UNITED STATES
DEPARTMENT OF LABOR
MINE SAFETY AND HEALTH ADMINISTRATION

District 11

ACCIDENT INVESTIGATION REPORT
Surface Area of Underground Mine


Fall of Material
Structural Failure


No. 5 Mine, I. D. No. 01-01322
Jim Walter Resources, Incorporated
Brookwood, Tuscaloosa County, Alabama


November 10, 1995

By

Walter W. Deason
Coal Mine Inspector

Donald T. Kirkwood
Civil Engineer, Pittsburgh Safety and Health Tech. Center

Terence M. Taylor
Civil Engineer, Pittsburgh Safety and Health Tech. Center


Originating Office - Mine Safety and Health Administration
135 Gemini Circle, Suite 213, Birmingham, Alabama 35209-5842
Michael J. Lawless, District Manager

GENERAL INFORMATION

The No. 5 Mine, Jim Walter Resources, Incorporated, is located in Brookwood, Tuscaloosa County, Alabama. This is a vertical shaft mine approximately 2,100 to 2,500 feet in depth which liberates approximately 20,000,000 cubic feet of methane in a 24-hour period. The mine has three continuous mining machine sections on development work and one longwall section on retreat mining. The mine employs 373 miners and has a daily production of 8,000 to 10,000 tons of raw coal. The preparation plant operates two to three shifts per day, five to six days per week. The preparation plant processes 4,477 tons of coal daily, with a washer reject of 44 percent. The last completed Mine Safety and Health Administration health and safety inspection was conducted from July to September 1995. An ongoing Mine Safety and Health Administration health and safety inspection is being conducted which began in October, 1995.

Company officials are listed below:

William Carr.......................................President
K. J. Matlock......................................Vice President
James M. Smith....................................Vice President
W. H. Weldon.....................................Treasurer
John F. Turbiville...................................Secretary

DESCRIPTION OF THE ACCIDENT

On Friday, November 10, 1995 at approximately 7:00 a.m., the day shift preparation plant crew gathered in the dining area of the plant and C.A. Squire, Plant Foreman, gave out the work assignments.

Thomas Alton Emfinger, Heavy Equipment Operator (victim) and Ronald Holmes, Heavy Equipment Operator, were assigned to haul refuse rock from the refuse bin. Normal haulage operations continued throughout the morning. The cab door on the rock truck that Emfinger was operating would not stay closed, so he took the truck to the shop to be repaired. He obtained an International 330 rock truck and returned to the rock bin area.

Emfinger and Holmes took turns eating lunch. After lunch, they emptied the rock bin. Holmes left the rock bin area to get fuel for his Terex Pan. When Holmes returned to the rock bin, Emfinger was parked out from beneath the rock bin on the west side. Holmes pulled his pan to within 30 feet of the rock bin on the east side. Holmes sat there for approximately ten minutes and at approximately 12:45 p.m., Emfinger backed his truck beneath the rock bin. Holmes stated that when the rock chute door started to open, he saw and heard the cone section of the rock bin begin to fall and land on the cab of Emfinger's rock truck.

There were approximately 150 tons of refuse rock in the bin. The weight broke the front axle on the truck and mashed the cab and frame down toward the ground and the rock covered the cab area where Emfinger was located.

There were no communication devices in the pan so Holmes drove to the washer area for help. Rescue efforts began and the victim was trapped in the cab of the rock truck for approximately two hours and 53 minutes. He was pronounced dead on the scene by the doctor in attendance.

PHYSICAL FACTORS INVOLVED IN THE ACCIDENT

The refuse bin at Jim Walter Resources, Inc., Mine No. 5 was designed and constructed by McNally Pittsburg Manufacturing Corporation. According to the dates on the design drawings No. E-4801-42, E-4801-43, and E-4801-44 the refuse bin was constructed in 1976 and 1977. The refuse bin structure is 70 feet high and consists of four main support columns. They support the 300-ton refuse bin, a hydraulically operated gate mechanism at the bottom of the cone, the end section of the 36 inch wide refuse conveyor belt and supporting structure, and an enclosure housing the conveyor belt head drive, motors, and controls. The compass direction north is on the side of the bin opposite the conveyor structure. The conveyor structure attaches to the bin at a point approximately 58 feet above the column base. For clarity, the four main columns will be referred to as the northeast, northwest, southeast, and southwest columns. The vehicular traffic pulled under the bin from the east toward the west.

