Note: Descriptions are shown in the official language in which they were submitted.
CA 02654611 2011-08-09
METHOD OF REMOVING SOLIDS FROM BITUMEN FROTH
FIELD OF INVENTION .
[0002] This invention relates generally to the removal of asphaltenes, other
solids and
water from bitumen froth that is generated in the processing of mined oil
sands in
producing hydrocarbon feed stocks suitable for refining.
BACKGROUND OF THE INVENTION
[0003] In the processing of mined oil sands, bitumen froth is generated
through a
combination of water/aqueous extraction, air flotation, and deaeration
processes.
Typically this deaerated bitumen froth takes the form of a bitumen emulsion
containing
approximately 60 wt% bitumen, 30 wt% water, and 10 wt% mineral solids (e.g.,
sand,
clay).
[0004] The contaminants, such as water (carrying corrosive metal chlorides)
and mineral
solids in the bitumen froth have to be removed before the bitumen can be
further upgraded
or refined. For example, the total water and solids content (bottom solids and
water
analyzed by centrifugation) has to be less than about 0.5 vol. % in order to
meet typical
specifications for being transported by pipeline. Furthermore, the bitumen
must contain
only minimal amounts of impurities (e.g., less than a few hundred parts per
million of ash
product, the metal-containing compounds) if it is to be used as a feedstock
for high-liquid
conversion hydrocracking processes.
[0005] Traditionally, the solids and water have been removed from the bitumen
froth
using a naphthenic process. In the naphthenic process naphtha is added to the
bitumen
froth to reduce its density and viscosity. A series of inclined plate
separators and
centrifuges are then utilized to remove the bulk of the water and solids. This
process
typically results in a diluted bitumen product containing up to 3.5 vol.%
solids and water.
This bitumen product would likely require additional upgrading before it could
be
pipelined or used as a refinery feed stock.
[0006] Recently, a paraffinic process has been proposed as a more effective
means to
remove the solids and water from bitumen froth. In the paraffinic-froth
treatment (PFT)
process, a light aliphatic hydrocarbon solvent (typically pentane, hexane, or
pentane/hexane blends) is blended with the bitumen froth. It has been shown
that if the
1
CA 02654611 2009-02-18
solvent concentration exceeds a certain critical level (the level depending on
the solvent)
certain components of the hydrocarbon (bitumen) known as asphaltenes will
precipitate
out of solution. Further experimental observations have shown that these
asphaltenes,
together with water droplets and solids dispersed in the bitumen, form
aggregates as long
as the onset conditions for asphaltene precipitation are satisfied. The
resulting aggregates
(or flocs) can be readily removed from the bitumen froth in gravity-based
settling
processes without the need for centrifuges or inclined-plate separators. This
gravity-based
settling process can generate a bitumen product containing less than 0.5 vol.
% solids and
water, meeting typical hydrocarbon pipeline specifications.
[00071 The primary criteria to assess product quality after treatment of the
bitumen froth
are water and solids content in the solvent-diluted bitumen. The water content
of batch
samples is often analyzed using Karl Fischer titration and solids content is
often analyzed
as ash content or filterable solids using standard procedures (see for example
ASTM D473
and D4928). These procedures are time and labor-intensive and generally do not
lend
themselves to automatic on-stream implementation or permit timely adjustments
to the
operating conditions. Therefore, direct, real-time monitoring and control of
the settling
process in paraffinic froth treatment is not achieved with existing
instrumentation and
techniques.
[00081 In particular, Canadian Pat. No. 2,350,001 (the `001 reference)
discloses a process
controlled by monitoring the height of an interface (hydrocarbon/water) within
a process
vessel and adjusting flow rates of the vessel's intake and withdrawal streams
based on the
interface height. The `001 reference fails to disclose the use of particle
size distribution.
Further, such an approach is undesirable because it requires long periods of
time (e.g.,
hours) to respond to changes in the feed composition and changes in the level
of the vessel
may relate to parameters other than product quality and particle size.
[00091 U.S. Pat. App. No. 2004/0084623 (the `623 reference) discloses the use
of near-
infrared spectroscopy and chemometrics to determine asphaltene content,
solvent-to-
bitumen ratio, and density of a bitumen-solvent stream. The `623 reference
fails to
disclose the determination or use of particle-size distribution to control a
bitumen stream
composition. Further, the method of the `623 reference is undesirable because
it requires
the development of a database of "calibration models," which would require
significant
time and effort.
