CA1232854A - Use of a submersible viscometer in the primary separation step of the hot water process for recovery of bitumen from tar sand - Google Patents

Abstract

"USE OF A SUBMERSIBLE VISCOMETER IN
THE PRIMARY SEPARATION STEP OF
THE HOT WATER PROCESS FOR
RECOVERY OF BITUMEN FROM TAR SAND"

ABSTRACT OF THE DISCLOSURE
The hot water process is controlled in response to viscosity measurements taken in situ in the middlings in the primary separation vessel. The viscosity in the middlings is found to vary.
Therefore, the layer of maximum viscosity is located and the viscosity at this depth is monitored. Adjustments are made to the process to keep this maximum viscosity below a pre-determined limit.

Description

~23Z854

2 This invention relates to an improvement of the flotation-

3 sedimentation step, for recovering bitumen from a tar sand slurry in

4 a primary separation vessel, which step forms part of a conventional tar sand plant circuit. More particularly, it relates to the manner 6 in which the viscosity of the middlings is measured and to the 7 utilization of the measurements so obtained to guide adjustments to 8 the process conditions.

Tar sands, also referred to as oil sands and bituminous Al sands, contain a heavy oil usually referred to as bitumen.
12 There are tar sand deposits, in the Athabasca region of 3 Alberta, which are today being commercially exploited. In connection 4 with these operations, the tar sand is first mined and the bitumen is then extracted from the mined tar sand by a process called the hot water 16 process. The extracted bitumen is subsequently upgraded by refinery-17 type processing, to produce synthetic crude.
18 The tar sand is a mixture of sand grains, connate water, 19 fine minerals of the particle size of clay, and bitumen. It is commonly believed that the connate water envelopes the grains of sand, the fine 21 solids are distributed in the water sheaths, and the bitumen is trapped 22 in the interstitial spaces between the water-sheathed grains.
23 The hot water process is now well described in the patent 24 and technical literature. A schematic of the circuit is shown in Figure 3.

1 In broad summary, this process comprises first conditipnjng 2 the tar sand, to make it amenable to flotation-sedimentation separation 3 of the bitumen from the solids. Conditioning involves feeding mined tar 4 sand, hot water (180F), an alkaline process aid (usually Noah), and steam into a rotating horizontal drum, wherein the ingredients are agitated 6 together. Typically, the amounts of reagents added are in the following 7 proportions:
8 tar sand - 3250 tons 9 hot water - 610 tons Noah - 4 tons (20% Noah) 11 Enough steam is added to ensure an exit temperature of the mixture 12 from the drum of about 180F. The residence time in the drum is 13 typically about 4 minutes.
14 During conditioning, the mined tar sand (in which the bitumen, connate water and solids are tightly bound together) is 16 converted into an aqueous slurry of porridge-like consistency, wherein 7 the components are in loose association.
18 The slurry leaving the drum is screened, to remove oversize 19 material, and then flooded or diluted with additional hot water. The diluted slurry typically comprises 7% by weight bitumen, 43% water, and 21 50% solids. Its temperature is typically 160 - 180F.
22 The diluted slurry then is transferred to the primary 23 separation step, wherein it is temporarily retained in a large separation24 vessel having a cylindrical upper section and conical lower section. (This vessel is hereafter referred to as the "PSV" - for 'primary separation 26 vessel'.) The vessel is similar to a thickener and has a rake system 27 in its lower end to assist in discharging the sand bed which accumulates 28 there. The slurry is retained in the PSV for about 45 minutes in a 29 quiescent condition.

~2328S~

1 During this interval, air bubbles, incorporated into the 2 dilute slurry during conditioning, attach themselves to the bitumen, which 3 is in the form of flecks or globules. Most of the aerated globules are 4 buoyant and they rise through the slurry, to collect at the upper surface in the form of a froth. This froth us referred to as primary froth.
6 Most of the coarse solids, primarily being sand particles, 7 sink through the slurry, are concentrated in the conical bottom end of 8 the vessel, and are discharged through a bottom outlet. This stream 9 is discarded as tailings (known as the 'Primary tailings').
Not all of the bitumen becomes sufficiently aerated so 11 as to rise and join the primary froth. Some of this non-buoyant 12 bitumen is lost with the primary tailings. Most of it, together with a 13 large part of the fines, collects in the mid-section of the PSV. This 14 aqueous mixture is termed "middlings".
A drag stream of the middlings is withdrawn from the 16 vessel and is fed into sub aerated flotation cells, wherein secondary 17 separation is practiced. Here the middlings are subjected to 18 vigorous agitation and aeration. Bitumen froth, termed "secondary 19 froth", is produced.
Typically, the primary and secondary froths have the 21 following compositions:
22 Primary (% by weight) Secondary (% by weight) 23 Bitumen 66.4 23.8 24 Solids 8.9 17.5 Water 24.7 58.7 123Z8S~

