8 Pipeline Material, Orientation, and Bends
1
INTRODUCTION
A major advantage that pneumatic conveying systems have over alternative mechanical conveying systems is in flexibility in continuous pipeline routing. Pipelines can run horizontally, and with bends in the pipeline, flows can go vertically up or vertically down, with little restriction on numbers of bends or distances. Pipelines inclined upwards are not generally recommended and so flow in inclined pipelines is examined. Up to now pressure gradient has been discussed in global terms of pressure drop available and distance over which a material must be conveyed, with high pressure gradients being required for dense phase conveying. Data is included in this chapter to show how pressure gradient varies with conveying parameters for horizontal and vertical conveying in both dilute and dense phase flows. Conveying parameters were introduced in the previous chapter for pipeline bore and conveying distance. In this chapter scaling parameters are presented for other pipeline features including vertical flow. The influence of conveying parameters on pressure drop across bends is considered, for both dilute and dense phase flow, and losses are presented in terms of both a pressure drop and an equivalent length. Pipeline material is also considered, with particular reference to the use of flexible rubber hose.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
242 2
PRESSURE GRADIENT DATA
All the conveying data presented so far has been for total pipeline systems. This is usually obtained from a test facility comprising a pipeline test loop that generally includes horizontal pipeline lengths, a number of bends and possibly an element of vertical lift. The pressure drop data in the conveying characteristics presented has been for the entire pipeline. In order to isolate the effect of any individual element of pipeline, such as a straight section of horizontal or vertical pipeline, or a bend in the pipeline, pressure tappings must be fitted into the pipeline. Some of these issues are considered in this chapter but are considered in more detail in Chapter 23. 2.1
Horizontal Conveying
Typical conveying data for flow in a horizontal section of pipeline is presented in Figure 8.1. The data is for barite, which is often used as a drilling mud powder. This material has a particle density of about 260 lb/ft3 but despite this it will be seen that the material could be conveyed at solids loading ratios in excess of 100 and at low velocity. For drilling purposes it is used as a very fine powder and so has very good air retention properties in this form. As a consequence of the air retention properties the material will convey in dense phase flow.
Solids Loading Ratio
120
JOO
60 h
40
40 o
20 20
a
'ressure Gradient Ibt7in 2 perl00ft 0 0
50
100
150
Free Air Flow Rate - ftVmin Figure 8.1
Pressure gradient in horizontal flow for barite in 2 inch bore line.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Material, Orientation, and Bends
243
The data in Figure 8.1 is presented in exactly the same form as the conveying characteristics, with material flow rate in Ib/h plotted against free air flow rate in fVVmin. The family of curves plotted is now of pressure gradient in lbf/in 2 per 100 ft length of pipeline rather than a pressure drop for the total pipeline. Lines of constant solids loading ratio are also included as these are simply straight lines through the origin as before. The juxtaposition of these two sets of curves on the one plot is particularly useful for illustrating once again the problem of maintaining flow in dense phase with increase in conveying distance. 2.1.1
Long Distance Conveying
As expected, it will be seen that as the solids loading ratio increases, the pressure gradient increases. At a solid loading ratio of about 100 the pressure gradient is approximately 10 lbf/in 2 per 100 ft length of pipeline. With a limit on air supply pressure because of air expansion problems, and the consequent need to step the pipeline, the scope for long distance dense phase flow is strictly limited. This does not take account of the additional pressure drop due to bends and sections of vertically upward pipeline that might need to be included either. For longer distance conveying there must be a compromise and this is to convey at a lower value of solids loading ratio where the pressure gradient is lower. In Figure 7.2Id, magnesium sulfate conveyed over a distance of 2500 ft is presented and the maximum value of solids loading ratio, with a conveying line pressure drop of 30 lbf/in 2 , is only about 1 '/2. 3
VERTICAL CONVEYING
Apart from the difficulty of finding a suitable wall or structure on which to mount a vertical pipeline for testing purposes, a test loop needs to be used, unless two conveying systems are available, one conveying to the other. The former was used for the test work reported here [1]. An advantage of this method is that the pipeline must go down as well as up and so data can be obtained for both sections of pipeline in every test run. A sketch of the test pipeline is given in Figure 8.2 together with dimensional details. A high pressure top discharge blow tank was used to feed material into the pipeline. The layout of the test facility was such that the material was conveyed vertically down first and then vertically up. The fall and rise elements of the pipeline were both 53 ft long. The total pipeline length was about 185 ft. Two pipelines were available; one of two inch and another of three inch nominal bore, both following an identical routing. Typical conveying characteristics for the total pipeline system are presented in Figure 8.3 [2]. These are for barite conveyed through the three inch bore pipeline. Barite can be conveyed in dense phase, as was illustrated in Figure 8.1 and so conveying with air supply pressures up to 30 lbf/in 2 was possible and solids loading ratios of well over 100 were achieved.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
244
Solids Loading Ratio
120
o o o
X30 'onveying 80 Line Pressure Drop Ibf/in 2
20 E
40
10 Bend Number
0
0
100
200
300
400
Free Air Flow Rate - ft3/min Figure 8.2 Details of test pipeline used for vertical conveying.
Figure 8.3 Conveying characteristics for barite in figure 8.2 pipeline of 3 inch bore.
