Pittsburgh Safety & Health Technology Center
P.O. Box 18233
Pittsburgh, PA 15236
| U.S. Department
of Labor |
Mine Safety and
Health Administration Pittsburgh Safety & Health Technology Center P.O. Box 18233 Pittsburgh, PA 15236 |
|
Mine Waste and Geotechnical Engineering Division
December 13, 1999
MEMORANDUM FOR LEE D. RATLIFF
Assistant District Manager, North Central
District
Metal and NonMetal Mine Safety and Health
THROUGH: EDWARD J. MILLER
Chief, Pittsburgh Safety and Health Technology
Center
KELVIN K. WU
Chief, Mine Waste and Geotechnical Engineering
Division
DONALD T. KIRKWOOD
Supervisory Civil Engineer, Mine Waste and
Geotechnical Engineering Division
FROM: TERENCE M. TAYLOR
Civil Engineer, Mine Waste and Geotechnical
Engineering Division
GHARIB IBRAHIM
Civil Engineer, Mine Waste and Geotechnical
Engineering Division
SUBJECT: Executive Summary, Relief Valve Testing for
Explosion Investigation at Kaiser Aluminum and
Chemical Corporation's Gramercy Plant, MSHA
I.D. No. 16-00352
On October 7, 1999, and November 16, 1999, relief valves from Kaiser Aluminum
and Chemical Corporation's (Kaiser) Gramercy Plant were examined and tested at
Valve Actuation and Repair Company's (VARCO) facility in Monessen, Pennsylvania.
VARCO is a licensed VR (Valve Repair) shop in accordance with the National Board
of Boiler and Pressure Vessel Inspectors. Testing was witnessed by the Director
of the Pressure Relief Department of The National Board of Boiler and Pressure
Vessel Inspectors. Representatives from Kaiser, the United Steelworkers of
America, and the Plaintiffs were present for the testing. Four of the seven
relief valves from the manifolds protecting Digestion Area Flash Tanks 7, 8, and
9 were taken as evidence for evaluation. These tanks were three of the five
sealed vessels which were involved in the explosion on July 5, 1999. The four
valves were 8xlO's manufactured by Farris-Teledyne. Two 39 2 were assembled by
Farris-Teledyne and two by Groth Corporation. Each valve was given a unique
identifier. The valves were labeled MSHARV1039, MSHARVU1, MSHARVU2, and
MSHARVU3. The first two were stamped for air service, and the later two for
steam service. The valves had identifying plates with stamped set pressures
ranging from 47 to 49 pounds per square inch.
The testing on October 7, 1999, entailed a pretest inspection, and placement of
the valves on the test media stand. The valves rated for air service were placed
on the air stand, and the valves rated for steam service were placed on the
steam stand. The valves were tested in their as-found condition with the
exception of removing some loose scale in the inlets and outlets. On November
16, 1999, the valves were disassembled and an internal inspection was conducted.
The following is a summary of the test results and conditions observed for each
of the four valves:
MSHARVU1 (serial number 307178-l-A10): this valve was stuck
partially open and leaked excessively when placed on the air test stand. The
valve was pressured to 48.9 p.s.i., and no pop was detected. According to
records received by Groth Corporation, this valve was originally set for a
pressure of 47 p.s.i. The assembler or valve repair (VR) seals where missing.
Erosion marks caused by product flow were present between the nozzle and the
disk. Product was found: on the inlet nozzle seat, on the outside ring of the
nozzle near the disc, adhered to the disc holder, on top of the disc holder,
on top of the guide, and in the body of the valve. These conditions were
evidence that the valve had leaked during prior service. During disassembly,
the product build-up was interfering with the removal of the guide and the
disc holder. It therefore would have adversely affected (increased) the valve
set pressure.
MSHARV1039 (serial number 312350-A1O): This valve was stuck open 3/4"
by large pieces of product between the inlet nozzle and the disc. It was not
tested, as it would have leaked excessively. The valve was originally set for
49 p.s.i., with an air stand. The assembler seals were missing. Product was
found: in the inlet nozzle, in the bonnet, in the body of the valve, on the
inlet nozzle disc seat area, in the outlet of the valve, on top of the disc
holder, and on top of the guide. During disassembly, the product buildup was
interfering with the removal of the guide and the disc holder. It therefore
would have also adversely affected the valve set pressure.
40 3 MSHARVU2 (serial number 3166237-2-A10): This valve was tested on
the steam stand and had an average set pressure of 48.35 p.s.i. This was in
close agreement with the 49 p.s.i. set pressure indicated on the valve
nameplate. Product was found on the valve internals, in the lower section of
the nozzle, and on the valve bonnet. The nozzle ring was stuck in position.
The assembler seals were missing.
MSHARVU3 (serial number 316237-1-A1O): This valve was originally set at
49 p.s.i. for steam service. The bonnet had a large dent and the body was
packed with product. The set pressure was therefore not tested. The assembler
seals were missing from the cap. Product was found: clinging to the body of
the bonnet, on the internals, on the inlet nozzle, and on the tip of the
nozzle and the disc. Product scale on the inlet nozzle caused a 58% reduction
in valve flow capacity.
CONCLUSIONS
The observed conditions of valves MSHARVU1 and MSHARV1039 were evidence that the
valves had leaked while in service, and therefore may have been blocked out
prior to the explosion. The product build up in these valve would have adversely
affected their set pressure.
The set pressure for relief valve MSHARVU2 was consistent with the value
indicated on its nameplate.
The set pressure for valve MSHARVU3 was undetermined. The flow capacity of this
valve would have been reduced by 58% by the large amount of hardened product
scale on the inlet nozzle orifice.
The springs in each of the valves were within specification for the pressures
stamped on the name plates.
Based on the condition on valves MSHARVU1, MSHARVU2, and MSHARV1039, it appeared
that they were not being maintained in good repair.
Attachment
cc:
E. Teaster
M. Skiles
S. Michalek
Files