2
CA 02654611 2009-02-18
100101 An article by Kurt Leschonski titled "Online Measurement of Particle
Size
Distributions in Gases and Liquids" from 1974 (the Leschonski reference)
discloses a
survey of technologies available for measuring particle-size distribution
(PSD) and
mentions the appropriateness of rapid response. The Leschonski reference does
not
disclose the use of on-line, real-time optical measurements. In fact,
Leschonski teaches
away from such an approach by listing "image analysis" of PSD as not
appropriate for on-
line analysis. See e.g., Table 1, item 4.
[00111 One current practice is to collect liquid samples from the product
stream and
perform standard analytical tests to determine water, mineral solids, or
asphaltene content.
These analyses take hours to complete, so the data are not available to make
real-time
process control decisions. Extensive measures must be employed to ensure the
sample is
representative and stable and care must be taken to address the safety
concerns associated
with the sampling from a pressurized feed stream containing hydrocarbon
fluids.
Furthermore, this type of data only reveals whether or not product
specifications were
being met at the time of sampling. The results do not provide much information
regarding
the gravity-based settling process nor do they indicate how close to the edge
of an
operating envelope one may be operating.
SUMMARY OF THE INVENTION
[00121 The described invention relates to a paraffinic solvent, gravity-based
process for
removing solids content from bitumen froth streams comprising the steps of: a)
providing
particle-sizing instrumentation in a bitumen froth inlet stream to a gravity
settling vessel
subsequent to the addition of the paraffinic solvent, the solvent addition
forming
aggregates; b) measuring a representative particle size distribution of
entrained aggregates
with said instrumentation; c) determining mathematically the settling rate of
the
aggregates from said particle size distribution; d) collecting operating data
while repeating
one or more times steps a) - c) after changing one, or a combination of two or
more, of the
process conditions of the gravity-based process; e) establishing a set of
desired process
operating conditions based upon particle size and settling rate from the data
generated in
step d); and, f) operating said paraffinic solvent, gravity-based process for
aggregate
removal by adjusting process operating conditions to optimize the settling
rate based upon
the set of desired operating conditions established in step d).
[0013] In one or more preferred embodiments the process is preceded wherein an
initial
start-up process precedes the process of claim 1, said start-up process
comprising: 1)
3
CA 02654611 2009-02-18
r r
isolating the particle-sizing instrumentation from the bitumen process stream,
2) then
introducing hot process gas to the particle-sizing instrumentation for
purging, 3)
withdrawing the process gas after purging but without passing into the bitumen
froth
stream, 4) continuing said purging and withdrawing until process temperatures
are
reached, 5) and then reopening the flow of the bitumen froth stream through
the particle-
sizing instrumentation.
[0014] In further embodiments, the exposed portion of the instrumentation in
the inlet
stream is coated with an anti-fouling agent prior to the introduction of the
bitumen froth in
said inlet stream in a) of paragraph 0008 above. Additionally, it is preferred
that in situ
cleanings be conducted in the manner of the start-up options, with hot process
gas or
without, as described further below.
[0015] Thus the invention provides a method/methodology that would enable the
measurement and control of the settling rate via particle size distribution
(PSD)
measurement associated with the gravity settling of asphaltene aggregates and
solids in
bitumen froth-paraffinic solvent process systems on-line. This method provides
a unique
process control tool that enables one skilled in the art to maintain the
stringent product
quality specifications.
BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The present invention and its advantages will be better understood by
referring to
the following detailed description and the attached drawings in which:
Figure 1 is a schematic for a paraffinic froth treatment process embodiment
according to the invention where particle-sizing instrumentation is installed
as a loop off
of the bitumen froth feed stream after the addition of the paraffmic solvent.
Figure 2 is a graphic illustration of the measured increase in median particle
size
with both increasing temperature and increasing solvent-to-bitumen ratio.
Figure 3 is a graphic representation of the calculated increase in hindered
settling
rate of a median particle (as determined from the particle size distribution)
with increasing
temperature and solvent-to-bitumen ratio.
Figure 4 is a graphic illustration of the measured particle size distribution
in terms
of the frequency of occurrence (%) of particles having chord lengths between 0
and 1000
microns when measured through a clean particle sizing instrument window.
4
CA 02654611 2012-03-09
Figure 5 is a graphic illustration of the measured particle size distribution
in terms
of the frequency of occurrence (%) of particles having chord lengths between 0
and 1000
microns when measured through a fouled particle sizing instrument window.