McNally Pittsburgh's Drawing No. E-4801-42 identifies the refuse bin as a 300-ton capacity structure with typical refuse material listed as having a unit weight of 100 lbs./cu.ft. Using the 100 lbs./cu. ft. weight and the calculated volume of the refuse bin including the cone and cylinder sections, the estimated weight for a fully loaded bin would have been in excess of 340 tons.

The four columns were designed as W10x49 steel sections for the bottom 42 feet and W10x33 sections for the upper 28 feet. The main horizontal steel members which extend between the four columns, were designated as W12x27 sections 18 feet above the column base (design drawing elevation 151 feet), W21x55 sections 40 feet above the column base (elevation 173 feet 3 inches ), W18x45 sections 58 feet above the column base (elevation 191 feet 7 inches), and W10X21 sections supporting the roof approximately 70 feet above the column base (elevation 203 feet 6 inches).

The refuse conveyor belt dumped the rock and refuse into the bin at drawing elevation 194 feet. The material dropped through an opening in the floor at the elevation 191 feet 7 inches, which was 11 1/2 inches beneath the end of the belt. The refuse bin consists of two main sections. The upper section of the refuse bin was a 12 feet high, 21 feet 10 inches in diameter, steel cylinder constructed out of 1/4 inch thick steel plates. Welded beneath this cylinder was a truncated, steel cone. The cone was 15 feet 6 inches high and had a 4 foot diameter opening in its bottom. The cone was constructed of 1/4 inch plate steel with 1/4 inch hardened steel liners along the inside. The gate for regulating the flow of material from the bin was attached to the bottom of the cone.

At the drawing elevation 173 feet 3 inches, four W21x73 sections were connected across the corners of the main W21x55 beams forming an equilateral octagon with 9 feet 6 3/8 inch long sides. The compression ring beam was composed of two 1/2 inch plates acting as flanges and the section of the 1/4 inch cylinder plate between them, acting as the web. The bottom flange of the compression ring beam rested on the beams forming this octagon. This 1/2 inch compression ring bottom flange was welded perpendicular (horizontal) to the vertical, 1/4 inch cylinder plate approximately 3 inches above the bottom of the cylinder plate. A second 1/2 inch horizontal plate was welded around the circumference of the vertical cylinder plate, 2 feet above the first plate. The compression ring beam also had sixteen, 1/2 inch thick, vertical, bearing stiffener plates welded between its upper and lower flanges. These stiffeners were placed in pairs, one foot apart, centered over the eight locations where the lower flange of the compression ring beam rested fully over the 21-inch deep beam comprising the octagon. The cone section was welded to the inside of the cylinder plate at the same level that the lower flange of the compression ring beam was welded to the outside of the cylinder plate.

The forces exerted by the weight of the waste rock inside of the bin and the weight of the cone and its liners were carried through the cone, to the compression ring beam. The compression beam was supported by the octagon shaped frame of the 21 inch deep beams. The forces in the octagon beams were then transmitted down through the four main columns. The compression ring beam also supported the cylinder section laterally against the forces of the material inside the cone.

The bin was designed so that mobile equipment could enter either from the east or west side between the main columns. The trucks would pull directly under the center of the cone. The gate would be opened remotely and the truck would be filled with refuse and rock. The gate would then be closed remotely, and the truck would exit. No indicators of the material level within the bin were originally installed, nor required. The bin was later equipped with a high level rock indicator.

1991 Modification

The size of the material that was dumped into the refuse bin varied from fine dust to large, 18 inch rocks. The refuse and rocks were dumped from the conveyor belt at a height of 37 feet above the bottom of the cone. The impact of the large rocks on the cone, the abrasion from sliding material, and the corrosive nature of the wet, acidic environment within the bin led to continual deterioration of the inside of the bin. The area that would deteriorate most rapidly was the inside of the cone section and that is why the cone was fitted with hardened steel, armor plates on its inside. Despite the additional protection afforded by these armor or wear plates, the cone section was typically the first area to show signs of serious wear or deterioration.