Once again this illustrates the conveying potential of relatively small bore pipelines in that material flow rates of over 100,000 Ib/h were achieved. For the total pipeline the form of the conveying characteristics is little different from that for other pipeline systems presented in earlier chapters. In order to obtain pressure gradient data for the two test sections there were 15 pressure tappings (seven along the down section and eight along the up section). The first and last tappings at each section were placed about five feet from the bends in order to ensure that any upstream or downstream effects would have minimum influence on the pressure readings. At each location a ring of four tappings was used and all four were coupled to a common point. Results from two tests carried out with a fine grade of pulverized fuel ash (fly ash) in the two inch bore pipeline are presented in Figure 8.4 [2, 3]. The horizontal axis represents the length of pipeline (see Figure 8.2) from the bend in which the flow is horizontal to vertically down (bend number 4), to the bend in which the flow is vertically up to horizontal (bend number 7). The first section, therefore, represents the flow vertically down, along which there were seven pressure tapping locations, and the second section represents the flow vertically up, along which there were eight pressure tapping locations. The vertical axis represents the pressure of the conveying air. The solid lines drawn represent the linearized dependence, from the measured values of pressure, while the dotted lines represent an approximate development of the pressure in the region where the pressure was not measured.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
245
Material, Orientation, and Bends
20 \-
16 Ofl 3 ta eo
12
•O C
-a c
T3 c
CQ
CQ
D
3
4
U
8 9 10 11 12 13
14
I
40
Figure 8.4
60
80
100 Distance - feet
120
140
160
Typical pressure gradient results obtained with pulverized fuel ash.
It will be noticed that in one case the pressure gradient in the vertically down section was negative, while in the other it was positive. In the vertically up flow the pressure gradient was negative in each case, but there was a significant difference between the two tests. The influence that conveying conditions can have on the values of pressure gradient are considered below. 3.1
Flow Vertically Up
Pressure gradient data obtained in this way for the vertically upward flow of barite in the three inch bore pipeline is presented in Figure 8.5. The barite was conveyed over a wide range of both air and material flow rates, and some forty to fifty individual tests had to be carried out in order to provide the necessary pressure gradient data to obtain the plot or performance map shown in Figure 8.5. Once again solids loading ratios well in excess of 100 were achieved with this material. The data is plotted in terms of a pressure gradient in lbf/in 2 per 100 feet of vertically up pipeline. If the data is compared with that for the horizontal pipeline in Figure 8.1, which also relates to the conveying of barite, it will be seen that the pressure gradient values are significantly higher.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
246 140
120
100
Solids Loading Ratio
120
80
Pressure Gradient - '"' M/in 2 perlOOft 24
32
40
E
40
"c3
20
100
200
300
Free Air Flow Rate - ftVmin Figure 8.5 line.
Pressure gradient data for barite conveyed vertically up in 3 inch bore pipe-
Material flow rates are also very much higher but this is because the data is for a larger bore pipeline. It is by comparing sets of data such as this that scaling parameters can be determined, but ideally they need to be of the same bore pipeline, and so this is considered later in this chapter. The data can also be compared with the conveying characteristics presented in Figure 8.3. Figure 8.3 was generated from exactly the same test program as that for the data in Figure 8.5. One was plotted from the total pipeline pressure drop data and the other from the pressure gradient data derived from the pressure tapping readings. A comparison of the two will show that the slope of the pressure gradient lines on Figure 8.5 is very different from the slope of the lines of constant pressure drop on Figure 8.3. Figure 8.5 is for the vertically upward section of pipeline in isolation, while Figure 8.3 is for the total pipeline system, including nine bends. The influence of bends is also considered later in this chapter. Two additional sets of data are included to reinforce the nature of the curves. These are for cement and fly ash, both conveyed vertically up through a two inch bore pipeline. There are clearly differences between the two sets of data but following the comparative data presented in Chapter 4 this will not come as a surprise. The unknown factor is how the different elements of the pipeline contribute to the overall differences observed. Data for the cement is presented in Figure 8.6a and that for the fly ash in Figure 8.6b.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
247
Material, Orientation, and Bends
Solids Loading Ratio
Solids Loading Ratio
60
60 Pressure 140 Gradient lbf/in 2 per 100ft
50
o: o
Pressure • Gradient Ibf/m1 per 100ft
\
£ 40 0)
3
a
*30 _o
o E 1 20 o
E 320
12
20
10
10 0
(a)
16
50
100
50
150
Free Air Flow Rate - fVVmin
(b)
100
150
Free Air Flow Rate - ftVmin
Figure 8.6 Pressure gradient data for flow vertically up in a two inch nominal bore pipeline, (a) Cement and (b) a fine grade of fly ash. 3.1.1 Scaling Parameter In the majority of pneumatic conveying system pipelines the proportion of horizontal conveying is very much greater than that of vertical conveying. A scaling parameter, therefore, is required in terms of an equivalent length of straight horizontal pipeline. In order to provide a comparison between the data for conveying vertically up and conveying horizontally, and hence to obtain the necessary scaling parameter, a rectangular grid was placed on the various sets of pressure gradient data. The grid was set at corresponding values of air and material flow rates, and the ratio of the pressure gradient values obtained from the vertical and horizontal data were determined. The results of this process, carried out for barite in two inch bore pipeline, are presented in Figure 8.7a. They are presented on the same axes, together with the solids loading ratio lines, so that any pattern in the values with respect to conveying conditions could be determined. From this it will be seen that the ratio of the pressure gradient for vertically upward flow to that for horizontal flow varies from a minimum of about 1-9 to a maximum of about 2-4 and that the predominant value is about two.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
248
50
o
83(
Ica20
'B lio
40
(a)
80
120
160
Free Air Flow Rate - ftVmin
40
(b)
80
120
160
Free Air Flow Rate - ftVmin
Figure 8.7 Ratio of vertical to horizontal conveying line pressure drop data for flow in two inch nominal bore pipeline, (a) Barite and (b) fly ash. It can be seen that the relationship obtained covers a very wide range of conveying conditions. A similar analysis, carried out with fly ash in a two inch bore pipeline is presented in Figure 8.7b. It will be noticed that there is very little variation in this ratio from minimum to maximum values of conveying air velocity and from minimum to maximum values of solids loading ratio. The only deviation from a mean value of about two would appear to be at the two extreme limits of the pressure gradient curves, where the data is least reliable. This, therefore, shows that the pressure drop in conveying vertically up is approximately double that in horizontal conveying, for given conveying conditions, over the entire range of conveying conditions. 3.2 Flow Vertically Down In the majority of pneumatic conveying systems, flow vertically down usually occurs only when the pipeline is routed over some obstruction such as a road or railway line. In these cases the influence of the vertically downward section is generally disregarded. It is essential, however, that the additional bends required are taken into account. It is with mining that long vertical pipelines come into their own, both for conveying vertically down as well as vertically up. The removal of muck from the boring of vertical mine shafts is often undertaken pneumatically. Much of the coal
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
249
Material, Orientation, and Bends
mined around the world is obtained from deep mines. With mechanization of coal face operations in the 1970's the mining capability exceeded the hoisting capability of winding gear and so additional means had to be found for extracting the additional capacity. Pneumatic conveying was widely used for this purpose and the conveying of coal 1650 feet vertically upwards at 110,000 Ib/h was quite common in the 1970's [4]. Back-filling of mined out areas is generally a requirement and cement and fly ash are widely used for this purpose. These materials, therefore, are often conveyed vertically down mine shafts. Because of the vast quantities of fly ash being produced around the world from power generation with coal, and the environmental problems associated with the material, the disposal of fly ash in this way is being considered more widely [3]. The longest pipelines conveying material vertically down are probably in South Africa. Ice is used in many deep gold mines as a heat transfer medium for cooling ventilation air. Ice making plant is located at the surface level and the ice produced is pneumatically conveyed over distances up to three miles, with vertically down distances up to about 7900 feet [5]. Pressure gradient data for the pneumatic conveying of cement vertically down in the Figure 8.2 pipeline of two inch nominal bore is presented in Figure 8.8. Although the form of the data is similar to that for the other pressure gradient data presented, it will be seen that signs have been added to the values.