Figure 6 is a graphic representation of the measured particle size
distribution in
terms of the frequency of occurrence (%) of particles having chord lengths
between 0 and
1000 microns when measured, first, before fouling, second, after fouling, and
third, after
in-situ cleaning.
Figure 7 is a graphic illustration of median particle size over time predicted
from a
model based on earlier measurements and correlated to the temperature and
solvent-to-bitumen ratio for actual running (solid line) and the actual median
particle size
as measured in situ over the same time period (large dots).
[0017] The invention will be described in connection with its preferred
embodiments.
The invention is recited in the claims and should not be limited to preferred
embodiments
described herein.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
[0018] The process of the present invention relates to where real-time
measurement of
particle-size distribution in a process stream is used to make process control
decisions in a
gravity-based settling process. The process is particularly suited for the
high-temperature
paraffinic froth treatment process proposed for use in bitumen oil sand
developments.
[0019] Particle-sizing instrumentation ("PSI") is placed in the inlet stream
to the
process, downstream of, or subsequent to, the point where paraffinic solvent
is introduced
to the bitumen froth. The particles (asphaltenic flocs composed of water,
mineral solids,
precipitated asphaltenes, solvent, and sometimes, some bitumen) formed when
the solvent
is introduced to the froth are sized by the instrumentation prior to entering
the settling
vessel. Typical instrumentation may comprise a probe or a sample window, or
equivalent,
in the inlet stream, or more preferably, in an auxiliary loop, see Fig. 1,
where at least a
portion of the inlet stream, preferably all, is diverted for measurement by
passing by the
probe or window, before being returned to the inlet stream.
[0020] The standard paraffinic bitumen froth settling process comprises mixing
a
bitumen froth stream with paraffinic solvent which preferentially dissolves at
least a large
-5-
CA 02654611 2009-02-18
portion of the bitumen in the stream. The then combined stream of bitumen,
water,
solvent and entrained solids is conducted to a settling unit where the solvent
and much of
the bitumen rise to the top and are extracted in an overhead stream while much
of the
water and the entrained solids are withdrawn from the bottom of the unit. The
process
temperatures used in the invention process are typically in excess of 50 C but
less than or
equal to about 100 C, more typically in a range of from 60 C to and including
80 C.
Though temperatures lower than 50 C can be used, the amount of solvent used
must be
increased to achieve the desired amount of asphaltene precipitation. The
operating
pressures of the bitumen froth stream and in the settling unit are typically
from about 100
to 150 psi (689.5 to 1,034 kPa); however, the embodiments described in this
patent are not
limited to these pressures.
[00211 The settling of particles (including aggregates) in a settling unit may
occur under
hindered or unhindered conditions, depending on the concentration and sizes of
particles
in the settling unit. Under free settling conditions, the particles settle
independently and in
unhindered fashion, with the settling rate mainly dependent on particle size.
Under
hindered settling conditions, the particles are sufficiently close to one
another to restrict or
hinder passage or egress of water and solvent from between particles. The
particle settling
rate decreases with increasing concentration and size of solid particles, and
the settling
rate is retarded by particle interference. The settling rate used in the
practice of the
present invention can be either an unhindered settling rate or a hindered
settling rate,
depending on conditions in the settling unit. Since hindered settling
conditions are
applicable to most settling units of paraffinic solvent, gravity-based
processes for
removing solids content from bitumen froth streams, the present disclosure
discusses use
of hindered settling rates. However, persons skilled in the art having benefit
of this
disclosure will recognize that an unhindered settling rate could be used if
the particles in
the settling unit do not experience significant hindered settling conditions.
Persons skilled
in the art would know how to mathematically calculate unhindered settling
rates.
[00221 A preferred method for measuring a representative particle size
distribution and
determining mathematically from that a hindered settling rate is based on
using a chord-
length distribution (CLD) (or a suitable ensemble average of multiple
distributions) of the
particles. When the CLD is recorded by the PSI, it can be converted to a
volume-based
distribution of particle sizes using algebraic relationships well known in the
art. Generally
speaking, an ensemble average, a moving average (preferred), or other
statistic
6
CA 02654611 2009-02-18
measurement based upon the individual CLD's will give better results than a
single
instantaneous statistic. Alternatively, more sophisticated models can be
applied such as
that offered by J. Worlitschek, T. Hocker, M. Mazzotti, "Restoration of PSD
from Chord
Length Distribution Data Using the Method of Projections onto Convex Sets",
Particle &
Particle Systems Characterization, 22(2):81-98, 2005. Additional variables
such as focal
length can also be considered using an approach like that used by J.
