Final Report - Fatality #12 - August 17, 2012

OVERVIEW

On August 17, 2012, Pierre (Sonny) Mezidor, Cement Equipment Operator, age 58, was killed when the silo roof he was working on collapsed. The cement roof slab, beams, and the grout in the beam pockets failed causing the roof and the equipment on the roof to fall into the 3/4 full silo. Rescuers responded and Mezidor was recovered from the silo on September 4, 2012.

Mezidor was on top of Silo 12 to measure the level of material in the silo. Miners had to manually check the silo levels several times each shift because the radio transmitter level indicator and the high level indicators for the silo were not functional.

The accident occurred due to management’s failure to correct defects on the silo where the victim was working. The roof decking on Silo 12 was inadequately attached to the roof beams, causing the beams to become unstable and buckle; the grout under the beam ends was too thick and some of the grout in the beam pockets had cracked and delaminated; and the shear stirrups were placed too far below the beam ends to prevent the grout in the beam pockets from failing in shear. Also, in 2004 a roof beam was cut when a penetration was made into the roof to install an automatic level detecting device. This cut significantly reduced the load carrying capacity of the beam. These defects led to the collapse of the silo roof slab that was supporting the victim. In addition, adjacent Silo 11 had experienced a partial roof failure in 2011 that caused the roof to bulge and caused significant cracking in the reinforced concrete. Management did not adequately investigate the causes and conditions surrounding the Silo11 partial roof failure and therefore did not identify similar conditions existing under the roof of Silo 12.

Management also allowed the silos to operate with defective aeration systems and the presence of a large rathole (partly caused by those defective systems). When the cement would bridge over the rathole and then collapse during discharge, this would create significant suction loading and detrimental vibrations on the roof slab, beams, grout in the beam pockets, and welds. In addition, other equipment including high level indicators and an automatic level detection device were inoperable, making manual measurement necessary and the over pressure events more likely. The over pressurization in the silo resulted in upward pressures on the silo roof slab and its support system (puddle welds and beams), and damage to those components.

When combined, these equipment-related defects and structural deficiencies contributed to the collapse of the roof slab.

GENERAL INFORMATION

PENNSUCO Cement Plant, a hydraulic cement facility operated by Tarmac America, LLC, and owned by Titan Cement Company SA, is located in Medley, Miami-Dade County, Florida. At the time of the accident, the principal operating official at the plant was Kevin Baird, General Plant Manager. The plant operates 24 hours a day, 7 days a week. Total employment is 132 persons. An adjacent quarry, PENNSUCO Quarry, supplies the plant with raw material. PENNSUCO Quarry employs 164 persons.

A series of belt conveyors transport the raw material to the kiln, where the rock is crushed, sized and mixed with the other raw materials. The kiln heats the mixture to about 2,700 degrees F to produce the clinker. The clinker is conveyed to another part of the plant where it is ground with a small amount of gypsum into a powder. This powder, or Portland cement, is then fed into one of 12 silos.

The Mine Safety and Health Administration (MSHA) completed the last regular inspection at the plant on August 2, 2012.

DESCRIPTION OF ACCIDENT

On the day of the accident, Pierre (Sonny) Mezidor started his shift at 7:00 a.m. Following a routine 15 minute safety meeting, Mezidor talked with Dolores Otero, Control Room Operator, who briefed Mezidor on the status of the silos. Mezidor then went to the maintenance trailer and gave a fan, that needed repair, to Alejandro Ortiz, Cement Shift Repairman.

About 7:40 a.m., Otero called Mezidor on his radio and told him to go to Silo 12 to confirm a measurement of material in the silo. The mine operator’s inventory record showed the silo to be nearly full, but a measurement taken one hour before the collapse showed the silo to be nearly empty. Mezidor stopped at the loadout control room at the base of Silos 1-9, talked with the miners there, and then used the elevator to travel to the top of Silos 1-9 at 8:05 a.m. Mezidor walked across Silos 1-9 and climbed the ladder on the side of Silo 10. He crossed Silos 10 and 11 and reached Silo 12 shortly before 8:10 a.m. Silo 12 collapsed soon after Mezidor reached the silo. Surveillance video showed the collapse occurred at 8:11 a.m.

Lazaro Sainz, Cement Repairman, and Reynerio Martinez, Cement Repairman, were working on a conveyor at the new plant when they saw the dust collector fall into Silo 12. Martinez called Otero at the control room to let her know the dust collector on the roof of Silo 12 fell into the silo. Otero informed Martinez that Mezidor was measuring Silo 12. Sainz tried unsuccessfully calling Alberto Hernandez, Maintenance Supervisor. Otero tried contacting Mezidor several times on the radio.

