Slurry Pipeline Design, Testing and Practice

1
Slurry Pipeline Design, Testing and Practice

• Presented by
• Graeme Addie and Lee Whitlock
• GIW Industries, Inc.
• Bob Hagler
• Hagler Systems

2
Slurry Pipeline Design

• FIRST NEED TO KNOW
• CONCENTRATION FIXED OR CAN BE VARIED
• IS PIPELINE EXISTING DIAMETER CAN IT BE VARIED
• IS SLURRY SETTLING OR NON-SETTLNIG TYPE
• HOW MUCH CAN
• SLURRY SIZE
• TYPE
• CONCENTRATION, ETC.
• VARY/ CHANGE

3
Solids Transport Ratefor Different Pipe
Diameters, Velocities and Concentrations
4
Demarcation Between Settling and Non-Settling
5
Slurry System Characteristics
6
When the solids are less than 80 micron, the
slurry is usually non-settling
7
For Non-Settling Slurry
Design flow can theoretically be very low but
energy dictates need be higher capital cost
also Also, if laminar, any large particles may
settle
8
Non-Settling Slurry Analysis
Rabinowitsch and Mooney have shown that the rate
of strain of a fluid particle adjacent to the
pipe wall can be expressed as
(11)
for all laminar flows in a pipe. In eq. 11
(12)
The importance of the Rabinowitsch and Mooney
proof is that of establishing the scaling law
(13)
for all steady uniform laminar flows in a pipe.
In other words, both T 0 and 8V/D can be
determined from experiment with the result that
the experimentally determined friction, eq. 13,
can be applied to other pipe sizes and/or
velocities provided that the same non-Newtonian
fluid is involved in both the test are the
design. Here, the shear stress distribution is
linear varying from a maximum at the boundary to
zero at the centerline of the pipe, that is,
(3)
In which dp/dx the pressure gradient along the
axis of a horizontal pipe.
9
Where We Can Only Get a Small Sample, We Can Use

• Rotating Viscometer
• Extrusion Rheometer

10
Extrusion Rheometer
11
Non-settling slurries have a zero flow wall shear
stress and a laminar and turbulent region.
t
(PSI)
12
Non-settling slurries arevery concentration
dependent
1.45 Red Mud Alcan 1.13 Phosphate Slimes
13
Settling slurries of 100 micron to 5 mm size
solids,have a minimum head loss and a deposit
velocity
Settling Slurries
14
The deposit velocity can be calculated
15
Pipeline Deposit Velocity
16
Settling Slurry Modeling Approach

• Wilson and Clift used the concept of a
  stratification ratio R and the threshold of
  turbulent suspension to show that the total
  excess friction pressure gradient (im - if) is
• where the stratification ratio
• and
• A' and B are properties of the slurry
• Smd specific gravity of the delivered mixture
• if pipe friction due to carrier liquid only.

17
Stratification in Settling Slurries
To estimate the variation of R with pipe size and
mixture velocity, we return to the work of Wilson
(5). Wilson and Watt proposed that, below a
particular mixture velocity termed the threshold
of turbulent suspension, Vs, the solids are fully
stratified. If the mixture velocity is increased
above Vs, then part of the solids pass into
pseudohomogeneous suspension. The value of Vs is
estimated as where Vt and d are the terminal
velocity and diameter of a conveyed particle, and
?w is the (Moody) friction factor for flow of the
carrier liquid alone (i.e. 2gDiw/V2m). For
mixture velocities below Vs, flow is fully
stratified and R 1. For Vm gt Vs, R can be
estimated where values of m close to 1.7 have
been determined (6). For particles which are
sufficiently fine to ensure Vs lt Vm over the
whole.
(9)
(10)
18
Settling Slurry Transport Mechanisms
Sliding
Lifted
19
Analysis of Settling Slurry Test Data
Behaviour of masonry-sand slurry (d50 0.42
mm) In 203 mm and 440 mm pipe, after Clift et al.
(1982)
20
Settling Slurry Pipe Friction Approach
21
Pipe Friction Constant From Pipe Tests
22
Wilson Book Method
The equation for the approximating power law is
readily obtained in terms of V50, giving the
relative solids effect as
The coefficient 0.22 represents the value of the
relative solids effect at the point where Vm
equals V50, as mentioned previously. The power M
is about 1.7 for slurries with a narrow particle
grading (Clift et al., 1982)
Based on data obtained for various slurries
tested at the GIW Hydraulic Laboratory, the
resulting expression is
Prepared by CRA GIW Industries, Inc.
22
23
GIW has Developed Different Models for Predicting
Head Loss
Definition sketch for limit of stationary deposit
zone.
24
The GIW Models are Some of the Best
Available.In a Long Pipeline with Difficult
Slurry, it is Best to do Tests.
25
GIW Slurry Test Loops3, 4, 6, 8, 12, 18
and 20
26
Observing Deposit Velocity at Glass Pipe Section
27
US Army Corps. of EngineersSlurry Tests