Worlitschek and M.
Mazzotti, "Choice of the Focal Point Position Using Lasentec FBRM", Particle &
Particle
Systems Characterization, 20(1):12-17, 2003.
[0023] Alternatively, other algorithms can be applied. See, for example, E. J.
Hukannen,
and R. D. Braatz, "Suspension Polymerization Using In Situ
Laser Backscattering and Video Microscopy", presented at 2003 Lasentec
Users Conference, Feb 26, 2003; and J. Worlitschek, T. Hocker, and M.
Mazzotti,
"Monitoring of Particle-Size Distribution Using Lasentec FBRM", presented at
2002
Lasentec Users Conference, Feb 24, 2002. In the present work, a semi-empirical
model
based on laboratory and pilot-plant experiments with bitumen froth/solvent
systems, sand
particles, and glass beads was used to refine the calculated particle size
distribution
("PSD") based on CLD. See Example 1 below.
[0024] The hindered settling rate of the system of particles can be determined
by
applying suitable correlations available from the literature, see for example,
J. H. Masliyah, "Hindered Settling in Multi-Species Particle System", Chem.
Eng.
Science, 34:1166-1168, 1979; or B. Xue and Y. Sun, "Modeling of Sedimentation
of
Polydisperse Spherical Beads with a Broad Size Distribution", Chem. Eng.
Science,
58:1531-1543, 2003. One preferred approach is to apply the Richardson-Zaki
correlation
to the volume-based PSD. See J. F. Richardson and R. A. Meikle, "Sedimentation
and
Fluidisation Part III: The Sedimentation of Uniform Fine Particles and of Two-
Component
Mixtures of Solids", Trans. Instn. Chem. Engrs. 39: 348-356, 1961; and J. F.
Richardson
and W. N. Zaki, "Sedimentation and Fluidisation: Part 1 ", Trans. Instn. Chem.
Engrs.
32: 5-53, 1954. The capacity of the paraffinic-froth treatment process is
ultimately
determined by the settling rate of the aggregates in the settling vessels. The
settling rate of
each aggregate in the process fluid can be determined based upon its size, but
it is
typically more convenient to base process control on the settling rate of the
median,
average, or other suitable statistical description of aggregate size rather
than on the size of
each individual particle.
7
CA 02654611 2009-02-18
[0025] Parametric studies can be used to correlate the observed aggregate
size, and
therefore settling rate, with a desired set of process operating conditions
such as
solvent/bitumen mass ratio (S:B), flowrate, pressure, and temperature.
Operators may
then adjust one, or a combination of two or more, of the process conditions
(e.g., S:B,
flowrate, pressure, and temperature) to affect the desired change in settling
rate of the
system. From this data, a parametric set of operating conditions can be
established based
upon particle size and settling rate.
[0026] Characteristic particle-size distribution patterns have been identified
which
signify that the optical window of the PSI has become partially or wholly
obscured or
coated with process fluid. Typically under optimal operating conditions the
particle count,
e.g., number of particles recorded by the PSI per second, will be a smooth
curve in an
exemplary range from about 2,000 to 6,000 particles per second, typically
about 3,500,
plus or minus from about 1,500 to 2,500. When the particle count begins to
show spikes,
especially large spikes, in the number of particles being counted, for example
above about
10,000, or especially if above about 15,000 particles per second, up to as
much as 40,000,
or more, then the optical window is likely being obscured by adhering
particles.
[0027] Start-up procedures have been developed which minimize the tendency of
the
process fluid to adhere to and to obscure the instrument's optical window.
Furthermore,
an in-situ cleaning process has been developed to remove this obscuring layer.
A
combination of aromatic solvent (e.g., toluene or mixed aromatic solvents)
washes and
process gas (e.g., nitrogen, methane, helium, argon, natural gas or its
component gases,
etc.) purges of the piping containing the instrument has been shown to be
effective in
cleaning the probe window. See the further description below.
[0028] In a particularly preferred embodiment, an anti-fouling compound may be
used
as a coating to the exposed portion of the instrumentation, or its window,
which reduces
the potential frequency and severity of fouling. Furthermore, a coated window
is easier to
clean using the in situ process than an uncoated window.