Solids Loading Ratio
60 Pressure Gradient Ibf7in 2 perl00 40
\
O
20
20
50
Figure 8.8 bore pipeline.
100 Free Air Flow Rate - ftVmin
150
Pressure gradient in vertically down flow for cement in two inch nominal
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
250
At high values of solids loading ratio the pressure gradient is negative which means that there is a rise in pressure along the length of the pipeline, rather than a pressure drop. Where the pressure gradient is 16 lbf/in 2 per 100 ft of pipeline, for example, it means that at the bottom of the 53 ft vertical fall in the test facility the pressure will have risen by about 8'A lbf/in 2 . It will also be seen that some of the pressure gradients are positive which means that there is a pressure drop along the length of the vertical fall. The magnitude of the pressure gradient varies with solids loading ratio, and pressure rise for the flow vertically down increases with increase in solids loading ratio. At a solids loading ratio just below about forty the pressure gradient is zero, which means that the material is conveyed with no pressure drop whatsoever under these conditions. At lower values of solids loading ratio there is a pressure drop and this covers the entire range of dilute phase conveying. Two additional sets of data are included to reinforce the nature of the curves. These are for barite and fly ash, both conveyed vertically down through the same two inch bore pipeline. Data for the cement is presented in Figure 8.9a and that for the fly ash in Figure 8.9b. It will be seen that all three materials follow a very similar pattern, although material flow rates differ, as might be expected. The zero pressure gradient curve is also consistent in occurring at a solids loading ratio of about 35 in each case. Solids Loading Ratio Solids Loading Ratio
60
Pressui - Gradient 2 lbf/in per 100 ft
80
Pressure Gradient Ibf/in 2 per 100ft
o o. o •
1 40
100
^^. 140
60
140 120,100 80
50
120
\
£40
w a
* 30 _o
o E
1
E 20
120
20 10 0
(a)
-2
+2
10 50 100 150 Free Air Flow Rate - ft3 / min
0
(b)
50 100 150 Free Air Flow Rate - ft3 / min
Figure 8.9 Pressure gradient data for flow vertically down in a two inch nominal bore pipe, (a) Barite and (b) fly ash.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Material, Orientation, and Bends
251
For conveying vertically down, therefore, materials capable of dense phase conveying could be conveyed very long distances with a relatively low air supply pressure. A particular advantage is that in conveying materials such as cement and fly ash down a mine shaft, the pressure generated at the bottom could be high enough to automatically convey the material to underground mine workings another 5000 ft distant. A problem with this, however, is in the sizing of the pipelines, for the velocity of the material at the start of the horizontal run may be too low as a result of the high pressure generated. This point is considered further in the next chapter on Stepped Pipeline Systems. 3.3
Minimum Conveying Air Velocity
Much has been said about minimum conveying air velocities and conveying line inlet air velocities. The data presented so far has essentially been for total pipeline systems comprising horizontal and vertical sections of pipeline, and bends. Minimum conveying air velocities in vertically upward flow are lower than those for horizontal conveying but in a mix of orientations in the one pipeline it is usually difficult to take the benefit of this into account and so the worst case of velocity requirements for horizontal conveying are usually specified. In horizontal pipelines particles that drop out of suspension, or saltate, will come to rest on the bottom of the pipeline. With an increase in the thickness of the saltated layer the cross sectional area will reduce and there will be a corresponding increase in conveying air velocity. Depending upon the nature of the material this may result in a steady equilibrium situation. More often than not, however, the saltated layer will be formed into dunes and these will be swept up and block the pipeline, often at a bend in the pipeline. In vertically upward flow this process is referred to as choking. When particles drop out of suspension, usually in the boundary layer at first, where the velocity is lowest, they will enter into free fall. At velocities at which particles will settle on the bottom of the pipeline in horizontal flow, particles are likely to be reentrained in flow vertically up because of impact with other particles moving up and the general turbulence. As a consequence minimum conveying air velocities can be lower for vertically upward flow. In mining situations, as discussed earlier, this can be used to advantage where there will be very long runs of vertical pipeline. Where the majority of a pipeline runs horizontal it is more difficult to take advantage of this fact. 4
INCLINED PIPELINES
There is little published information on the advisability of using inclined pipelines. Much of it is anecdotal, but as it is generally experiential it would generally be wise to avoid pipelines that incline upwards. An inclined section of pipeline may well reduce the overall length of a pipeline but their use is not recommended.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
252 4.1
Chapter 8 Upward Incline
The general consensus of opinion is that pipelines inclined upwards should be avoided and that for any vertical rise, a combination of horizontal and vertical sections only should be used. The problem relates essentially to low velocity conveying and the influence that an inclined line might have on the minimum velocity. There is, of course, the additional issue of pressure drop When saltation occurs in a horizontal pipeline, particles will be deposited on the bottom of the pipeline, as mentioned above. In a pipeline inclined upwards, however, particles dropping out of suspension will be more mobile and will tend to roll backwards. The saltated layer will readily form dunes and these will result in pipeline blockage Although the minimum conveying air velocity for vertically upward flow is lower than that for horizontal flow, the minimum conveying air velocity for pipelines inclined upward is higher than that for horizontal pipeline. If it is known that the velocity in an inclined section of pipeline will be high there should be no risk of blockage. It is also understood, however, that the pressure drop in a pipeline inclined upwards is much greater and so on this basis it would be better to keep to horizontal and vertical sections for any vertical rise required. The scaling parameter is one for horizontal flow and two for vertically upward flow. At an angle of inclination of about 60° the scaling parameter is a maximum and is slightly greater than that for vertically up flow at 90° [6]. 4.2