1 It will be noted that the secondary froth is considerably 2 more contaminated with water and solids than the primary froth. One 3 seeks to minimize this contaminations the froth stream requires 4 downstream treatment, to remove solids and water, before it can be fed to the upgrading process.
6 It is therefore desirable to operate the process so that 7 as much of the bitumen as possible reports to the primary froth.
8 In summary then, the contents of the PSV may be described 9 as existing in the form of three sequential layers. it the base one has the tailings - this is primarily sand with some water and a minor Al amount of bitumen entrained therein. Above this layer, one has the 12 middlings - this is water containing fines and insufficiently buoyant 3 bitumen . But passing downwardly through the middlings are many coarse 4 sand particles and rising through the layer are some buoyant bitumen globules. And at the top , one has the froth.
16 Of particular interest are the well-aerated bitumen 7 globules, which should rise and form the primary froth, which is the 18 main commercial product of the process. These globules must make 19 their way up through the middlings.
If the middlings are too viscous the well-aerated bitumen 21 globules may fail to achieve the needed upward velocity and may end 22 up being discharged with the primary tailings or being withdrawn 23 with middlings for treatment in the secondary separation circuit. If 24 the globules exit with the primary tailings, they are lost entirely from the process. If they are removed to secondary recovery, they 26 will be recovered in the form of poor quality froth.

1232~354 1 At this point, it us appropriate to point out: (1) that 2 the nature of the tar sand feed is variable; and (2) that the capability 3 of the hot water process to extract the contained bitumen is significantly 4 affected by the nature of the tar sand feed.
More particularly, the tar sand may contain a relatively 6 high content of bitumen and a relatively low content of fines. This type 7 of feed is referred to as "rich" tar sand. Alternatively, the tar sand 8 may be relatively low in bitumen and high in fines. Such a feed is 9 referred to as "lean" tar sand.
Typically, a "rich" tar sand can have a composition as 11 follows 12 14.44% bitumen 13 0.36% water 14 85.2% total solids Typically, a "lean" tar sand can have a composition as 16 follows:
17 7.56% bitumen 18 0.5% water 19 91.84% total solids.
The percentage fine solids (-44~ solids in the total 21 solids) can range from 5% for rich tar sands to as high as 25% for some 22 lean tar sands.
23 In general, the rich tar sand feeds yield high primary 24 froth recoveries. The lean feeds give low primary froth recoveries.
Stated otherwise the lean feeds are difficult to process with the hot 26 water extraction procedure; they do not contain much bitumen and such 27 bitumen as they do contain is difficult to extract.

isles 1 This is partly because the lean feeds contain many fines, 2 which interfere with the flotation-sedimentation separation taking place 3 in the middlings layer of the PSV. In~additionj the flexor globules 4 of bitumen which appear in the PSV middlings, when lean tar sand is the feed, are minute compared to the globules that are there when the 6 tar sand feed is rich. These minute flecks do not rise as readily as 7 the larger flecks.
8 If the fines content in the middlings becomes high. the 9 flotation mechanism can literally become inoperative. There is so little primary froth being produced that the process performance is unacceptable.
Al In this instance, the contents of the PSV may have to be jettisoned and 12 the process started up again.
13 There are a number of courses of action open to the 14 operator by which he may adjust and alleviate undesirable process conditions in the PSV arising from the nature of the tar sand feed. For example, he 16 can:
17 - adjust the rate of Noah addition; or 18 - adjust the rate of water addition to the conditioning 19 or flooding steps or - blend some better quality tar sand in with the lean 21 tar sand, to provide a blended feed; or 22 - vary the residence time or temperature in the 23 conditioning drum.
24 A crucial matter, though, is to know when to make these adjustments and to what extent the adjustment should be made. This 26 requires that a process parameter be monitored which parameter gives 27 the operator a useful guide on which to base the adjustments.