[0029] Turning now to Figure 1, particle-sizing instrumentation 12, preferably
a
Lasentec FBRM by Mettler-Toledo that uses a focused-beam reflectance
measurement
technique, in a particle distribution sizing process loop 20, is placed in the
inlet stream 1
before a first settling vessel 2. This particle distribution sizing process
loop may be much
smaller than depicted, and may treat only a representative sampling of the
bitumen process
stream, up to the entire stream in smaller operations. Placement in the inlet
piping helps to
8
CA 02654611 2009-02-18
assure that a representative sample of the bitumen froth has been obtained. In
the first
Froth Settling Unit 2, the heavier, larger aggregates settle and are taken out
in a
concentrated stream 4. This stream is then again diluted with solvent by
solvent stream 5
and enters a second Froth Settling Unit 6 where the heaviest particles and
water are
removed as a tailings stream from the bottom. A thus increased-solvent stream
is taken
off as stream 7 and introduced into the inlet stream 1 at a point well before
it enters the
particle size distribution process loop 20 comprising the particle-sizing
distribution
instrument 12 and particle size distribution process streams 8 and 13. The now
solvent-
containing bitumen froth stream 1 can be diverted, for instance by closing of
a gate valve
V7 (though gate valves are discussed, any effective open/shut-off valve would
be suitable
as well), through an open gate valve V5, and becomes particle-sizing
instrumentation
process stream 8. Means of introducing a purge process gas stream 9 is
provided by a gate
valve V1 and additional means of introducing a solvent stream 10 is provided
by a gate
valve V2. A drain is typically located at a low point in the loop 20, stream
11, which can
be isolated by a gate valve V3. The particle-sizing instrumentation (PSI) 12
then operates
on the provided process stream 8. The instrument measured process stream 13 is
then led
and introduced back into the bitumen stream 1, for example through a gate
valve V6
before entering the first Froth Settling Unit 2 (FSUI). A waste stream 14
taken off from
stream 13 is provided which can be opened or closed with a valve V4, this
being used
typically for purging or to adjust the pressure in the loop.
[0030] As noted above, an anti-fouling coating, such as
dichlorodimethylsilane, may be
applied to the exposed portion of the particle-sizing instrumentation 12, or
window, in the
inlet stream prior to the introduction of the bitumen froth to reduce the
frequency and
severity of fouling. Such windows are typically made of hard mineral
substances that can
transmit long wavelength light, for example, red laser sourced light. The
instrumentation
used in these examples was constructed of sapphire material. The coated window
was
observed to be easier to clean than an uncoated window using the in situ
process. This
coating can be applied by initially cleaning a probe's optical window, for
example, with
toluene (or any aromatic solvent, or mix) and alcohol washes and then
immersing the
probe in a 10 vol% solution of the coating material, e.g.,
dichlorodimethylsilane dissolved
in toluene, for 15 minutes or more. The probe is then removed and allowed to
air dry.
The immersion/air dry cycle may be repeated one or more times if a thicker
coating is
desired. Other suitable anti-fouling coating compounds include known coating
materials
9
CA 02654611 2009-02-18
that are stable and transparent under the operating conditions of the PFT
process, e.g.,
TEFLON , and others.
[0031] After coating, if the preferred method is used, the particle-sizing
instrumentation
is exposed to a stream containing the bitumen froth, upstream of the first
vessel, e.g., Froth
Settling Unit 2 (FSU1), but downstream of the point where the solvent stream 7
is
introduced to the bitumen froth stream 1.
[0032] A preferred start-up procedure preceding the general process steps can
be
followed. Said start-up process can comprise: 1) isolating the PSI 12 from the
bitumen
process stream [opening V7 and closing V5]; 2) then introducing hot process
gas as
stream 9 to the PSI 12 [via VI] for purging; 3) withdrawing the process gas
after purging
as stream 14 [via V4] but without passing into the bitumen froth stream 1; 4)
continuing
said purging and withdrawing until process temperatures are reached; 5) and
then
reopening the flow of the bitumen froth stream through the PSI [opening V5 and
V6 and
closing V7]. The use of hot process gas is a preferred embodiment. Or more
specifically,
referring to Figure 1:
Starting with V7 open and V1, V2, V3, V4, V5 and V6 closed.
Open V4.
Open VI allowing hot process gas to flow through instrument loop 20 and PSI
and
discharge the gas stream to waste.
Gas flow is continued until the PSI has reached process temperature. The
duration of this
step may be specified as a minimum time or based upon temperature
instrumentation on or
in the PSI.
Once process temperature is reached, V4 is closed.