Downward Incline
The mechanism of flow in downward inclined pipeline is somewhat different and so there should be little difference in minimum conveying air velocity from that in horizontal pipelines. Saltated particles will tend to roll in the direction of flow and be re-entrained in the gas flow, rather than form dunes, at velocities just above the minimum conveying air velocity for horizontal flow. 5
PIPELINE BENDS
Although pipeline bends provide pneumatic conveying system pipelines with their flexibility in routing, they do have an impact on the performance of a conveying system. Determining the pressure drop due to bends in a pipeline, however, is not a simple matter. Apart from the influence of the conveyed material, the location of the bend along the length of the pipeline and the geometry of the bend are also likely to have an influence on the pressure drop across the bend. Data on the influence of bends may be required as an equivalent length rather than a pressure drop value. These issues are considered as well as the general influence of bends on conveying performance.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
253
Material, Orientation, and Bends 5.1
Classical Analysis
The difficulties of pressure measurement in pneumatic conveying system pipelines are highlighted most effectively with the problem of measuring the pressure drop across a bend in a pipeline. It is not just a matter of recording the pressure at inlet to and outlet from the bend and subtracting the two readings. This will give a totally false recording, being significantly lower than the actual value. It is necessary to record the pressure at regular intervals along the sections of pipeline both before and after the bend [3]. Part of the problem lies in the complexity of the flow in the region of a bend. The conveyed particles approaching a bend, if fully accelerated, will have a velocity that is about 80% of that of the conveying air. This velocity, of course, depends upon the particle size, shape and density, and the pipeline orientation. At outlet from a bend the velocity of the particles will be reduced and so they will have to be re-accelerated back to their terminal velocity in the straight length of pipeline following the bend. The situation is depicted in Figure 8.10. The pressure drop associated with this re-acceleration of the particles, therefore, is not registered in the bend, but occurs in the pipeline following, and so it must be taken into account as illustrated in Figure 8.10. The method by which the total pressure drop associated with a bend is determined is to instrument the pipeline before and after the bend with pressure transducers. Typical data for a white wheat flour is shown in Figure 8.11 [7].
¥ Ap in Bend * Ap in Line f I Following Bend 1 AP Total ^-^ \ 1
Approach
I | I
Bend
Following Straight
Distance Figure 8.10
Pressure drop elements and evaluation for bends.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
254
Conveyed Material - Flour 17
^c
16
a,
15
Slope = 2-7 lbf/in 2 per 10ft
I Pressure Drop Due to Bend = 2-0 lbf/in 2 I = 75ft
14
k*
-20
-10
0
10
20
30
40
50
Distance From Bend - ft
Figure 8.1 1
Pressure profile in straight pipeline either side of a steel bend.
The bend was tested in a two inch nominal bore steel pipeline and had a bend diameter, D, to pipe bore, d, ratio of 5:1. The bend was tested in the horizontal plane, with the flour conveyed at a solids loading ratio of about 32. The mean particle size of the flour was 78 micron, and the particle and poured bulk densities were 87 and 30 Ib/ft3 respectively. From Figure 8.1 1 it will be seen that the total pressure drop across the bend was about 2-0 lbf/in 2 . Since the pressure gradient in the straight pipeline, both before and after the bend was also available, the equivalent length of the bend could be determined. In the case presented this equivalent length was evaluated at about 75 feet. Re-acceleration of the particles may require a significant distance downstream of the bend, particularly if the particles have a large mass and density, and something of the order of a dozen pressure transducers would be required, as shown. In practical terms this pressure drop is a little high. With eight such bends in a pipeline it would require the output of a positive displacement blower just to negotiate the bends. The conveying air velocity at the bend was about 3500 ft/min and this is almost double that necessary to convey the flour at a solids loading ratio of 32. The data was obtained from a test on an instrumented pipeline in a laboratory facility and is used for illustration purposes. Systems, however, are frequently over designed, particularly if the designers are not certain of the minimum conveying air velocity value, and so there is often scope for improving the performance of existing conveying systems as a consequence.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
255
Material, Orientation, and Bends
5.2 Comparative Analysis An alternative, and potentially quicker, means of determining the energy loss associated with bends is to compare the conveying performance of two pipelines in which the same material has been conveyed. Ideally both pipelines should be of the same bore and preferably of a similar length and contain a different number of bends of the same geometry. By comparing the performance data of materials conveyed in the two pipelines it is possible to determine the influence of the additional bends. Conveying data obtained with barite conveyed through two such pipelines is presented in Figures 8.12a and b. Both pipelines tested were two inch nominal bore and all the bends in the two pipelines had a bend diameter, D, to pipe bore, d, ratio of 24:1. One pipeline was 340 feet long and incorporated nine 90° bends and the other was 330 feet long and incorporated seventeen 90° bends. The bends were uniformly positioned along the length of the pipelines, and there was sufficient length of straight pipeline before every bend to ensure that the material was fully accelerated to its terminal velocity [8]. Solids Loading Ratio