Otero began telling miners over the radio that there was a problem on Silo 12 and that Mezidor was missing. At approximately 8:20 a.m., Otero called Kevin Baird, General Plant Manager, to report the roof of Silo 12 had collapsed and that Mezidor was on the silo when it collapsed. Baird went to the loadout and met with Hernandez, Jeff Harris, Assistant Cement Production Manager, and other miners, and took the elevator to the top of Silos 1-9. They reached the top of Silo 11 at approximately 8:30 a.m.

Harris contacted Marco Burgoa, Technical Manager, soon after reaching the top of Silo 11. Burgoa arrived on top of Silo 11 approximately 8:40 a.m. The fire department was already on-site when Burgoa made it to Silo 11 and rescue efforts began.

INVESTIGATION OF THE ACCIDENT

MSHA was notified of the accident about 8:30 a.m. on August 17, 2012, when Burgoa initially called MSHA’s national office in Arlington, Virginia. MSHA’s National Call Center was notified at 8:45 a.m. by a telephone call from Rene Hernandez, PENNSUCO Cement Plant. The National Call Center notified Patrick Sharp, Southeastern Health Specialist, and an investigation started the same day. At 9:31 a.m., MSHA issued an order under the provisions of 103(j) of the Mine Act to ensure the safety of the miners. At 2:55 p.m., soon after the first Authorized Representative (AR) arrived at the mine, this order was modified to section 103(k) of the Mine Act.

MSHA sent two inspectors from the Bartow Field office to secure the accident scene until the accident investigation team could arrive. Before the end of the day, the first two members of MSHA’s accident investigation team, from Kentucky, arrived at the mine and made a physical inspection of the accident scene. The following day additional accident investigation team members arrived on the mine site. During the course of the next 28 days, the accident investigation team assisted in the recovery of the miner, interviewed employees, inspected recovered debris, and reviewed documents and work procedures relevant to the accident. MSHA conducted the investigation with the assistance of mine management and employees.

The Miami-Dade Police Department and Fire Department had members on the property from the time they were notified until the victim was recovered on September 4, 2012. Throughout the investigation, these two local Departments participated in the daily discussions with MSHA and assisted the mine operator in moving forward with the rescue and recovery efforts.

DISCUSSION

The conclusions in the body of this report are compiled from the interview findings and the technical investigations. Details of the engineering and technical investigation that support the findings of this report are located in Appendix C.

Location of the Accident

The accident occurred on Silo 12. It was one of three silos (10, 11and 12) that were constructed in 1981. These three silos were interconnected and oriented in a line. Silo 11 was the middle silo and it shared a common wall with Silo 10 on its northern side and Silo 12 on its southern side.

Surveillance and Time of Collapse

Security cameras monitoring the plant captured the collapse from a distance. The time of the collapse, ascertained from viewing the footage, indicated the roof failed at 8:11 a.m. Based on the footage, it appeared that a majority of the dust plume, once the roof collapsed, came from the east side of the silo. Just prior to the collapse, there appeared to be a haze in the vicinity of the dust collector. It was not clear whether this was sun reflections off the east side of the collector or a localized haze of dust leaking near the collector that was being illuminated by the rays of sunlight.

Weather

The weather on the day of the accident was clear with a slight breeze and a temperature about 85 degrees Fahrenheit. Weather was not considered a contributing factor to the accident.

Rescue and Recovery

After the collapse, the cement product level near the silo walls was approximately 35 feet below the roof. The victim was found at a depth of approximately 51 feet below the top of the silo in the region of the silo slightly north of the center. He was below most of the debris, indicating that he likely fell off the slab as it was tilting into the silo, but before it pulled completely free from the ends of the perimeter vertical reinforcing bars that were bent into the silo. The depth to the recovery point was in stark contrast to the depth of 148 feet measured about an hour before the collapse. This information suggested that the taped reading of 148 feet was into a flow channel, which then filled in once the peaked perimeter of cement build-up collapsed into the channel void at the time of the failure.

PENNSUCO contracted G&R Minerals, a construction company from Birmingham, Alabama, to manage the recovery operations. Two cranes were on site by the end of the first day of the recovery. One crane was used to lift debris out of the silo. The second one had a man basket attached that could lower persons in and out of the silo. For the first eight days of the debris removal, the process involved cutting and breaking large pieces of material to a size the crane could lift out of the silo.