Confirm instrument loop pressure matches or exceeds line pressure.
Close VI.
Open V5.
Open V6.
Close V7
[0033] If hot process gas is not available, the start-up process may comprise
the
following: (1) isolating the PSI 12 from the bitumen froth stream; (2) opening
a waste gas
withdrawal valve located in the PSI loop; (3) heating the PSI by applying heat
to said
instrumentation and at least a portion of connecting fittings and piping until
process
temperatures are reached; (4) increasing pressure in the PSI loop by closing
said
CA 02654611 2009-02-18
withdrawal valve and opening a process valve in said loop; (5) and then
reopening the
flow of the bitumen froth stream through the PSI 12. Or, more specifically
referring to
Figure 1:
Start with V7 open and V1, V2, V3, V4, V5 and V6 closed.
Open V4.
Bring the PSI to process temperature by applying heat by means of steam lance,
heat
tracing, or other suitable heat source(s). The duration of this step may be
specified as a
minimum time or based upon temperature instrumentation on or in the PSI.
Once the PSI has reached process temperature, close V4.
Open V1 allowing process gas to pressurize instrument loop.
Confirm instrument loop pressure matches or exceeds line 1 pressure.
Close V1.
Open V5.
Open V6.
Close V7.
[0034] The described start-up procedures were developed to minimize the
tendency of
the process fluid to adhere to and to obscure the PSI's optical window. Pre-
heating of the
piping and instrument to process temperature by circulating heated gas,
external
application of steam, heat tracing, or other heat source and pressurizing the
instrument
loop with dry gas equilibrates the instrument loop with the system before the
PFT process
fluids are introduced. In this way, when process fluid is introduced, there is
minimal
change in temperature or pressure and solvent vapor will not condense as
droplets on the
probe's optical window.
[0035] A characteristic PSD was identified which signified that the probe
window had
been at least partially obscured by fouling. As noted above, particle counts
increasing
above about 10,000 counts/second, especially above 15,000 counts/second, were
indicative of the fouling. This was repeatedly confirmed when the probe was
removed
from the process piping and a sticky, black film was observed on the probe's
optical
window.
[0036] In situ cleaning of the probe's window can be conducted without
disrupting the
process, when it has been determined that the probe's optical window has
become
obscured, see, for example, Figure 5. This procedure can be completed as often
as needed
to keep the probe's window clean and measurements accurate. Referring again to
11
CA 02654611 2009-02-18
Figure 1, a PSI cleaning process follows, that process comprising: (1)
isolating the PSI
loop 20 from the bitumen froth stream 1; (2) opening a waste gas withdrawal
valve V4
located in the PSI loop 20; (3) introducing a solvent stream 10 to the PSI 12
and (4)
withdrawing the solvent as stream 14 after passing through the PSI 12; (5)
monitoring the
particle size distribution statistics in the PSI; (6) continuing steps (1)
through (5) until the
observed particle count is less than 100 counts/sec; (7) then stopping the
introducing and
the withdrawing of solvent (steps 3 and 4); (8) draining any accumulated
solvent in the
PSI 12; (9) flushing with process gas, and adding heat and any pressure needed
with said
gas to match operating conditions; (10) stopping the introduction and
withdrawal of
process gas; and (11) then re-opening the flow of the bitumen froth process
stream to the
PSI 12. More specifically, the process can be conducted as follows:
Open V7.
Close V6 and V5.
Open V4 to vent the instrumentation loop.
Open V2.
Pump solvent through the PSI 12 and the solvent exiting as waste stream 14.
Monitor
PSD statistics. Continue flowing solvent until observed particle count is less
than 100
counts/sec, preferably less than 10 counts/sec. Typically, satisfactory
results were
obtained in 15 minutes or less.
Once particle count has been reduced, close V2.
Close V4.
Open V3 and allow remaining solvent in instrumentation loop to drain.
Open V1.
Blow out remaining solvent through drain (V3).
Follow start-up procedures described above (paragraphs [0023] or [0029].
[00371 In an alternative embodiment, not shown, the PSI loop 20 can be
installed in the
diluted bitumen product line 3, instead of line 1 as illustrated. Thus valves
V1, V2, V3,
V4, V5, V6, V7 and streams 8, 9, 10, 11, 12, 13, 14, with the PSI 12, would be
placed in
product stream 3 after FSU1 2. The process would be operated in the manner
described in
this patent and would provide advantages of lesser fouling of the PSI and less
need for in
situ cleaning of the PSI probe's window.