Solids Loading Ratio
.60
50 „ Conveyin
- Line - Pressure
40 I Drop - Ibf7in 2
Conveying -Line Pressure
30
20 0
|'° 0
0
(a)
50
100
150
Free Air Flow Rate - ftVmin
0
0
(b)
50 100 150 Free Air Flow Rate - ft3/min
Figure 8.12 Conveying data for barite in two inch bore pipelines of approximately the same length. Pipeline with (a) 9 bends and (b) with 17 bends.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
256
Conveying characteristics for the barite in the pipeline with nine bends are presented in Figure 8.12a and for the pipeline with seventeen bends in Figure 8.12b. Barite was chosen so that a very wide range of conveying conditions could be examined, from low velocity dense phase to high velocity dilute phase. The conveying data has been presented on the same axes for both pipelines, with material flow rates up to 50,000 Ib/h considered for each, and it will be seen that for the pipeline with 9 bends, conveying line inlet air pressures up to 50 lbf/in 2 were employed, but this had to be increased to 60 lbf/in 2 for the pipeline with 17 bends. The only essential difference between the two pipelines is eight bends and so the difference between the two sets of data can reasonably be attributed to eight bends. With complete sets of conveying characteristics obtained for the same material conveyed through two pipelines of approximately the same length, but with different numbers of bends, it should be possible to compare the results and determine the influence that the bends have, since the influence of pipe bore and conveying distance have been isolated. The comparison is based on the mass flow rates of the barite achieved for given values of air flow rate and conveying line pressure drop. A grid was drawn on each set of conveying characteristics and the ratio of the barite flow rates was determined for every grid point. In order to determine whether there is any pattern in the value of this ratio, with respect to conveying conditions, the values corresponding to the grid points have been plotted on the conveying characteristics for the pipeline with 17 bends. These are shown in Figure 8.13. 0-85
0-75
0-80
40
o o
0-70
2 30
I 20
0-65
Material - barite
? _o fan
"3
10
mp in 330 ft x 17 bends
064
a
mp in 340 ft x 9 bends 50
Figure 8.13 nine bends.
0-69
100 150 Free Air Flow Rate - ftVmin
067
064
200
Ratio of material flow rates in pipeline with 17 bends, to pipeline with
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Material, Orientation, and Bends
257
It will be seen that the material mass flow rates for the line with 17 bends are lower in every case, varying from 88% to 64% of the value obtained with the pipeline with only nine bends. Figure 8.13 shows very clearly that bends have relatively little influence when conveying at very high solids loading ratios with low air flow rates, but have a very significant effect when conveying at low solids loading ratios with high air flow rates. Solids loading ratios have not been shown on Figure 8.13 but this information is available from Figures 8.12a and b. This also shows that no single value can be applied to allow for the influence of bends in a pipeline. An allowance will quite clearly depend on the conveying conditions. Figure 8.13, however, shows that the influence of conveying conditions on the effect of the bends is very uniform and consistent, and so it should be possible to determine a simple relationship between the allowance to be made and some parameter that defines the conveying conditions. 5. 2. 1 Equivalent Length The next stage in the analysis is to assign an order of magnitude, or value, to the allowance to be made for the bends. For this purpose an equivalent length is probably the best way of allowing for the added resistance. An equivalent length of straight horizontal pipeline in feet is therefore required, so that this can be added to the existing pipeline length to give the total equivalent length of the pipeline [8]. As the equivalent length will vary with conveying conditions it is necessary to superimpose regular grids on the two sets of conveying characteristics, as presented earlier and to evaluate the value at every grid point established. The equivalent length of the bends can be determined with a model that relates material flow rate and equivalent conveying distance for a pneumatic conveying pipeline. Such a model was presented in Chapter 7 with Equation 7.10: ......
-
where rhp = mass flow rate of material Le
-
0)
- Ib/h
= equivalent length of pipeline - ft
and subscripts 1 and 2 refer to different pipelines of the same bore The equivalent lengths of the two pipelines will be:
and
Le, Le2
= (340+ 9b) = (330+176) where
b
ft ft
..........
= equivalent length of straight horizontal pipeline per bend
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
- ft
(2)
258
Chapter 8
Substituting Equation 8.2 into 8.1 and re-arranging gives: mn,
330 + lib
m.
340 + 9 b
(3)
It is this ratio that is plotted on Figure 8.13. The only unknown in this equation, therefore, is b. The equivalent length, therefore, will increase in a pattern similar to that shown for the ratios on Figure 8.13. An analysis of the data produced the relationship shown in Figure 8.14. The entire program of test work and analysis was repeated with cement, in place of the barite, and a very similar set of results was obtained. The correlation is in terms of a single parameter, which is conveying line inlet air velocity, which makes its application very convenient. This would indicate that the location of the bend along the length of the pipeline is only of secondary importance, despite the fact that the conveying air velocity will increase along the length of the pipeline. It might, however, be that the difference in particle velocities across the bends do not vary significantly with their position along the length of the pipeline. From Figure 8.14 it will be seen that equivalent lengths of bends can be as low as 5 ft per bend for low velocity dense phase conveying, with conveying line inlet air velocities of 600 ft/min. For dilute phase conveying, however, with a conveying line inlet air velocity of 4000 ft/min, for example, the equivalent length is about 80 ft per bend. The curve continues to rise to higher values of equivalent length with further increase in velocity.
