Report of the Expert Panel on the Technical

[pic 1]

Expert Panel:

Peter K. Robertson, Ph.D. (Chair) Lucas de Melo, Ph.D.

David J. Williams, Ph.D.

G. Ward Wilson, Ph.D.

December 12, 2019

EXECUTIVE SUMMARY

At approximately 12:28 p.m. local time on January 25, 2019, tailings dam B-I at Vale S.A.’s Córrego do Feijão Iron Ore Mine (“Dam I”), located 9 kilometers (km) north-east of Brumadinho, in the state of Minas Gerais, Brazil, suffered a sudden failure, resulting in a catastrophic mudflow that traveled rapidly downstream.

This dam failure is unique in that there are high quality video images of the event that provide insight into the failure mechanism. The videos clearly show a slope failure within the dam starting from the crest and extending to an area just above the First Raising (the Starter Dam). The dam crest dropped and the area above the toe region bulged outwards before the surface of the dam broke apart. The failure extended across much of the face of the dam and collapse of the slope was complete in less than 10 seconds, with 9.7 million cubic meters (Mm3) of material (representing approximately 75 percent (%) of the stored tailings) flowing out of the dam in less than 5 minutes (min). The material in the dam showed a sudden and significant loss of strength and rapidly became a heavy liquid that flowed downstream at a high speed. The videos show that the initial failure was relatively shallow and was followed by a series of rapid shallow slips with steep back slopes that progressed backwards into the tailings impoundment. Based on these observations, it is clear that the failure was the result of flow (static) liquefaction within the materials of the dam. The significant and sudden strength loss indicates that the materials within the dam were brittle.

The failure is also unique in that it occurred with no apparent signs of distress prior to failure. High quality video from a drone flown over Dam I only seven days prior to the failure also showed no signs of distress. The dam was extensively monitored using a combination of survey monuments along the crest of the dam, inclinometers to measure internal deformations, ground-based radar to monitor surface deformations of the face of the dam, and piezometers to measure changes in internal water levels, among other instruments. None of these methods detected any significant deformations or changes prior to failure. Post-failure satellite image analyses indicated slow and essentially continuous small downward deformations of less than 36 millimeters per year (mm/year) were occurring on the dam face in the year prior to the failure, with some acceleration of deformation during the wet season. In the lower part of the dam, the deformations measured in the 12 months prior to failure included horizontal deformations ranging from 10 to 30 mm. Such deformations are consistent with slow, long-term settlement of the dam, and would not alone be indicative of a precursor to failure.

The dam’s construction history provides insight as to the possible reasons for the failure. The dam was constructed using the upstream construction method over a period of 37 years in 10 raises. No new raisings were constructed after 2013, and tailings disposal ceased in July 2016. The Starter Dam contained features that impeded drainage through the toe. No significant internal drainage was installed during the construction of later raisings, other than small drainage blankets below some of the later raisings, and chimney drains in some of the upper raisings. The drainage blankets and chimney drains in the later raisings were in response to observed seepage from the dam face above the toe during construction. The initial design of the dam established a relatively steep slope. After the Third Raising, a setback was constructed to straighten the alignment of the dam crest. The setback reduced the overall slope of the dam but moved the upper portion of the dam closer to

the pond and also closer to the future internal water level. Pre-failure aerial and satellite images show that, at times during the life of the dam, water was close to the crest of the dam, resulting in weak tailings close to the crest and interbedded layers of fine and coarse tailings within the dam. The setback also moved the upper portion of the dam over weaker, finer-grained tailings.

The lack of significant drainage features, coupled with the presence of less permeable fine tailings layers within the dam itself, resulted in the dam having a high water level. Seepage from the dam face above the toe was observed periodically from as early as the Fourth Raising. Despite tailings deposition ceasing in mid-2016, review of piezometers installed within the dam showed the water level within the dam did not reduce significantly after tailings deposition ended. The water levels in the upper portion of the dam were slowly dropping, but remained high in the toe region. This was predominately due to the high regional wet season rainfall combined with limited internal drainage in the dam. Deep horizontal drains (DHPs, based on their Portuguese language term) were installed in early 2018, and a total of 14 was installed, mostly along the toe of the setback. Following an incident during the installation of DHP 15, further installation of DHPs was not pursued.

Data from pre-failure geotechnical investigations were significant and included drilling, sampling, cone penetration tests (CPTu), field vane shear tests (FVT), and in situ shear wave velocity (Vs) profiles. These data provided detailed information regarding the nature, consistency, and distribution of materials, as well as water pressures, within the dam. These data, combined with available aerial and satellite images, allowed the Panel to develop detailed two-dimensional (2D) and three-dimensional (3D) stratigraphic profiles of the materials within the dam.

The Panel carried out new investigations to provide additional information about the materials within and under the dam. One significant finding was that the tailings within the dam had very high iron (ferrous) content (greater than 50%), with very little quartz (less than 10%). The high iron content gave the tailings a high total unit weight of approximately 26 kilonewton per cubic meter (kN/m3). The historical CPTu data, together with appropriate unit weights and water pressures, indicate that the tailings were predominately loose, saturated, and contractive at large strains. Advanced laboratory testing carried out as part of the Investigation on representative reconstituted samples of the tailings showed brittle strength loss behavior and indicated the presence of bonding. This brittle strength loss also was observed in historical data from FVTs, CPTu, and in some historical laboratory test data. Scanning electron microscope (SEM) images analyzed by the Panel attributed this brittle behavior to bonding, most likely due to iron oxidation. The advanced laboratory testing also showed that loose samples of tailings would accumulate strain under constant load. This accumulation of strain under constant load is referred to as creep. In summary, the tailings were loose, predominately saturated, and bonded. The bonding rendered the tailings stiff and potentially brittle. Combined, these features resulted in a material that had potential for significant and rapid strength loss with ongoing strain. The stiff, brittle character of the tailings is consistent with the lack of observable deformations prior to failure and the sudden, rapid response at failure.