EXAMPLES
12
CA 02654611 2009-02-18
[0038] Example 1. This example 1 demonstrates variation in median particle
size and
hindered settling rate with changes in the solvent-to-bitumen ratio
(hereinafter "S:B") or
temperature.
[0039] A parametric study was conducted in a 15 barrel per day (bitumen froth
feed)
pilot plant. Commercial bitumen froth obtained from Syncrude Canada Ltd.
(nominally
60 wt% bitumen, 30 wt% water, 10 wt% mineral solids) was fed to a paraffinic
froth
treatment ("PFT") process using a 60/40 vol% blend of iso/normal pentane as
the
paraffinic solvent.
[0040] PSI 12 (Lasentec FBRM by Mettler-Toledo) was placed in the feed piping
to the
first froth settling unit (FSU1) 2 as indicated in Figure 1. This location is
downstream of
the point where solvent in stream 7 is introduced into the bitumen froth feed
1 but
upstream of the FSUI 2. A dichlorodimethylsilane coating had been applied to
the
probe's window (using the process described in paragraph [0030] prior to
insertion into
the process piping to reduce the propensity for fouling. Initial baseline
readings of the
clean PSI indicated 84 counts/second. Process fluid was introduced into the
PSI piping
using the preferred start-up procedure described above in paragraph [0028]
with hot
nitrogen as the process gas.
[0041] The process was maintained at a number of predetermined set points
(temperature and S:B) for a minimum of 5 minutes each, preferably 20 minutes
or more
during which time chord length distributions (CLD's) were recorded, once every
5
seconds. Data were stored on a computer for later analysis. During the study
the S:B ratio
was varied from 1.4 to 1.8 and the temperature was varied from 60 to 80 C.
[0042] Particle size distributions (PSDs) and statistics which included
median, average,
and particle count were monitored real-time during the parametric study. If
fouling of the
probe window was indicated (as identified using the procedure described in
paragraph
[0022]), the cleaning procedure described above was completed before the study
was
resumed. Typically, the probe operated for 2 or more hours before becoming
fouled and
requiring cleaning.
[0043] The PSDs collected at each process set point (or a subset) were
ensemble
averaged and statistics were computed. Three curves representing the median
particle size
in microns as a function of S:B ratio are shown in Figure 2, representing
temperatures of
60, 70, and 80 C, respectively.
13
CA 02654611 2009-02-18
[00441 This median particle size was used to compute the hindered settling
rate, uh, of
the median particle at each process condition (temperature and S:B) using the
Richardson-
Zaki correlation (see paragraph 0020) given by the equation
Uh =u~(1_q$)",
where,
4gd p - PL
U 3PLCD
Re = PLUfd
u
24 Re < 0.1
Re
CD 124 Re (1 + 0.14 Re '") 0.1 < Re < 1,000 ,
0.445 Re > 1,000
d is the median particle size in meters, and the numerical values are given in
Table 1. These computed hindered settling rates are shown in Figure 3.
Table 1. Parameters used in computation of hindered settling rate
Parameter Symbol Value
Particle density ^p 1060 kg/m3
Liquid density ^L 680 kg/m3
Particle volume fraction 0 0.123
Richardson-Zaki exponent n 5
Liquid viscosity ^ 0.8 centipoise
Gravitational constant g 9.81 m/s2
This curve can be used to establish operating guidelines, specifically the
maximum flux
rate, or throughput, of the FSU 1 (see Fig. 1). For example, from Fig. 3 where
S:B is 1.6
and the temperature is 70 C, we would expect the median particle size to
settle at a rate
within the range from about 1,400 to 1,600 mm/min. In this case, operation of
the pilot
plant was constrained so that the flux rate within the vessel would not exceed
1,400
mm/min to prevent the particles from being carried into the product stream. We
can also
determine that if temperature were increased to 80 C the maximum flux rate
could also be
increased to about 1,800 mm/min, increasing the throughput of the process.
Alternatively,
14
CA 02654611 2009-02-18
this same effect could be obtained without increasing the temperature by
increasing the
S:B from 1.6 to approximately 1.75.
[0045] Example 2: This example demonstrates the characteristic PSD associated
with a
fouled or obscured optical window.
[0046] A parametric study was conducted as in Example 1 including PSI
placement,
probe window coating and start-up procedure.
