak liquefiable layer.

Thinking about the state of the art in liquefaction most of the new methods were published between 2008 and 2014 by the Berkley university group. Newer research would only increase the amount of material inferred to be liquefiable. Provide new correlations for liquefied shear strength ratios in the 0.05 to 0.10 and peak or yield shear strength ratios between 0.2 and 0.25, which is close to what the authors determined.

Now that we have some hindsight, the failure geometry from the video shows a large failure surface from the top beach area starting upstream of the compacted perimeter area to a lower bench. This was not shown in the authors stability analysis. Instead the failure surfaces shown started on the middle of the slope through an inferred perimeter dam and out along the bottom of the slope in an offset bench area.
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

JStephen
  • Question
Bridge collapse- Brazil
Replies
2
Views
2K
Sym P. le
Sym P. le
JohnRBaker
  • Locked
  • News
A plane crash in Brazil, near Sao Paulo...
Replies
15
Views
274
Alistair_Heaton
Alistair_Heaton
azcats
  • Locked
  • Question
Rapidan Dam - MN 1
2
Replies
22
Views
387
Heaviside1925
Heaviside1925
Alistair_Heaton
  • Question
Power infrastructure failure cuba
Replies
10
Views
3K
LittleInch
LittleInch
M
  • Question
2023 Osprey crash in Japan - failure report
Replies
1
Views
2K
drwebb
drwebb

Part and Inventory Search

Sponsor

Back
Top Bottom