1~328S~

It has heretofore been broadly taught in the prior art 2 that the viscosity of the middlings can be monitored and maintained 3 within staged ranges, to optimize the primary bitumen froth recovery 4 from the PSV. This teaching appears in Canadian patent 889,823~ filed by Grubbily et at. Also of interest are Canadian patents, 889,825 6 and 841,581 7 However, in accordance with conventional practice, the 8 viscosity has been monitored in one of the following ways:
9 - withdrawing a sample from the middlings drag stream and measuring the sample viscosity with an appropriate Al instrument; or 12 - lowering a sampler into the middlings, taking a grab 3 sample, and measuring the sample viscosity with an 4 appropriate instrument; or - applying density measurements to either of the 16 foregoing samples and assuming that the viscosity 17 varies proportionately with the density.
18 Now, there are certain shortcomings associated with 19 these prior art practices.
If one samples the middlings drag stream, one must assume 21 that this sample - taken at one level of the PSV (there is usually 22 only a single outlet in the PSV wall) - is representative of the entire 23 column of PSV middlings.
24 When one attempts to measure the viscosity of this sample, 25 one is dealing with a mixture of sand, oil, clay, and water. The sand 26 and ail begin to settle and rise instantaneously. In addition, the 27 mixture is not static. It is impossible to duplicate the flow and 28 turbulence conditions which exist within the PSV.

~232854 1 Perhaps for these-reasons, the industry has moved toward 2 measuring the density of the sample and assuming that the trend of 3 viscosity will follow the trend of density.

In the fundamental step of this invention, the viscosity 6 of the middlings is taken in situ in the PSV with submersible viscometer.
7 In the testing which led up to this invention, when this 8 was done the following discoveries wormwood:
9 (1) that the viscosity varies strikingly at various depths in the middlings in the PSV;
11 (2) that while the in situ-measured viscosity in the 12 PSV may vary significantly, the density of the 3 middlings when measured in connection with grab 14 samples may vary very little - therefore there does not appear to be a useful correlation between the 16 two that may be relied on; and 17 (3) that the viscosity measurements obtained in situ 18 vary significantly from those obtained by taking 19 grab samples at the same depth in the PSV and measuring the viscosity of the grab samples in a 21 conventional instrument external of the PSV.

~23Z8S~

1 Stated otherwise, it has been found that it is necessary to 2 measure the viscosity of the middlings in the dynamic environment of the 3 PSV contents, in order to obtain reliable and useful measurements. It is postulated that the currents which arise in the PSV (from the continuous entry of fresh slurry, the withdrawals of the tailings and middlings streams, 6 and the influences of dropping solids and rising bitumen), together with the 7 presence of the solids at the point of testing, combine to create a unique 8 and depth-variable viscosity regime in situ which differs in kind from that9 which may be measured in grab samples and drag streams. It is this unique in situ viscosity regime which must be monitored in order to give the desired 11 guidance for process control.
12 In a preferred embodiment, one may "hunt" out the maximum 13 viscosity level in the middlings in the PSV by moving the submersible 14 viscometer vertically and taking measurements at different levels. One then alleviates the undesirable process conditions by monitoring 16 viscosity at this level and making one or more process adjustments, 17 as previously described, to control said maximum viscosity and bring it 18 close to a predetermined desired value.
19 Broadly stated, the invention is an improvement in the primary separation step of the hot water process for extracting bitumen from tar sand 21 in a primary separation vessel, wherein the bitumen floats upwardly in a 22 tar sand slurry to form a froth layer, the coarse solids drop to form a 23 tailings layer, and a middlings layer is formed between the froth and the 24 tailings. The improvement comprises: providing a submerged viscometer in the middlings layer and actuating said viscometer to measure the viscosity 26 of the middlings at one or more levels in the vertical column of middlings 27 and produce signals, external of the vessel which are indicative of said 28 measurements; taking sufficient measurements to determine the viscosity of 29 the region of maximum viscosity within the middlings layer; and adjusting the viscosity of the middlings in response to said signals to maintain the 31 maximum viscosity in the column below a predetermined value, whereby the i2~2~354 1 flotation of the bitumen through the middlings layer to the froth layer is 2 substantially enhanced.