[0047] When the process fluid was first introduced past a clean PSI, PSDs
similar to
that shown in Figure 4 were observed. In this case 3,736 chords/second were
recorded by
the probe. However, after 140 min of operation, a PSD like that shown in
Figure 5 was
observed. The high particle count, 15,831 chords/s, and noisy distribution are
indicative
of a fouled probe. The probe was removed from the instrumentation loop and
examined.
A thick, sticky, black film was observed, confirming that the probe surface
was obscured.
[0048] Example 3. This example demonstrates the effectiveness of the described
cleaning process.
[0049] In this test, the PSI (Lasentec FBRM by Mettler-Toledo) was placed in
the feed
piping 1 to the first-stage settling vessel (FSU 1) 2 as indicated in Figure
1. As before, this
location is downstream of the point where solvent in stream 7 is introduced
into the
bitumen froth feed but upstream of the FSU1 2. A dichlorodimethylsilane
coating had
been applied to the probe's window as for Example 1 prior to insertion into
the process
piping to reduce the propensity for fouling. Initial baseline readings of the
PSI indicated
less than 100 counts. Process fluid was introduced into the PSI piping using
the preferred
start-up procedure described above for Example 1.
[0050] CLD's similar to that shown in Figure 6 were observed for the first
approximately 90 minutes of operation (curve 20). After 90 minutes, however, a
CLD
characteristic of a fouled probe was observed (curve 21). At 99 minutes, the
following
cleaning procedure was completed. Referring to Figure 1:
Open V7.
Close V6 and V5.
Open V4 to vent the instrumentation loop.
Open V2.
Pump solvent (toluene) through the PSI and into waste.
CA 02654611 2009-02-18
[0051] The CLD statistics were monitored and solvent was pumped until the
observed
particle count fell below 100 chords/sec. In this case it took 5 minutes of
pumping toluene
to adequately clean the window.
Close V2.
Close V4.
Open V3 and allow remaining solvent in instrumentation loop to drain.
Open V 1.
Blow out remaining solvent through drain (V3).
[0052] The start-up procedures of paragraph [0028] were followed again.
[0053] Once the cleaning procedure was completed the observed CLD was
substantially
restored to its original distribution (curve 22) as shown in Figure 6. This
demonstrated
that the cleaning procedure was effective in restoring the probe to its normal
operation
without the need to physically remove it from the instrument piping or disrupt
the feed to
the process.
[0054] Example 4. This example demonstrates the potential for process control
using
the PSI and modeling based on its measurements.
[0055] This test was again conducted as in Example 1.
[0056] Data (CLDs) were collected with the PSI over a continuous 2-day period.
The
PSD was then computed using the procedures described previously at 4 hour
intervals.
The median of these PSDs distribution was determined and is shown in Figure 7
(large
dots).
[0057] The data collected in Figure 3 were regressed to create an empirical
model of the
form:
L = aT +bT2 +cS+dS2 +eTS+ f ,
where L is the median particle size in microns, T is the temperature in C, S
is the
solvent-to-bitumen ratio (dimensionless), and a, b, c, d, e, and f are
adjustable
parameters determined through least-squares regression. (Note alternate
expressions could be used for the regression incorporating more or fewer
independent variables.)
[0058] The temperature and solvent-to-bitumen ratio in the pilot period were
also
recorded during this same time period. These values (instantaneous values) of
temperature
and solvent-to-bitumen were used to predict the median particle size using the
expression
above. The resulting values are shown in Figure 7 (black line). The reader can
see that
16
CA 02654611 2012-03-09
with minimal exception, the correlation obtained during the parametric study
of Example 1
predicts the variation in median particle size as recorded by the PSI. In
cases where
significant deviation exists, the observed discrepancy is due to local (in
time) variability in
the composition of the bitumen froth fed to the system and does not represent
an inherent
weakness in the method.
[0059] This expression (or one similar) could be implemented into a computer
or other
digital control system. If the current particle size in the process as
measured by the PSI is
undesirable, the appropriate temperature and/or solvent-to-bitumen ratio could
be
computed from the expression and the process set points could be changed. The
PSI
would then provide the necessary feedback to confirm that the process change
had the
desired influence on particle size. Such an approach does not require accurate
knowledge
of the composition of the bitumen froth since the real-time feedback and
control allows the
operator to maintain a desired particle size, and therefore settling rate, by
changing
temperature and/or solvent-to-bitumen ratio. This demonstrates that the PSI
and the
procedures described above could be implemented in a control system.
[0060] The invention is recited in the claims and should not be limited to
preferred
embodiments described herein.
-17-