Several times during the first eight days of the recovery, the Miami-Dade Fire Department lowered their cadaver dog and trainer into the silo to search for Mezidor.

After the eighth day of debris removal, a clam shell was used to remove the loose material. The clam shell had the capacity to remove up to 15 cubic yards (around 15 tons of cement) per lift. As the extraction progressed, the center portion of the silo material moved down at a constant rate. However, the outer edge of the material remained firm. On the eleventh day, the operator lowered a Brokk machine into the silo to break the hardened material from the outside edges. In two days, hardened material was scaled down to the center material.

During the clam shell extraction phase of the recovery, an average of 10 tons of material per 30 minute cycle was removed from the silo. This cycle included the removal of the material from the silo, dumping the material into a waiting dump truck at the base of the silo, and delivering the material to a flattened area to be dumped on the ground. A crew of miners was at the dump site with rakes and shovels moving the debris around to locate the victim. Cadaver dogs and their trainers were at the dump site during this entire process.

At the end of August, 2012, tropical storm Isaac dropped more than 10 inches of rain at the accident site. The excessive wind speeds and rain from the storm stopped the use of the cranes and slowed recovery efforts for two days.

At approximately 2:30 a.m. on September 4, 2012, 18 days following the accident Mezidor was located. The Miami-Dade Medical Examiner pronounced the victim dead at the mine site.

General Information Silos 10, 11, 12

Silos 10, 11, and 12 were cement storage silos designed by R.S. Fling and Partners, Inc.1 in 1980 and built by S&W Construction Company2 using reinforced concrete. The silo was constructed in 1981 for the original plant owners, Lone Star Industries, Inc. Law Engineering Testing Company was engaged to serve as the owner’s representative for the foundation construction phase of the three silos. The silos were partially interconnected and oriented in a line. Silo 11 was the middle silo and shared a common wall with Silo 10 on its northern side and Silo 12 on its southern side. The silos were approximately 189 feet high, as measured between the foundation and the top of the silo roof. They had an approximately 56 foot outside diameter (photo 1). The silo walls were 14 inches thick and reinforced vertically with 2 layers of No. 4 bars at 14 inches on centers. The horizontal reinforcing hoops varied in size from No. 4 to No. 8 bars and spacing from 6 to 12 inches on centers from the top of the silo to the bottom. At each elevation there were two layers (an inner and outer hoop).

The internal storage of the silo was less than the overall volume of the cylinder, because in each of the silos there was a 34.6-foot-high conically-shaped steel hopper supported by an elevated reinforced concrete floor. The floor was supported by 14 internal columns varying from 36 to 42 inch diameter. Thehoppers were sloped at a 55º angle and had a single 6 foot diameter bottom discharge. The hopper plates were 3/16 inch thick at the top of the hopper and 3/8 inch thick at the bottom. The plates were ½ inch thick where they passed through the elevated floor. While the elevated floor supported the hopper plate directly as it passed through the floor, the hopper was supported above the floor by a lean concrete fill material that formed to the conical shape of the hopper plates. The lean fill rested on the elevated floor. The actual live cylindrical storage area above the hopper was 135.2 feet high.

The combined volume of the live cylinder and conical hopper as per the calculations by R.S. Fling was 327,451 cubic feet. Each silo could store 16,147 tons of Type I/II cement powder, as per the silo inventory datasheet provided by the mine operator. The cement powder had a stored density of 94 pounds per cubic foot.

The silo roof was a reinforced concrete slab with a varying thickness (photo 2). The slab was designed to be 4½ inches thick at the edge and 11½ inches at the center to create a slope for drainage. These thicknesses included the 1½ inches of depth of the corrugated metal floor decking (specified as 16/16 GA) that supported the wet concrete while it hardened during initial construction. The gauge ratio referred to the thickness of the top hat plate and the thickness of the flat bottom sheet of the embossed cellular decking, which were both 16 gauge (equal to 0.06 inches). The center to center spacing of the corrugations in the roof decking was 6 inches. The top hat geometry was B-LOK. The decking sheets were as much as approximately 31 feet long and spanned north to south over multiple roof beams. The roof slab was reinforced by four layers of reinforcing steel, with two of the layers being oriented north to south and the other two layers east to west. The bars were specified as No. 4 bars on 18 inch centers. The concrete compressive strength was specified at 3,500 psi and the reinforcing steel at 60,000 psi yield strength. The roof slab was designed for its self-weight plus 50 psf3 live loading. The slab was tied to the silo walls via continuity of the vertical reinforcing steel bars. Specifically, the two layers of vertical bars in silo walls were extended three feet above the top elevation of the wall. These extensions were then bent over into the roof slab to become part of the slab pour. Aside from this continuity of the wall reinforcing into the roof slab and the bearing around the perimeter, the primary means of support for the slab came from the six underlying steel roof beams.