1000
2000
3000
4000
Conveying Line Inlet Air Velocity - ft/min Figure 8.14 Influence of conveying line inlet air velocity on equivalent length of long radius 90° steel bends for conveying barite.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Material, Orientation, and Bends
259
5.2.1.1 Coefficient of Restitution Although the data in Figure 8.14 relates to barite, the entire process was also carried out with cement, as mentioned above. There were obviously differences in the conveying characteristics between the barite and cement conveyed through the two pipelines but the analysis carried out produced an almost identical result in terms of equivalent length [8]. It is suspected that many other materials will follow this same pattern in terms of equivalent length. However, it is believed that the value of the coefficient of restitution between the particles and the bend wall might well be an additional influencing parameter. If materials having a high value of coefficient of restitution impact against a bend the velocity of the particles on leaving the bend will not be as low as those for materials such as flour, barite and cement. As a consequence the energy loss across the bend will not be as high, particularly for higher velocity flows. This point is considered further, later in this chapter, when the analogous situation of conveying materials through rubber hose is investigated. 5.2.2 Pressure Drop An alternative presentation of the data in the form of conveying characteristics is presented in Figure 8.15. The bend loss here is expressed in terms of lbf/in 2 per bend. It will be seen that the most significant parameter is air flow rate, and hence conveying air velocity, with losses varying from about '/2 lbf/in 2 per bend in low velocity dense phase flow to 2/4 lbf/in 2 per bend in high velocity dilute phase flow over the range of air flow rates considered. 50
Solids Loadin Ratio
o o
x'^,100
,80 60
2 40 j5
i 30 I
20 S3
Pipeline Bore = 2 inch
a
10
0
Bend Loss -*•• 40
"" lbf/in
^ 80
120
/bend
160
Free Air Flow Rate - fWmin Figure 8.15 Influence of conveying conditions on pressure drop for barite conveyed through long radius 90° steel bends.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
260
From the pipeline and conveying parameters for the flour, presented in Figure 8.11, the air flow rate was about 155 ftVmin and the solids loading ratio was given as 32. If this data is plotted on Figure 8.15 for the barite it will be seen that the pressure drop would be about 2 lbf/in 2 which is the same value as that reported on Figure 8.11. This is despite the difference in bend geometry. 5.2.2.1 Comparative Values It will be noted that in terms of equivalent lengths the spread of values over the range of conveying conditions considered is of the order of 20:1 from Figure 8.14, but in terms of pressure drop values from Figure 9.15 it is only about 5:1. The values in terms of pressure drop are much closer because pressure gradient values in dense phase are very much higher than those for dilute phase. The use of scaling parameters for evaluating pneumatic conveying system performance and capability is very different from that of summing pressure drop values for individual elements of the pipeline. 5.3
Bend Geometry
The majority of the work reported in this chapter has been undertaken with long radius bends having a D/d ratio of about 24:1. Tests with the wheat flour related to a bend with a D/d ratio of 5:1 but the data agreed quite closely with that for barite in the long radius bends. The influence of bend geometry on the air only pressure drop for bends was considered in Chapter 6 with Figure 6.6 and this is reproduced here in Figure 8.16 for reference. From this it will be seen that it is only with very short radius bends that pressure drops will be high for air only.
Rough Pipes f = 0-0075
Smooth Pipes /i= 0-0045
10
20
30
Ratio of D/d Figure 8.16
Head loss for 90° radiused bends.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
40
Material, Orientation, and Bends
261
Bends having a wide range of geometries are employed in pneumatic conveying system pipelines. Short radius bends and tight elbows are cheaper and easier to install than long radius bends. Blind tees are often used in pipelines in which abrasive materials are conveyed. In order to determine the influence of bend radius on pressure drop and conveying performance a program of tests was carried out with a range of bend geometries [9]. A pipeline was specially built with a double loop in the horizontal plane, in which the bends at the corners could be replaced. The pipeline included eleven 90° bends and seven of these could be conveniently changed. The pipeline was 165 ft long and of two inch nominal bore. A fine grade of fly ash was used as the conveyed material to ensure that tests could be carried out over as wide a range of conveying conditions as possible. A sketch of the pipeline is given in Figure 8.17 for reference. The central group of seven bends, positioned in the corners of the double loop were arranged so that bends of different geometry could be conveniently incorporated. The location of the bends is indicated on Figure 8.17 and were chosen since there was a reasonable length of straight pipeline before the bend to ensure that the fly ash was accelerated to its terminal velocity before meeting the next bend. The group of seven bends, all having the same geometry, were all changed for each test program. Tests were carried out with sets of long radius bends having a bend diameter, D, to pipe bore, d, ratio of 24:1; with short radius bends (D/d = 6); elbows (D/d = 2); and with blind tees. A proportioned sketch of the different bends tested is given in Figure 8.22. Return to Hopper
Supplementary Air t Discharge from Blow Tank
Figure 8.17
Pipeline used for bend geometry tests.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
262 Elbow
Figure 8.18
Long radius
Short radius
Sketch of bends tested.