To address this problem, in 1991 Jim Walter Resources designed a modification to the bin which would allow the cone and its liner plates to be periodically replaced. This modification is illustrated on the design drawings prepared by JWR's designer Ira Lanstrum, numbered 735 A-71355, 735 A-71356, and 735 A-71357. These drawings are identified as applying to the refuse bin at JWR's No. 7 Mine, however, the company stated that they were also used in the fabrication and installation of the refuse bin modification at JWR's No. 5 Mine in 1991. Selman Fabricators assembled the replacement cone sections in 1991 and Professional Iron Workers (PIW) installed the cone in July 1991.

The modification (Dwg. 735 A-71355) called for cutting the original cone off at a point 7 5/16 inches below the attachment point of the cone to the cylinder, described above. A steel flange (top flange) 1/4 inch thick, and 48 gusset plates were welded to the bottom of the section of original cone at the cut line. The drawings called for the top flange to have 136, 13/16 inch diameter bolt holes drilled on 6 inch centers.

In 1991, the cone was fabricated in two sections, an upper cone section and a lower cone section. The upper cone section (Dwg. 735 A-71357) was fabricated in eight tapered and curved segments. Each segment was 9 feet 10 13/16 inches high, 8 1/2 feet wide on its top, and 4 feet wide on its bottom. Each of the eight segments had flanges welded to its top and bottom. The drawings called for both flanges to have 13/16 inch diameter bolt holes drilled through them on 6 inch centers. Another flange was welded to each side of each upper cone segment. The side flanges also had 13/16 inch diameter bolt holes drilled through them, but these holes were on 12 inch centers. The eight, upper cone segments were joined together with bolts through their side flanges. Once the eight segments of the upper cone section were joined, the entire upper cone section was attached to the flange welded to the cut line on the original cone. The lower cone section (Dwg. 735 A-71356) was fabricated as a single section with flanges on its top and bottom. The drawings called for both flanges on the lower cone section also to have 13/16 inch diameter bolt holes drilled on 6 inch centers. The lower cone section was 5 feet high with a 9 feet 9 inch inside diameter on the top and a 4 feet inside diameter on the bottom. The lower cone section was installed as a single piece to the bottom of the upper cone section. The hydraulic gate assembly attached to the bottom of the lower cone section. The specifications on the drawings called for all bolt connections to be made with A-325, 2 inch long, 3/4 inch diameter steel bolts.

1995 Cone Replacement

Jim Walter Resources decided in 1995, as they had in 1991, that because of the wear on the cone section of the rock bin, it was necessary to replace the cone. The drawing (735A-71355, 71356, and 71357) prepared by Ira Lanstrum, which had been used for fabrication and installation of the cone section in 1991, were again used to fabricate and install the new cone section in 1995. The entire upper cone section was replaced in 1995. The top flange, which had in 1991 been welded to the original cone at the cut line was not replaced in 1995. As in 1991, Professional Iron Workers (PIW) was retained to install the newly fabricated cone section. The new cone sections were fabricated in 1995 by Caine Steel. The cone was replaced during the miner's vacation between July 17 and July 22, 1995.

Deterioration of the Compression Ring

The compression ring beam transfers the majority of the load from the weight of the refuse and rock within the bin and the self weight of the cone out to the main structure of the refuse bin. The compression ring beam consists of two 1/2 inch plates acting as flanges and the sections of the 1/4 inch cylinder plate between them, acting as the web. There were several locations observed where the web of the compression ring was completely corroded and/or worn through leaving large holes. In areas where the metal had rusted through, permanent deflections of up to 1/2 inch were measured between the beam's two flanges. Severe deterioration such as this affected the compression ring beamūs ability to transfer loads from the cone to the main structure and more importantly the characteristics of this load transfer would have been altered. Large areas of deterioration of the web, such as those observed, led to non-uniform stiffness within the compression ring beam. This non-uniform stiffness resulted in localized stress concentration in areas surrounding the deterioration. These areas of concentrated stresses would likely have lead to non-uniform stresses within upper connections of the cone including the weld of the top flange to the cone at the cut line and the bolted connections between this flange and the flange on top of the upper cone section.

