Furnace Tapping 2018, Edited by J.D. Steenkamp & A.Cowey
Furnace Tapping 2018, Edited by J.D. Steenkamp & A.Cowey
Furnace Tapping 2018, Edited by J.D. Steenkamp & A.Cowey
Cowey,
Southern African Institute of Mining and Metallurgy, Kruger National Park, 14–17 October 2018
# &
Anglo American Platinum, South Africa
The Anglo Converting Process (ACP) forms part of the Anglo American Platinum smelting
operations based in South Africa. The ACP converting vessels are two identical circular
furnaces, each with one slag and two matte tap-holes. Each tap-hole has three graphite inserts,
with a 75 mm tapping channel through which the molten material flows. Planned shutdowns
at ACP are driven by maintenance requirements for replacement of the graphite inserts and
sacrificial steel faceplates. Initiatives to extend the life of slag tap-hole graphite inserts started
in 2010 and included measuring insert dimensions and tolerances and ensuring that there
were no gaps during installation. From 2013, after internal benchmarking with other smelting
operations, smaller drill bits were used to open the tap-holes and in 2014 drilling the slag tap-
holes open was implemented. In 2016, the specifications of the graphite inserts were evaluated
and improved further. Combined, these initiatives have extended the graphite inserts’ life,
and the maintenance interval for tap-hole repairs has been extended, thereby delivering
improved availability on the ACP.
The Anglo Platinum Converting Process (ACP) is located in the Waterval Smelter
complex in Rustenburg, North West Province, South Africa. ACP is where all the various
Anglo American Platinum streams converge, making it a crucial part of the platinum
pipeline. The converting furnace is based on the Ausmelt technology and consists of a
top submerged lance configuration (Figure 1). Matte from the four primary electric
smelting furnaces of Anglo American Platinum is injected down the converter lance
together with oxygen-enriched air to produce an Fe- and S-deficient high-grade matte.
The process is continuous and converts the approximately 40% Fe in the furnace matte
down to approximately 3% Fe in the converter matte product.
There are two converters, viz. phase A (commissioned in 2002) and phase B
(commissioned in 2006). Each converter comprises an air-cooled hearth system, a water-
cooled crucible (a low-pressure waffle copper cooler in a circular arrangement), and a
high-pressure boiler system with their individual auxiliaries. Each furnace is operated
for 2–-3 years while the other is on maintenance. Converter matte is tapped and slow-
cooled in the slow-cool aisle before it is crushed and sent to the Rustenburg Base Metal
Refinery for further processing. The slag tapped from the sole slag tap-hole is
granulated, dried, and sent to the slag cleaning furnace (also located in the Waterval
complex) for recovery of the base metals and platinum group metals (PGMs) in the
converter slag (Viviers and Hines 2005). The converting process produces an SO2-rich
off-gas, which is treated at the acid plant to produce sulphuric acid.
49
FURNACE OPERATION
Matte is tapped from the furnace through two tap-holes (Figure 2). The matte tap-hole
elevation is in line with the top of hearth skew so that all the molten matte can be tapped
out, while slag is tapped at the opposite side. The slag tap-hole elevation is 800 mm
higher than the matte tap-holes and slag is tapped out of the furnace using a single slag
tap-hole.
Tap-hole opening is conducted using a drilling machine and oxygen lances. The tapping
channel is drilled to 500 mm, and the remainder is then lanced open. At the slag tap-
hole, the tapping channel is drilled open and instances where lancing is required are
seldom.
Offgas Lance
Matte
Slag
Bath
(1) Planned maintenance shutdowns at ACP are generally scheduled around the
lifespan/integrity of the tap-hole graphite modules, referred to as ’inserts’. In
particular, the slag tap-hole graphite inserts are of concern since these wear faster
than to the inserts installed in the matte tap-holes. There is a distinct difference
in the slag tap-hole channel between the phase A and phase B converters in that
the insert sizes of phase A are smaller than those of phase B. The hole diameters
(75 mm baseline) are identical and the tap-hole channel lengths are
approximately the same. Poor insert integrity poses a number of safety risks,
with the possibility of molten material coming into contact with the water-cooled
copper tap-block and resulting catastrophic failure of the tap-block,
(2) Excessive wear of the insert hole causes high slag flow into the granulation tank,
which affects the water to slag ratio in the granulation and thus results in poor
slag granulation. Typical slag flows range between 3–5 t/min. Slag flow rates
greater than 5 t/min are considered high and pose the risk of ’bangs’ due to low
water: slag ratio.
(3) The use of mud guns to close a running tap-hole introduces an additional risk in
that the mud gun may not have sufficient pressure (typical mud gun pressure is
120–150 bar) to successfully plug the tap-hole with tap-hole clay, as a result of
excessive wear in the hole.
This paper focuses on the work done to improve the lifespan of the graphite inserts in
order to extend campaign life. The campaign life in the context of this paper is the period
between slag insert change.
50
The ACP tap-hole channel is 535 mm deep. The operation is forced to stop and the
converter has to be drained completely (typically 5 hours), blown cold, and drilled safe
before inserts can be changed. By comparison, on the primary six in-line furnaces there
is provision to maintain power on during shallow slag tap-hole repairs, which is not
possible on the ACP converter due to the size of the tapping channel. The slag and matte
inserts are identical. consisting of three graphite blocks. The design of the tap-block is
such that the graphite inserts form a conical shape when installed, viz. shapes BB, CC,
and DD (identification codes) (Figure 3) for phase A and shapes X, Y, and Z (Figure 4)
for phase B slag side.. A very thin layer of Jadeset is applied between each block during
installation to plug any small gaps that may exist between the graphite inserts on
installation. The inserts sit in a copper host block that is water-cooled.