A complete set of conveying characteristics was obtained with the fly ash conveyed through the pipeline for each of the four different sets of bends. The conveying characteristics obtained with the long radius bends and the blind tees are presented in Figure 8.19. Solids Loading . Ratio ^~-^ -Conveying Line • Pressure Drop '- lbf/in 2
" Conveying Li _ Pressure Drop
\ 32 24
10 0
(a)
50 100 150 Free Air Flow Rate - ftVmin
0
(b)
50 100 150 Free Air Flow Rate - ftVmin
Figure 8.19 Conveying characteristics for fly ash conveyed through the pipeline shown in figure 8.21 having, (a) Long radius bends and (b) blind tees.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
263
Material, Orientation, and Bends
If these two sets of conveying characteristics are compared it will be seen that over a large area of conveying conditions an increase of about 50% in pressure drop is required in the pipeline with blind tees to achieve the same material flow rate in the pipeline with long radius bends. If material flow rates are compared for a given value of conveying line pressure drop it will be seen that the flow rate achieved in the pipeline with blind tees is approximately half of that achieved in the pipeline with long radius bends, particularly at high air flow rates. It must be recalled that these two sets of conveying characteristics relate to the same pipeline, for the 165 ft length of pipeline and four of the bends are exactly the same in each case. These differences, therefore, are due entirely to the change in geometry of only seven of the bends in the pipeline. To provide a full comparison of the different sets of bends the conveying characteristics were compared over the entire range of conveying conditions. The comparison was based on the pressure drop required to achieve a specified material flow rate for a given air flow rate. To do this a grid was drawn on each set of conveying characteristics at regular increments of both air and material flow rates, and the conveying line pressure drop at every grid point was noted [9]. The results of this analysis are presented in Figure 8.20 with both the blind tees and the short radius bends compared with the long radius bends. The long radius bends have been taken as the datum for reference. Solids Loading Ratio
Solids Loading Ratio
f p
A
Pshoit rad
A
Plong radius |
,00% i QQ
50
100
60
10
0
(a)
50
100
150
Free Air Flow Rate - ftVmin
0
(b)
150
Free Air Flow Rate - ftVmin
Figure 8.20 Comparison of performance of long radius bends, (a) With blind tees and (b) short radius bends.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
264
Chapter 8
The numbers on these plots are essentially the ratios of corresponding pressure drops, in terms of a percentage increase, or decrease where there is a negative sign. It will be seen that the pressure drop with the blind tees was about 40% greater than that for the line with the long radius bends, whether for the low velocity dense phase or the high velocity dilute phase conveying of the material. In terms of energy considerations, therefore, blind tees could not be recommended for pneumatic conveying system pipelines with this type of material. A comparison of the short radius with the long radius bends is given in Figure 8.20b and from this it will be seen that there is little difference between the two, although at low values of both air and material flow rates the short radius bends performed better. In terms of bend selection, therefore, long radius bends would only be recommended if there were particular needs for such bends in terms of erosive wear resistance and the minimizing of material degradation. A comparison of the elbows with the long radius bends showed an overall increase in percentage ratios of about 15% over those shown on Figure 8.20b for the short radius bends [9]. The data overall, therefore, shows a very close correlation with the data for air only in Figure 8.16, with respect to the influence of bend geometry on pressure drop. 5.3.1 Pocketed Bends Bradley [10] undertook a program of tests with a 90° pocketed bend and reported that the pressure drop was only marginally better than that for a blind tee. The pocketed bend was of the vortice variety and was tested in a similar manner to that discussed in relation to Figure 8.11. 5.4
Bend Location
The general recommendation is that a reasonable length of straight pipeline should proceed a bend in a pipeline, particularly the first bend in a pipeline following material feed into the pipeline. This area in a pipeline is particularly critical because the conveying air velocity is at its lowest and the material is generally fed into the pipeline at zero velocity. Ideally the particles should be accelerated to their terminal velocity. With large and high density particles this requires a relatively long distance. If this is not possible, and particularly if the first bend is a blind tee or pocketed bend, it may be necessary to increase the conveying air velocity to compensate and this, of course, will increase the energy requirements for the system. By similar reasoning no two bends in the pipeline should be spaced too closely together, particular in the low velocity area in a pipeline. Long radius bends are also to be avoided if the first bend must feed vertically up. The problem here is analogous to inclined pipelines. A long radius bend in a horizontal to vertically up orientation will have a significant section of pipeline on an incline, and a higher conveying air velocity may be required for material to negotiate such a geometry.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Material, Orientation, and Bends 6
265
PIPELINE MATERIAL
Although all the data so far has related to the pneumatic conveying of materials through steel pipelines, not all pipelines are made of steel. Rubber hose is widely used in pneumatic conveying systems, both for pipeline and bends, and in systems where a degree of natural flexibility is required, such as in vacuum off-loading and mobile systems. By virtue of its natural resilience rubber hose can often be used to particular effect in reducing erosive wear with abrasive materials and in minimizing degradation with friable materials. As a pipeline material it is particularly suited to the conveying of certain sticky and cohesive materials. For the off-loading of ships, that have self-discharging facilities, high pressures are generally employed in order to keep the discharge time to a minimum. With materials such as cement, conveying air pressures up to 100 psig can be utilized, and hose is available that will meet this requirement. A particular application is the transfer of drilling mud powders, such as barite, bentonite and oil well cement, from supply boats onto off-shore drilling platforms. As materials have to be off-loaded from boats in rough seas, a long length of hose is used to connect the discharge system on the boat with the fixed pipeline on the drilling rig. Road trucks and rail tankers are most conveniently off-loaded through lengths of flexible rubber hose, whether the vehicles are self off-loading or not. In these applications it would be impractical to use rigid metal pipelines because of the time required to achieve the necessary alignment. An unknown quantity, however, is whether the pressure drop for rubber hose will be any different from that of steel pipeline. 6.1
Pipeline Pressure Drop
In order to determine whether there is any difference in conveying performance between steel and rubber hose a program of tests was specifically undertaken. A 140 ft long pipeline of two inch bore steel pipeline that incorporated five 90° bends was used. Oil well cement was conveyed through this pipeline and its conveying characteristics were obtained. A 140 ft length of two inch bore rubber hose line was then strapped to the steel pipeline. By this means exactly the same routing and bend geometries were replicated. The oil well cement was then conveyed through this pipeline and its conveying characteristics were obtained. The two sets of conveying characteristics are presented in Figure 8.21. Oil well cement, like ordinary portland cement, is capable of being conveyed in dense phase and at low velocity and so the two sets of data cover a very wide range of conveying conditions. Tests were carried out with air supply pressures up to about 28 lbf/in2 gauge and so, as the pipeline was relative short, solids loading ratios up to about 200 were achieved.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
Chapter 8
266
Solids Loading Ratio -*• 200 160 130 100
80
24
50
Conveying Line Pressure Drop 18Q150]20 r .bt/in
SoUds Loa(jjng
00 Ratio
\
50
20
40
40
40 -
3 30 oi 13
16
20
S 10
S 10 0
0
(a)
50 100 150 Free Air Flow Rate - ftVmin
00
(b)
50
100
150
Free Air Flow Rate - ftVrnin
Figure 8.21 Conveying characteristics for oil well cement conveyed through 140 ft long pipeline of two inch bore of different materials, (a) Steel pipe and (b) rubber hose line.