During the interviews, numerous witnesses stated that it was known among the heavy equipment operators that an operator should not get out of his equipment at the rock bin due to rocks falling out of the cylinder on the north side. There were several incidents reported of minor equipment damage due to falling rocks and one incident of a miner being injured by a falling rock while walking up the stairs on the north side of the bin. During the interviews with management personnel, it was determined that they were aware of the problems and had ordered material to repair the holes in the cylinder. The material to repair the holes in the cylinder wall arrived on mine property approximately two weeks prior to the accident and preparations were being made for the repair work to be conducted.

Inspection of the 1991 Connection of the Top Flange at the Cut Line

The original cone was specified to be cut along a line that was approximately 7 5/16 inches below the attachment point of the cone to the cylinder in 1991. Cut depth measurements taken during the investigation varied from 7 to 9 inches around the perimeter. The top flange was then welded to the original cone at this cut line. Therefore, the top flange was approximately 13 1/2 inches above the bottom of the 21-inch deep beam making up the octagon structure at level elevation 173 feet 3 inches. This configuration left minimal space for welders to work when welding the upper side of the top flange to the cut cone. There were several areas where there would not have been sufficient room for the welders to directly see this area as they were welding it because their heads would not fit between the beam and the flange while welding.

Another problem relative to this weld was that once the original cone was cut, it is unlikely that the bottom of the cone section would have been perfectly round. The cone would have been subjected to many factors which over the preceding 14 years would have left it deflected from its original circular shape at the cut line level. These factors include the continual pounding from rocks falling off the end of the belt and minor deflections caused by the columns being impacted by equipment. Consequently, the top flange, which had been rolled to fit a circular cone with a straight cut line, would not have fit exactly against the cone.

Much of the weld between the top flange and the original cone failed during the cone fall. There were many areas observed that had minimal welds and a few areas that had no signs of a weld between the top flange and the original cone. In many areas of this flange connection, the weld material was uneven and non- uniform. Furthermore, there were several places where a steel bar had been placed between the top flange and the cut off cone. The upper side of this bar had been welded to the original cone and the lower side welded to the top flange. It appears obvious that this was done because the gap between the top flange and the original cone was too large to weld. This gap was further evident in areas where weld material flowed completely behind the gusset plates being welded to the cone because the flange and gusset plate were not tight against the cone. This void between the two metals affected the integrity of the weld.

This weld held the bin from the date the top flange was installed in 1991 until the binūs failure in 1995. However, the weld conditions described above would have created a situation where this connection was stronger in some areas than others. Particularly in the locations of the minimal welds and no welds, stress concentrations would have been created in adjacent areas where the weld was more competent.

Inspection of the 1995 Connection of the Upper Cone Section to 1991 Top Flange

In addition to the alignment problem resulting from attaching the fabricated cone replacement section with its circular top to the out of round existing cone, another misalignment was encountered in 1995. The fabricating drawings prepared for the cone modification made in 1991 and used again in 1995, called for the bolt holes to be placed on 6 inch centers on all horizontal flanges and 12 inch centers on the flanges running up the side of the upper cone section. In 1991, the bolt holes were apparently drilled on 6 1/2 inch centers on all of the horizontal flanges. This discrepancy actually did not affect the installation of the replacement cone in 1991 as the bolt holes in all of the horizontal flanges were drilled on the 6 1/2 inch centers. The bolt holes in both flanges were therefore aligned.

In 1995, the upper and lower cone sections were replaced, but the 1991 top flange welded to the original cone at the cut line, was not. The 1995 replacement cone was fabricated according to the specifications on the 1991 drawings, with bolt holes on 6 inch centers. As a result, when the installation of the replacement cone in 1995 was done, the bolt holes in the top flange and those in the upper flange on the upper cone section did not align as the bolt holes on the former were on 6 1/2 inch centers and on the latter 6 inch centers. In order to connect these two flanges, some existing bolt holes were enlarged and additional bolt holes were made by burning. The result was odd shaped holes, enlarged bolt holes, and unevenly spaced bolt holes. In at least a couple of areas, there was severe misalignment of the upper flanges causing the bolts through the upper cone section to completely miss the top flange welded to the original cone section. The result of these conditions was that fewer bolts were installed than specified in the design.