Z Y X
CC BB
DD
Figure 3. Graphite blocks phase A slag. Figure 4. Graphite blocks phase B slag.
The insert dimensions for phase A and phase B are shown in Table I. The phase B inserts
are significantly larger in diameter than the phase A graphite inserts, due to the different
slag copper block design.
The log sheet provides benefit in that when metallurgists are rotated into the tapping
floor section, it makes it easy to capture the measurements on a consistent basis. It is also
used to track insert change history in terms of measurements or any defects. It provides
51
new metallurgists with a learning ground in making decisions and understanding the
importance of the tapping process.
The log sheet entries start with measurements of each new block that is ready to be
installed. It contains the design length and diameter to help guide the actual
measurements. This assists in eliminating supplier errors and proactive decision-making
if the inserts are of incorrect size.
Photographs are taken before and after each block is broken out and hole diameters are
measured. The pictures are useful in aiding decision-making if there is a problem at
startup. Measurements are then taken of each new block installed. Here again the design
measurements are contained to guide the installation process, thus rectifying deviations
immediately.
A report is then compiled by the metallurgist using the information and deviations
explained. The log sheet is then signed off by the metallurgical and section engineers.
The insert life has been a journey of continuous improvement from when the converters
were first commissioned in 2002 (phase A) and 2006 (phase B). The insert lifespan has
significantly increased over the years through various interventions that the technical
metallurgists have implemented. The focus of this paper is on interventions from 2013
to 2017.
The change to a smaller drill bit increased the insert campaign life from 65 taps to about
110 taps.
From a review of data from the insert change log sheet, it was established that the inserts
would form a conical shape at the end of their campaign life due to wear and tear (Figure
5).
The original insert hole diameter was 75 mm. In 2015, 130 taps were achievable with a
13% increase in hole diameter on the first insert (front face) and a 46% increase in hole
diameter on the third insert (back end). The implementation of the finer grained inserts
enhanced the insert life to 180 taps for similar wear patterns.
The hole diameters measured across the three inserts from 2015 (Figure 6) and compared
to the finer grained graphite inserts hole diameters in 2016 (Figure 7) showed similar
results, with greater insert life (130 taps to 180 taps) on the finer grained graphite.
53
The slag flow rate is the first indicator of the extent of wear of the inserts. As previously
mentioned, typical slag flows range between 3 and 5 t/min. The peaks in the slag flows
are indicative of the inserts reaching end of life (Figure 8).
Between 2012 and 2013, phase A was operational, and the smaller drill bits reduced the
slag flows from an average of 3.6 t/min to 3 t/min and increased the slag taps from 65
to 110.
In 2014 phase B became operational and the slag flow rates increased to an average of
3.2 t/min. This is believed to be due to the different design of the copper slag tap-block,
which provides less efficient cooling than the phase A copper slag tap-block. Although
the flows had increased, the slag taps also increased, from 110 taps to 130 taps due to
54
drilling open and eliminating oxygen lancing. In November 2015 the finer grained
inserts reduced the average slag flow to 2.9 t/min and also increased the slag taps to 180.
Due to the fruitful results obtained from the finer grained inserts in the phase B campaign, it was
decided that these be trialled on phase A in 2017. The slag flows on phase A averaged 2.6 t/min,
with 180 slag taps.
The finer grained inserts trial conducted from November 2015 resulted in the campaign
lifespan increase from 130 taps to 180 taps (Figure 9) and extended ACP shutdowns from
4-weekly to 6-weekly. In some instances the inserts were proactively changed due to
other breakdowns that required a shutdown, thus offsetting the trial campaign.
55
The campaign life of the tap-block insert on the slag tapping channel at ACP has been
increased by 176%, thereby extending the time between shutdowns from 2 weeks to 6
weeks. This was achieved through interventions over 9 years. The move from lancing
the slag tap-hole to drilling open, the use of smaller drill bits, and more recently the use
of finer grained graphite brought about this improvement.
This paper is published by permission of Anglo American Platinum. The contributions
of our colleagues in technical and production management are gratefully acknowledged.
Viviers, P. and Hines, K. (2005). The new Anglo Platinum converting project. Proceedings of the
First Extractive Metallurgy Operators’ Conference, Brisbane, 7-8 November 2005.
Australasian Institute of Mining and Metallurgy, Melbourne.,
http://www.academia.edu/14747517/The_New_Anglo_Platinum_Converting_Project
Mbanza Sichone
Metallurgical Production Engineer (MPE), Anglo American Platinum
Mbanza is a chemical engineer who has worked in the
pyrometallurgical industry for over 15 years. She joined Anglo
Platinum in 2008, from Konkola Copper Mines in Zambia, and is
currently the metallurgical production engineer for the converting
process prior to that she worked in the ACP acid plant.
Ishmael Rambiyana
Process Metallurgical Engineer, Anglo American Platinum Ltd.
Ishmael is a process metallurgist at Anglo American Platinum. He
has worked the majority of his career at Anglo American Platinum,
gaining experience at the Converter Plant (ACP), Base Metals
refinery (MCP) and currently provides technical support for the
Acid plant at the Anglo American Platinum Waterval Smelter
Complex. Ishmael holds an M.Eng degree in Pyrometallurgical
Engineering from the University of Pretoria.
56