- lo -1~32~3S4 2 Figure 1 is a partially sectional side view of the 3 viscometer used in connection with the invention;
4 Figure 2 is a sectional: side view showing the viscometer suspended in the PSV of the pilot hot water process circuit used in 6 developing the invention;
7 Figure 3 is a schematic showing the hot water process 8 circuit, 9 Figure 4 is a plot of measured in-situ viscosity versus depth in the PSV at which the viscosity was measured, showing the Al variation in viscosity which is present in the PSV middlings at different 12 levels, for a single tar sand feed treated in two ways - one without 3 Noah addition and the other with Noah;
4 Figure 5 is a plot of measured density values for grab samples taken at different depths for the tar sand runs which generated 16 the data for Figure 4; and 17 Figure pa is a fanciful representation of the PSV contents 18 during the run in which Noah was not used;
19 Figure 6b is a fanciful representation of the PSV contents during the run in which Noah was used.

lZ3Z~354 2 The viscometer l used was of the oscillating torsional 3 pendulum type. The particular viscometer used was obtained from 4 Nametre Co., Edison, New Jersey and was identified as Model 7-006.
This particular viscometer Hess sphere 2 Which vibrates at a certain 6 frequency in air. When the viscometer is immersed in a viscous medium.
7 there is a change or diminution in vibration amplitude, Which is 8 related to the drag on the sphere. The additional power required, to 9 maintain the amplitude with the sphere immersed, at its value in air, is a measure of the viscosity of the medium.
Al The mode of operation of this instrument is explained in 12 an article entitled "New technique accurately measures low viscosity 3 on-line" in Control Engineering, July, 1975, pp. 39 - 40, which article 4 is incorporated herein by reference.
The viscometer l was enclosed in a waterproof housing 3.
16 Protective threaded bars 4, adjustable in length, were screwed into the 7 housing 3 and protruded downwardly beside the sphere 2, to protect it 18 against contact with the wall 5 and rake 6 of the PSV 7. A tube 8 was lo attached to the housing 3, whereby the unit could be raised and lowered - conductive leads 12 extended through the tube 8 to the disco-21 meter. The viscometer was adapted to produce a signal, indicative of 22 the change in vibration amplitude exerted by the PSV fluid, which signal 23 was a measure of the viscosity of the fluid in which the sphere 2 24 was vibrating.
The viscometer l is shown in Figure 2 as it was used in 26 the PSV 7. This PSV was a small, noncommercial pilot unit. However, 27 processing results in this pilot unit had previously been shown to 28 correlate with processing results in applicant's full scale commercial 29 PSVs.
The pilot PSV 7 was glass-sided, so that the action within 31 could be observed.

~232854 1 The PSV 7 was part of a circuit illustrated in Figure 3.
2 This circuit comprised a tumbler' 9j in which-tar sand was mixed With 3 hot water, Noah, and steam, and conditioned. The product slurry from 4 the tumbler 9 was diluted with additional hot water in a pump box lo.
The diluted slurry from the pump boy lo was transferred into the pSV 7 6 and retained there under quiescent conditions, to produce bitumen froth, 7 tailings, and middlings. Middlings were withdrawn from the PI 7 and 8 treated in a bank of sub-aerated flotation cells 11, to produce secondaryg froth and secondary tailings. The foregoing steps were conducted in accordance with conventional hot water process conditions.

11 Example l 12 The pilot circuit was used to process a tar sand 13 designated "A". This was known to be a poorly processing, lean feed.
14 Two runs were made during which the feed was treated by the hot water process. One run was carried out with Noah process aid having been 16 incorporated in the slurry; the other run was carried out without Noah.
17 Viscosity measurements were made during each run using the viscometer 1 18 at different depths in the middlings in the PSV 7. Two curves or plots 19 of measured in situ viscosity versus depth were developed. Plot 1 in Figure 4 involved the run without Noah. Plot 2 in Figure 4 involved 21 the run with Noah. The details of the conditions and primary froth 22 recovery results of the two runs are now set forth.
23 Tar Sand "A" composition:
24 9. I bitumen 3.2~ water 26 87.0% solids 27 21.3% fine solids' (expressed as % of -44~ solids 28 in the total solids) ~32854 Pilot Processing of Oil Sands "A"
2 Oil Sand Feed Ray lo - 630 g/s 3 Slurry-Temperature - 80C
4 Rate of Total Water Addition - 418 g/s Naomi Primary 6 Addition Bitumen recovery 7 (wt. %) (%) 8 0.000 g-5 9 0.025 22.1 As shown by plot 1 for the run without Noah, at a depth 11 of about 0.4 m in the PSV, the viscosity measured with the viscometer 12 was about I maps As the viscometer was lowered, the viscosity 3 increased rapidly to 110 maps at a depth of 0.8 my and then 4 diminished to about 80 maps at a final depth of about 1.2 m.
Thus the PSV contents, when the PSV was operating on 16 this lean tar sand A, were shown to be characterized by:
17 - a low viscosity at the upper end of the body of 18 contents (as very little primary bitumen froth was 19 generated by the poorly processing slurry in the absence of Noah);
21 - changes in viscosity with depth;
22 - and a "plug" or layer of high viscosity middlings 23 intermediate its ends.
24 The PSV contents were visually observed through the glass 25 wall of the vessel. Figure pa depicts what was observed. Again, there 26 was only a thin layer of primary bitumen froth at the top end of the 27 vessel contents and a viscous intermediate layer, which contained much 28 bitumen.