Supplemental larger size gusset plates were added around the perimeter of the cone at the eight support locations. These plates were welded perpendicular between the cone and the webs of the 21 inch deep beams in the octagon shaped frame. The welded connection between the components appeared to be irregular and in some locations the gusset plate did not meet flush with the cone.

Damaged Main Support Columns

The failure of the cone resulted in its impacting and damaging various structural components on the bin frame as it was falling. The walkway flooring and the channel support members at the level of the hydraulic gate (drawing elevation 151 feet), were completely destroyed by the falling cone. Three of the four main W12x27, horizontal support beams for this floor level were also damaged by the falling cone. The damage done by the falling cone appeared to be limited to those areas at this level and several bracing members below. There was damage to three of the four main W10x49 columns, but most of this damage appeared to have existed prior to the cone failure. The locations and alignment of the permanent deflection in the columns suggested that this damage was done by vehicular impact rather than by impact from the falling cone. Bracing had been attached to the bent columns before the bin failure. The post failure location of this bracing was an indication that most of the damage to these columns existed before the cone fell. Also, testimony by mine personnel clearly indicated that the columns were damaged by trucks running into them and that the southwest column was the most severely damaged prior to the bin failure. The deflection of the south west column was at a location 7 feet above its foundation and was measured to be 9 1/8 inches off vertical toward the west.

The elevations of the cross members supporting the cylinder were surveyed by JWR personnel following the failure. The measurements were taken at the four corners of the W21x55. The south west corner was at least 0.85 inches higher than the other corners. Since a survey was not conducted prior to the fall, comparison could not be made between pre and post failure elevations. Therefore, it was not clear whether this elevation difference was the original as-built level, the result of the truck impact on the columns, or the result of the cone impact on the lower level.

Gate Impact

The possibility of gate impact by the bed of the truck was investigated. The witness who viewed the cone fall stated that the victim's truck bed was not extended and did not hit the gate. The witness also reported that just prior to the accident he had cleaned out accumulated material on the ground with his scraper pan. To further investigate this possibility, a representative from the crushed truck manufacturer inspected the condition of the hydraulic hoses and the lift cylinders. The hoses were not blown and the cylinders were removed for testing. If the bed had been extended at the time of failure, presumably the hose would have blown or the cylinder would have failed. Based on the manufacturers findings, the bed was not extended when the cone landed on the truck. The investigating team did not find any markings on the side of the gate.

CONCLUSION

As previously mentioned, the location the cone cut off line specified for the 1991 retrofit design left minimal clearance for welding a flange plate to the original cone. The design also did not clearly specify what type of welded connection should be made at that interface. In addition to these design factors, the out- of-round shape of the cone and the uneven cut line adversely affected the quality of the weld that could be made between the top flange plate and cone. The alignment problems were further complicated when the flange installed in 1991 with predrilled holes of 6 1/2 inch centers was joined to a cone section in the 1995 with holes drilled on 6 inch centers.

According to witness testimony, the conditions of the compression ring and the top flange connection were not inspected when the decision was made to replace the cone below the flange in 1995. There also was no inspection of the construction during or after the 1995 replacement.

Based on the evidence and the final position of the cone, it appears that the failure initiated on the western side of the structure and progressed toward the east before the cone dropped. The failure occurred along the flange connection as the result of an over-stressing of this area. The primary contributing factors appeared to be missing bolts, oversized bolt holes, poor quality welds between the original cone section and the 1991 upper flange plate, and rusted holes above the compression flange. These conditions caused the load to redistribute to the stiffer areas where the connections were adequate. Ultimately, the load redistribution over-stressed the connections along the western side of the cone causing it to fall. The cone fell onto a truck, resulting in the fatality of the operator who had pulled under the refuse bin.

ENFORCEMENT ACTIONS

1. A 103-K Order Number 3397387 was issued to ensure the safety of miners until the investigation was completed.
   
   
2. A 104-D-2 Order Number 3397782 was issued because the rock bin structure was not being maintained in good repair to prevent accidents and injuries to employees, a violation of 30 CFR, Section 77.200.

Respectfully submitted by:

Walter W. Deason
Coal Mine Safety and Health Inspector

Donald T. Kirkwood
Civil Engineer

Terence M. Taylor
Civil Engineer

Approved by:

Michael Lawless
District Manager

Related Fatal Alert Bulletin:
FAB95C40