The silo roof beams spanned east/west and rested in wall pockets left open at the top of the wall (figure 1). Within the pocket, a bearing plate was installed. The beams’ ends transferred their reactions to the bearing plates and ultimately into the silo wall. There were a total of twelve pockets, six on the east side and six on the west side. During the investigation, these pockets were labeled as East 1 (E1) though East 6 (E6) and West 1 (W1) through West 6 (W6). The beams were likewise labeled as Beams 1 through 6, with Beam 1 being the southernmost and Beam 6 the northernmost. For consistent nomenclature, Beam 1 was supported by E1 and W1. Likewise, Beam 2 was supported by E2 and W2, etc. All the structural steel was grade A36. Beams 1 and 6 were W24x55 sections. Beams 2 and 5 were W27x84 sections and Beams 3 and 4 were W30x99 sections.

The outer Beams 1 and 6 were shallower beams, since they only spanned 39.2 feet. The middle two deepest Beams 3 and 4 spanned 54.8 feet, whereas intermediate Beams 2 and 5 spanned 50.2 feet.

A pair of short extension channels was bolted to the ends of each of the beams using between six to nine A325 high strength bolts. These channel extensions then rested on the bearing plates cast inside the beam pockets. The channels at the ends of Beams 1 and 6 were C10x15.3 and the channels at the ends of Beams 2 through 5 were C12x20.7. These deeper channels were needed for the middle four beams, since they were carrying higher end force reactions (shears). In addition, each of the ends of the six roof beams was coped (notched) near the location of the bolted connection with the channels. The length of the channel extensions was 13 to 16 inches.

1. R.S. Fling and Partners is no longer in business.
2. S&W Construction Company is no longer in business.
3. psf = pounds per square foot

Factors Causing the Collapse

As evident in the photos taken when the silo was being drawn down during recovery and based on disparities with the silo measurements, Silo 12 had a considerable amount of non-flowing static material that was partially cohesive and somewhat compacted. The non-flowing material occupied a considerable amount of the stored material and formed a large rathole (an area created in the silo where material moved mostly through the center and accumulated around the edges). With the restricted flow channel, the cement powder was able to bridge/arch over the opening during drawdown. When the arch collapsed, it created a large dynamic suction of the roof. Erratic flow over the life of the silo created dynamic suction and rebound loads (vibration) on the silo roof slab, roof beams, puddle welds and beam pockets.

Several factors contributing to the buildup of material and flow problems were attributed to the aeration piping system on the sides of the hopper, which was intended to help prevent rathole formation but had not worked for a long time. The silo was full for over three weeks prior to the accident, which allowed the material to settle and compact. Additionally, the air blaster at the hopper outlet was using ambient air exposed to moisture, rather than thermally dried air. The ambient air contained humidity, which caused the cement particles to begin to hydrate and stick together, thereby helping form a rathole.

The radio transmitter level indicator and the high level indicators were not functional, causing additional exposure to workers who had to measure the silo depths manually each day. On the morning of the accident, flow from Silo 12 unexpectedly stopped, even though the silo was found to contain a large amount of cement product. The truck loading operator could no longer fill customers from Silo 12.

The dust collector on Silo 12 was running constantly even though no new material was being pumped into the silo. This condition created a negative suction of 35 psf on the roof and the cement material, which was then partly lost when the silo port was opened for measurement. This change in pressure likely caused the cement powder in the upper reaches of the silo to destabilize and fall into the flow channel creating a significant suction load on the roof.