~Z328S4 1 The same tar sand A was then treated under the same 2 conditions as the Plot 1 run, except that in this second run a 3 conventional amount of Noah was used. The in-situ viscosity versus 4 depth results are shown by Plot II in Figure 4. At the top of the cell contents the viscometer 1 indicated aye viscosity (130 maps 6 indicative of the thick bitumen froth layer which was produced. As the 7 viscometer way lowered to 0.3 m, it passed through the froth-middlings 8 interface and the measured viscosity dropped off sharply. The viscometer 9 1 indicated that the viscosity continued to decline to a limiting value around 10 maps in the lower part of the vessel There was no "plug"
Al of highly viscous middlings to hinder the rise of the bitumen globules.
12 An improved primary bitumen froth recovery was obtained in this run 3 as compared with the first run. Visual inspection during the run 4 indicated that the PSV contents were of the form shown in Figure 6b.
There was a thick froth layer and no noticeable viscous layer laden 16 with bitumen.
17 Thus there was correlation between the results indicated 18 by the in situ viscometer measurements and PSV performance as indicated 19 by the primary oil recoveries.
During the two runs, several grab samples were also taken 21 at depths corresponding with some of those at which the viscometer 1 22 took in situ measurements. Attempts to measure viscosity representative 23 of conditions within the PSV, on withdrawn samples, resulted in failure.
24 The above-noted problems, that is, the ascent of bitumen in the sample jars, the rapid settling of coarse solids, and the impractical require-26 mints for reproducing the flow and turbulence currents of the PSV, 27 caused such measurements to be abandoned.

-- US --~Z32854 1 In summary, these results show that:
2 (1) use of the:submersible:viscometer produces results 3 that indicate thither are viscosity changes that 4 occur within a PSV.With depth;
(2) If high viscosity layers are developed in the PSV
6 middlings, they do trap bitumen and diminish primary 7 bitumen froth production, and 8 (3) These high viscosity layers can be eliminated by 9 adjusting process conditions, thereby improving primary bitumen froth recovery.
11 In use, the signals emitted by the viscometer 1, submerged 12 in the middlings, are monitored and the viscosity of the middlings 3 are adjusted by altering one of the aforesaid process conditions, to 4 maintain the maximum viscosity in the middlings column below a pro-determined value.

Claims (2)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY
OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. In the primary separation step of the hot water process for extracting bitumen from tar sand in a primary separation vessel, wherein the bitumen floats upwardly in a tar sand slurry to form a froth layer, the coarse solids drop to form a tailings layer, and a middlings layer is formed between the froth and the tailings, the improvement comprising:
providing a submerged viscometer in the middlings layer and actuating said viscometer to measure the viscosity of the middlings at one or more levels in the vertical column of middlings and produce signals, external of the vessel, which are indicative of said measurements;
taking sufficient measurements to determine the viscosity of the region of maximum viscosity within the middlings layer;
and adjusting the viscosity of the middlings in response to said signals to maintain the maximum viscosity in the column below a predetermined value, whereby the flotation of the bitumen through the middlings layer to the froth layer is substantially enhanced.

2. The improvement as set forth in claim 1 comprising:
moving the viscometer vertically within the column of middlings and locating and measuring the viscosity of the layer of middlings which has the maximum viscosity.