• Tests Run in 4 Loop
• Used Three Different Sand Sizes
• 650 micron
• 1250 micron
• 2250 mircon
• 4 Concentrations of each Sand
• (5, 10, 20, and 30 Cv)
• 5 Concentrations Clay for Each
• (0, 2, 4, 5, and 6 Cv)
• Total Number of Tests - 60

Prepared by CRA GIW Industries, Inc.
27
28
US Army Corps. Of Engineers Sand Results
Prepared by CRA GIW Industries, Inc.
28
29
Argyle DiamondSimulated Slurry Pipe Testin 250
and 300 mm Loop
Prepared by CRA GIW Industries, Inc.
29
30
Plugged PipelineWhile Running at CV49
Prepared by CRA GIW Industries, Inc.
30
31
Energy vs. Transport Rate for a 250 mm Pipe
Prepared by CRA GIW Industries, Inc.
31
32
Schematic of Hibernia Oil Platform
Prepared by CRA GIW Industries, Inc.
32
33
Hibernia Slurry Tests

• Phase 1 Determine the head loss characteristics
  of pumping magnetite ore in various pipelines.
  Results to be used to design distributor
  system.Average Solids S.G. 4.392Average
  Particle Size 25 mmA. Tests in combo 254 /
  305 mm loop at 0, 4, 8, and 16 by volume.
  Flow rates up to 550 L/sec.B. Tests in combo
  406 / 457 / 508 mm loop at 0, 6, 8, and 10
  by volume. Flow rates up to 1250 L/sec
• Phase 2 Test Full Scale Mock-up of Actual
  Distributor at concentrations of 0, 4, 8, and
  12 by volume. Flow rates up to 1800 L/sec.

Prepared by CRA GIW Industries, Inc.
33
34
Hibernia Distributor Slurry Test
Prepared by CRA GIW Industries, Inc.
34
35
Hibernia Slurry Tests
Prepared by CRA GIW Industries, Inc.
35
36
Syncrude Coarse RockReady for Loading in 450 m
Loop
Sliding
Lifted
37
Tar Sands Slurry Pipe Friction
Frictional headlosses for tar sand slurries at
40C, as determined in two laboratories
Delivered concentration for coarse rock in water,
Cr 0.066, D 490 mm
38
Dallas White Rock Lake

• 20 Mile Transportation of Sediment
• Slurry Analysis Performed by GIW for Terra-Mar,
  Consulting Engineers
• GIW Extrusion Rheometer tests performed to
  predict pipeline friction
• Silt NOT Sand
• Double Velocity required for Sand
• 3.25 times Friction for Sand
• 3.25 x 2 6.5 times the power
• Results (3) Pumps rather than (9) Pumps
• Results 5,000 HP rather than 27,000 HP
• Huge Project Savings