Although the original contractor generally provided the roof deck puddle welds on 12-inch centers, many of the welds were not of good quality. Specifically, while the weld metal was deposited on the top flange of the beam, the weld material had porosity and did not fuse with the steel in the beam in many instances. In some cases, it appeared that the burn holes in the decking did not penetrate both layers of decking. Therefore, the decking was not adequately attached at these areas. Both downward loading and previous instances of over pressurizing the silo (upward loading), likely caused the puddle weld connections to fail either through the weld, at the base of the weld, or through the decking material surrounding the weld. Aside from overload of the welded connections, fatigue (suction and rebound vibration loadings during flow) may have also contributed to some of the connection failures of the decking to the beams. Once the welded connections were lost, the beams were then not adequately braced to prevent lateral torsional buckling from downward loading, which likely occurred prior to the day of the collapse. Also, as mentioned above, if the welds were able to transmit the tension from the underside of the slab to the beams during an overpressure event, then the bottom flanges of the beams would be in compression. Since the bottom flanges were un-braced, the beams could fail in lateral torsional buckling, even from this upward direction of loading.

Post collapse investigation of the condition of identical Silos 10 and 11 found several roof beams buckled and detached from the roof slab. In addition, hardened cement was found on the top flanges of Silo 12 beam numbers 1, 2, 3, 4, and 6 near the buckled areas, indicating these beams were most likely buckled to some extent prior to the day of the collapse. A deflected buckled beam was unable to carry as much of the load as it did prior to the buckle, particularly if the buckle was severe enough to lose contact with the roof slab.

The top flange of Beam 3 in Silo 12 had been cut when a penetration was made after the silo was built. The cut reduced the section modulus by 30%, the compression flange cross sectional area by nearly 50%, and the general resistance to lateral torsional buckling. Hardened cement on the top flange of the beam and within the bent torch cut marks on the flange appeared to indicate that the beam had buckled some time prior to the day of the collapse.

In the original design, the 35 psf dust collector suction pressure should have been considered as a separate load in addition to the 50 psf live loading, rather than incorporating, by default, the suction within the 50 psf live loading. The original design did not specify supplemental bracing or minimal puddle weld connection strength to prevent lateral torsional buckling of the roof beams. In addition, a pressure relief system was not specified.

As the beams buckled, the end reactions on the beam pocket bearing plates increased when the load from one beam shed to adjacent beams. The flanges on the end channels were bent on at least one end of five of the six roof beams, indicating the beam ends were attempting to rotate prior to the final failure. Bending of the flanges applied point loading to the bearing plates. The buckling also likely caused an inward frictional drag force on the bearing plates, which was then transferred via the studs to the underlying unreinforced grout layer. Most of the beam pockets were found to be delaminated. The internal cracking from delamination reduced the shear resistance of the pocket by decreasing the length of the shear resistance path. In addition, as a result of a construction defect, the shear stirrups in the wall were placed too far away from the bearing plates, since the as-constructed thickness of the grout was much greater (11 inches instead of 1.5 inches) than shown on the engineering drawings. The delamination and the lack of shear stirrups within the shear plane of the pockets allowed at least two of the pockets, E2 and E3, to fail in shear and the bearing plate on W5 tilted down. When the east side pockets failed, the reinforced concrete slab could not support itself and the roof slab collapsed.

The investigators did not find any evidence or reports of previous structural inspection of any of the three silos. Even following an overpressure event in 2011 that caused the roof on Silo 11 to bulge up, and despite the significant damage, no structural inspection was conducted of the roof beams and pockets. An inspection at that time would have uncovered the problems with the buckled beams, puddle weld failures, and delamination in the beam pockets, which then would have caused concern regarding the integrity of adjacent Silos 10 and 12, and could have identified the deficiencies that led to the roof failure of Silo 12.

Training and Experience

Pierre Mezidor had 19 years of mining experience, all at this mine. A representative of MSHA’s Educational Field Services staff conducted an in-depth review of the mine operator’s training records. The training records for Mezidor were reviewed and found to be in compliance with MSHA training requirements.

ROOT CAUSE ANALYSIS

Investigators conducted a root cause analysis and the following root cause was identified.

Root Cause: Management failed to correct defects on Silo 12 where the victim was working. Specifically, the operator failed to determine the cause of a partial roof failure the previous year on an adjacent silo. Silos 10, 11, and 12 had not been built to specifications; maintained appropriately; and the operator added additional strain on the silo by allowing the silo to operate with an internal rathole and by subjecting the silo to overpressures during its life of operation. In addition, a modification had been made to one of the roof beams on Silo #12 that significantly weakened the roof.

Corrective Action: Silos 10, 11 and 12 at this mine have been under a Section 103(k) order since August 18, 2012, and have been the subject of an ongoing investigation. The operator must take the actions necessary to repair these three silos, and with the knowledge learned from this investigation prevent future unsafe conditions in the remaining silos at the mine. MSHA will require the operator to take appropriate actions to address the root cause of this accident to ensure that miners can safely work on or near Silos 10, 11, and 12.