From the two sets of conveying characteristics it will be seen that the nature of the curves is very different. With the steel pipeline there is a distinct pressure minimum point in the pressure drop curves. Conveying performance appears to be similar at low values of air flow rate but are widely different at high values of air flow rate. In order to compare the performance of the oil well cement in the two pipelines a grid was drawn on each of the sets of conveying characteristics, in much the same way as reported above for the program undertaken with bends of different geometry. The ratio of pressure drops for corresponding air and material flow rates were evaluated. The results of this exercise are presented in Figure 8.22. From Figure 8.22 it will be seen that there is a gradual increase in pressure drop for the rubber hose line, compared with that for the steel pipeline, with increase in air flow rate. The lines of constant percentage increase drawn on Figure 8.22 slope in the same way as the lines of constant velocity on the conveying characteristics, as illustrated on Figure 7.4, and so it is clearly a conveying air velocity effect. In dense phase flow at very low velocities there is little or no difference between the two pipeline materials, but with higher velocity dilute phase flow the pressure drop for flow through the rubber hose line is 50% greater than that through the steel pipeline.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
267
Material, Orientation, and Bends
+10
50
-1 x 100%
-10
40
£ 30 .
20 Material - Cement Pipe Bore - 2 inch Length- 130 feet
10
1
0
Figure 8.22
I
i
40
1
1
I
L
80 Free Air Flow Rate - ft3/min
120
160
Comparison of pressure drop data for steel and rubber hose lines.
The program of tests was repeated with another drilling mud powder (barite) and a similar set of results was obtained [11]. 6.2
Coefficient of Restitution
It is suspected that the coefficient of restitution between the particles and the pipeline wall plays an important part. Rubber, being resilient, will have a lower coefficient of restitution for impacting particles than steel. If the rubber absorbs more of the energy of impact of the particles than the steel, a greater pressure drop will result with the rubber pipeline, due to having to re-accelerate the particles from a lower velocity. This is why the pressure drop for flow through the rubber hose is greater than that through the steel pipeline, and since pressure drop increases with (velocity)2, this is why it increases with increase in conveying air velocity [11]. 7
EQUIVALENT LENGTH
Scaling, whether for system design or for undertaking a review of alternative conveying systems for a given duty, is generally undertaken in two stages. The first stage is to scale to the length and routing required and the second is to scale with respect to pipeline bore. Scaling with respect to length and pipeline routing is usually in terms of an equivalent length of the pipeline. The equivalent length incorporates vertical lift and bends, as well as horizontal pipeline, and is expressed in
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
268
Chapter 8
terms of horizontal length. A factor of two is suggested for a scaling parameter for vertically upward sections of pipeline. Equivalent lengths for bends were presented in Figure 8.14. For non radiused bends and tight elbows an additional allowance will have to be made. An additional allowance will also have to be made for rubber hose, but the data given here can be used in estimating appropriate values. Although the information presented relates to particular conveyed materials it must be appreciated that at this point in time there is no universal solution to the problem of designing pneumatic conveying systems and for determining the conveying capability of a pipeline. Different materials will behave differently, as was illustrated with total pipeline systems in Chapter 4.
REFERENCES 1.
2.
3. 4. 5.
6. 7.
8.
9.
10.
11.
12.
P. Marjanovic. An investigation of the behavior of gas-solid mixture flow properties for vertical pneumatic conveying in pipelines. PhD Thesis. Thames Polytechnic (now The University of Greenwich) London. 1984. D. Mills, J.S. Mason, and P. Marjanovic. The influence of product type on dense phase pneumatic conveying in vertical pipelines. Proc Pneumatech 2, pp 193-210. Canterbury. Sept 1984. D. Mills. Measuring pressure on pneumatic conveying systems. Chem Eng, Vol 108, No 10,pp 84-89. Sept 2001. J. Firstbrook. Operation and development of the pneumatic coal transportation system. Proc Pneumotransport 5. BHR Group Conf. London. April 1980. T.J. Sheer, R. Ramsden, and M. Butterworth. The design of pipeline systems for transporting ice into deep mines. Proc 3rd Israeli Conf for Conveying and Handling of Paniculate Solids, pp 10.75-80. Dead Sea. May/June 2000. D. Mills. A review of the research work of Professor Predrag Marjanovic. Proc 4th Int Conf for Conveying and Handling of Paniculate Solids. Budapest. May 2003. M.S.A. Bradley and D. Mills. Approaches to dealing with the problem of energy losses due to bends. Proc 13lh Powder and Bulk Solids Conf. pp 705-715. Chicago. May 1988. P. Marjanovic, D. Mills, and J.S. Mason. The influence of bends on the performance of a pneumatic conveying system. Proc 15th Powder and Bulk Solids Conf. pp 391399. Chicago. June 1990. D. Mills and J.S. Mason. The influence of bend geometry on pressure drop in pneumatic conveying system pipelines. Proc 10th Powder and Bulk Solids Conf. pp 203214. Chicago. May 1985. M.S.A. Bradley. Pressure losses caused by bends in pneumatic conveying pipelines: effects of bend geometry and fittings. Proc 14th Powder and Bulk Solids Conf. pp 681694. Chicago. May 1989. P. Marjanovic, D. Mills, and J.S. Mason. The influence of pipeline material on the performance of pneumatic conveying systems. Proc Pneumatech 4. pp 453-464. Glasgow. June 1990. D. Mills. Using rubber hose to enhance your pneumatic conveying process. Powder and Bulk Engineering, pp 79-87. March 2000.
Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.