39
Dallas Process Concept
40
Flow Properties of White-Rock-Lake-Sediment
Slurries
For Terra-Mar Consulting Engineers Determined
at GIW Hydraulic Laboratory
Four tests were performed in the GIW Hydraulic
Laboratory for the Terra-Mar, Consulting
Engineers. These four test were carried out in
the .43 ID Extrusion Rheometer and are
designated and differentiated as follows
41
0.43 ID Extrusion Rheometer Test
tau0 fn(8V/D) Test 4-97 Sm 1.30
42
Scaled Test Results
-dp/dx fn(V) Terra-Mar ID 19in.
43
Application of GIW Test Data
44
Hydraulic Gradeline 24 inch 20 Mile
SystemDredge, Unmanned Booster, and Instruments
45
Radio Link to Dredge Dallas White Rock Lake
46
Main Control Radio Link to Dredge, Phone Line
Link to Remote Booster
47
Remote Booster 6 miles Dallas White Rock Lake
48
System Make Up Water Dallas White Rock Lake
49
Automatic Flow Control Dallas White Rock Lake
50
Dredge SG Approx 1.1 Dallas White Rock Lake
51
Four Corners Mine Transportation Analysis
52
Long Distance Pipeline Pumps are Usually Not This
LargeCentrifugal Pumps are Usually Limited to
6 in Series in One Location
Prepared by CRA GIW Industries, Inc.
52
53
In Florida, Up to 11 Pumps in Series May be
Located Along a 21 Pipeline, 8.5 Miles Long
54
Phosphate Pump and Pipe Tests in the Field
55
Florida Phosphate Matrix Transport Costs
Existing Matrix Train 17, with 19 ID Pipe,
Except 10 WBC 46 (49)s LSA 62 Pit Pump Costs
at Different Concentrations (and TPH), 45,200
Feet (Same Motors, V.S. Pumps _at_ Max Speed, 300
Micron D50) (Assumes 1" Suction Liner Wear, 2"
Casing Belly Wear)
All costs are figured at 6000 hours unless
specified otherwise
56
Kaolin Hydro-transport Pipeline Design
Kaolin Product 37 miles of 12
pipeline d50 8 µm 98-99 lt 50 µm Some
200 µm quartz Cw 24 36 Closed Test
Loop Setup 400 feet long 12 steel pipeline
loop Observation section in inclinable
pipe Fully instrumented to measure both
pipeline and pump characteristics.
Prepared by CRA GIW Industries, Inc.
56
57
3 Pumps in Series Tachonite Tails Pump System
Prepared by CRA GIW Industries, Inc.
57
58
Power Failure Example
3.5 miles
3.5 miles
Pumps located in two pumping stations (3 in
series) with sumps
Prepared by CRA GIW Industries, Inc.
58
59
Piping Arrangements and Considerations
The simplest solution is often the best solution.
Prepared by CRA GIW Industries, Inc.
59
60
Startup on Part Full Line (or filling)
Pump comes up to speed and then operates at point
where H-Q curve of pump intersects the system
resistance for the amount of line filled.

• Unless throttle or controlled by reduced speed,
  operation will result in
• Increased Power
• Possible Cavitation
• If Pump H-Q curve is flat, increases will be
  larger

Prepared by CRA GIW Industries, Inc.
60
61
Startup (Sequencing) of Multiple Pumps in a Line

• After some time (steady flow) all pumps located
  at pit end of line.
• Shortly after starting all pumps located at pit
  end.
• After some time (steady flow) pumps located
  (spaced) along line.
• Shortly after starting pumps spaced along line.
• Starting booster before pit pump.

Prepared by CRA GIW Industries, Inc.
61
62
Water Hammer

• Occurs when rapid change happens
• May be initiated by
• startup
• shutdown
• rapidly closing valve
• water column closure
• Blocked suction
• Disturbance coming from the change
• is in the form of waves traveling outwards

Individual waves pass up and down at constant
velocity, neither being attenuated or undergoing
shape change by each other Waves are reflected
in full or part at each change of section. Net
pressure at a section is sum of pressure
heights Maximums occur at wave crossover points
or initiation or reflection locations.
Prepared by CRA GIW Industries, Inc.
62
63
Water Hammer Example
When cavitation occurs Vapor column created can
be pulled up line by inertia after pump
collapses. As line slows, pump will pick up
sending a second column after the
first. Subsequent closure of two columns will
result in transient.
A closure may be likened to a rapidly closing
value The transient in this case is Which can
be simplified to any air will cushion
Prepared by CRA GIW Industries, Inc.
63
64
Conclusions

• It is Possible to Use Centrifugal Pumps to
  Transport Solids in a Pipeline up to 30 Miles and
  More.
• It is Important to Categorize Slurries Accurately
  in Order to Identify the Most Energy Efficient,
  Lowest Wear Operating Velocity and Concentration.
• Test Lab Pipeline Tests are Necessary in Most
  Cases to Determine Pipe Friction and Pump
  Performance