CONCLUSION

The accident occurred due to management’s failure to correct defects on the silo where the victim was working. The roof decking on Silo 12 was inadequately attached to the roof beams, causing the beams to become unstable and buckle; the grout under the beam ends was too thick and some of the grout in the beam pockets had cracked and delaminated; and the shear stirrups were placed too far below the beam ends to prevent the grout pockets from failing in shear. Also, in 2004 a roof beam was cut when a penetration was made into the roof to install an automatic level detecting device. This cut significantly reduced the load carrying capacity of the beam. These defects led to the collapse of the silo roof slab that was supporting the victim. In addition, adjacent Silo 11 experienced a partial roof failure in 2011 that caused the roof to bulge and caused significant cracking in the reinforced concrete. Management did not adequately investigate the causes and conditions surrounding the Silo11 partial roof failure and therefore did not identify similar conditions existing under the roof of Silo 12.

Management also allowed the silos to operate with defective aeration systems and the presence of a large rathole (partly caused by those defective systems). When the cement would bridge over the rathole and then collapse during discharge, this would create significant suction loading and detrimental vibrations on the roof slab, beams, grout in the beam pockets, and welds. In addition, other equipment including high level indicators and an automatic level detection device were inoperable, making manual measurement necessary and the over pressure events more likely. The over pressurization in the silo resulted in upward pressures on the silo roof slab and its support system (puddle welds and beams), and damage to those components.

When combined, these equipment-related defects and structural deficiencies contributed to the collapse of the roof slab.

ENFORCEMENT ACTION

Issued to Tarmac America, LLC

Order No. 8720505 – issued on August 17, 2012, under provisions of Section 103(j) of the Mine Act. This Order was modified later that same day to Section 103(k) of the Mine Act. Nine additional modifications were made to this Order during the recovery operations after the accident. After the recovery of Mezidor on September 4, 2012, several additional modifications to this Order were made to allow continued investigations of the remaining silos. The conditions that contributed to the accident may still exist on the adjacent silos; therefore, the Order continues to remain in effect as of the issuance of this report.

Citation No. 8724257 -- issued under the provisions of Section 104(a) of the Mine Act for a violation of
30 CFR 56.14100(b):

A fatal accident occurred at this mine on August 17, 2012. The victim was on top of Silo #12 getting a depth measurement of the cement in the silo when the roof on the silo collapsed. Following the accident, investigators determined that the hopper aeration piping system was inoperable and that the hopper air blaster was using ambient air rather than thermally dried air to agitate the powder at the bottom of the silo. The moisture caused erratic flow of material and the stored cement hardened, forming a rathole inside the silo. The cement would bridge over the rathole and then collapse during discharge, creating significant suction loading and detrimental vibrations on the roof slab, beams, pockets, and welds. These conditions contributed to the eventual collapse of the silo. Defects on the hopper aeration piping system , the hopper air blaster, and the presence of a rathole were not corrected in a timely manner to prevent the creation of a hazard to persons working on the roof of Silo #12.

Citation No. 8724258 -- issued under the provisions of Section 104(a) of the Mine Act for a violation of
30 CFR 56.14100(c):

A fatal accident occurred at this mine on August 17, 2012. The victim was on top of Silo #12 getting a depth measurement of the cement in the silo when the roof on the silo collapsed. Following the accident, investigators determined that the roof decking on Silo 12 was inadequately attached to the roof beams, causing the beams to become unstable and buckle; the grout under the beam ends was too thick and some of the grout pockets had cracked and delaminated; and the shear stirrups were placed too far below the beam ends to prevent the grout pockets from failing in shear. Additionally, during a previous construction project, the top flange of one of the 30-inch center roof beams was accidently cut, considerably reducing the strength of this beam; and this significant damage was not repaired. These defects made continued operation of the silo hazardous to persons but the silo was not taken out of service and continued to operate. These defects caused the eventual collapse of Silo #12.

APPENDIX A

Persons Participating in the Investigation

Tarmac America, LLC Pennsuco Cement Plant

G&R Minerals

Ogletree, Deakins, Nash, Smoak & Stewart, P.C

Miami-Dade Medical Examiner Department

>Miami-Dade Police Department *

Miami-Dade Fire Rescue Department *

Mine Safety and Health Administration

*Participated in daily stakeholder meetings. Numerous other Miami-Dade police and fire personnel were involved in the rescue and recovery efforts.