Design of a slurry erosion test rig - download.xuebalib.com

Wear, 140 (1990) 331-344 331

Design of a slurry erosion test rig

J. B. Zu, I. M. Ihtchings and G. T. Burstein Lkpartmat of M&&L Scisnce and Metiwm, University of Ca?xbridge, Pembroke Street, Cannbridge CB2 342 &LIE)

(Received December 31, 1989; revised April 20, 1990; accepted April 30, 1990)

Abstract

The design of a jet impingement slurry erosion teat rig, built for laboratory use, is presented. This apparatus gives good control over many of the important test pammeters, such as impact velocity, solid particle concentmtion and impact angle. An ejector nozzle is employed to entrain sand particles from a sand bed into a stream of water to form a slurry; after impingement, the abrasive particles and the water phase are separated and recycled. This makes the rig simple, economical and easy to operate, and its pump and pipeline remain free from erosive wear. Experimental results are presented to illustrate the operation and performance of the rig.

1. Introduction

Slurry erosion has attracted much research interest over the last three decades, owing partly to problems encountered with the transport of minerals and other materials in large-scale pipelines. Research into wear by slurry erosion has been carried out mainly by generating and examining wear on test samples in laboratory-scale apparatus, to reveal the basic mechanisms of the wear process and to provide data for designing components. Slurry erosion test methods fall basically into two categories: pipe wear tests and laboratory simulation tests. Sedation testing is widely adopted because it is low in cost, relatively easy to set up and operate, and quick to produce results, although pipeline testing provides conditions which are closer to industrial practice.

In wear testing of pipes, pipe samples are fixed either in operating industrial pipelines or, in most cases, in closed loops. Wear caused by the shu-ry flow is then recorded by weighing or by monitoring the change in pipe wall thickness. Although some research has been carried out by pipe wear testing [ l-71, the high costs and the long times needed for detectable wear are major disadvantages. In contrast, simulation testing allows relative motion between a test sample and slurry to be generated more simply. The generally lower cost and flexibility of simulation testing, coupled with better control over the conditions to which the specimens are exposed, make it considerably more reliable for laboratory research into the mechsnisms of erosion and the influence of operating variables on wear rates.

0043-1648/90/$3.50 (0 Elsevier S~uo~ted in The Netherlands

332

A wide range of apparatus hss been designed to conduct laboratory- scale sluny erosion studies. For example, a ring of pipe rotating in a vertical plane about a horizontal central axis like a wheel and partly filled with slurry was used by Jacobs and Boothroyde [S]; Mistry 121 used a length of straight pipe, rotating about a horizontal axis and again containing slurry to generate erosive wear. Such apparatus can provide a useful alternative to full-scale pipeline tests. These techniques have disadvantages, however, such as the difllculties of mo~to~g the pattern of slurry flow relative to the inner surface of the sample pipe, of monitoring the erosion conditions, and of sample preparation.

In addition to the rotat~g-pipe systems, many types of pot testing eq~pment have been described. Lee and Clark [ 9] used a test rig in which specimens are attached to a vertical rotating shaft. The specimens move in a horizontal plane in a pot filled with slurry. Similar equipment was used by Tsai et al. f lo], with an impeller attached to the rotating shaft under the specimens, to stir the slurry and to prevent sedimentation of the erodent particles. A pot tester with two horizontal discs holding four specimens, rotating on concentric vertical shafts in opposite directions, has been reported by de Bree t% al. ill]. Schumacher [ 121 adopted a different type of pot tester in which three hubs, each capable of holding eight specimens, rotate in a vertical plane; the specimens are moved through sedimented erodent particles in a s~tion~ slurry. Such pot testers are easy to operate and provide rapid results for ranking resistance to slurry erosion of different materials. They have been successfully used to characterize erosion resistance and to provide data for selection of materials. Once again, the limi~~ons of these methods stem from the lack of control over important erosion parameters. The erodent particle concentration depends very much on the effectiveness of mixing the erodent particles in the slurry by stirring, either directly by the specimens, or by a separate impeller, In these systems it can be difficult to define the erodent particle concentration and to ensure its ~~~~. The booties of measuring impact velocity and angle are clear. Along the length of a specimen away from the rotation axis, the tangential velocity varies. This velocity tends to differ from the impact velocity, since the slurry moves to some extent with the rotating specimen. The complete of the relative flow con~tions between the shnry and the specimen surface renders difhculty in quantifying the impact angle and ascertaining its uniformity under these test conditions,

A different type of apparatus has been developed by Madsen [ 131, in which the shnry is moved past stationary specimens, up to 16 in number, by a rotating impeller. All 16 specimens undergo erosion simultaneously under the nominally similar test conditions, which makes the method very suitable for the comparison of the erosion resistance of dEerent materials. EIowever, the angle of impact cannot be defined with accuracy, and the rotation speed of the impeller provides the only measure of slurry velocity.

For many purposes the jet ~p~ement test me~od provides an iterative to the meson described above, in which the particle impact conditions can

be better defined. Various kinds have been reported, which can be divided into two groups: those which use pumps for slurry circulation [ 14-161 and those which do not. The main advantage of the latter is that unnecessary pump wear is avoided. Sagiies et al. [ 171 used a compressed gas system to force shnry from a tank through a nozzle. The rig adopted by Matsumura and coworkers [ 18, 191 employed an ejector nozzle to accelerate the shury, for which the following advantages were described. Damage occurred only on the surface of the specimen, and not elsewhere in the test apparatus. The amount of erodent used, specimen size and power consumption were all small. Four specimens could be tested simultaneously under nominally identical conditions. The particle concentration was easily controlled over a wide range and the reproducibility of the test results was excellent. Nevertheless the difiiculty of measuring the impact velocity and the fact that the tests are conducted on fully submerged specimens constitute signiilcant disadvantages.

In the present work, a shury erosion test apparatus has been developed which maintains most of the merits and overcomes the disadvantages of the rig employed by Matsumura and coworkers. Instead of being fluidized to form a uniform slurry, the erodent particles are sucked up into the ejector and well mixed with the carrying fluid before striking the specimen. Im- pingement takes place outside the ejector, which not only gives ease in controlling the velocity and angle of impact but also makes the specimens readily accessible.

2. The test rig

2.1. Description A schematic view of the rig is shown in Fig. 1. The driving fluid, water,

is circulated from the water tank (114 1 capacity) and fed to the ejector by a three-stage pump (three small series-linked centrifugal pumps, designed originally for domestic central heating systems). A region of low pressure is created inside the ejector, and a mixture of water and sand particles is sucked up through the vertical suction tube, the bottom of which reaches into the submerged bed of sand. This sand-water mixture is further mixed with the driving water, and the resulting shury is then accelerated through the exit nozzle and strikes the specimen, whose planar surface is held vertically in the transparent Perspex test chamber. The specimen holder, fitted with an impact angle gauge, can be rotated about its vertical axis. This makes it possible for the slurry which is ejected horizontally from the exit nozzle to impact the specimen at any pre-set angle between 0” and 90”. After striking the specimen the slung falls back into the tank. The sand particles sediment out while the water flows through a stainless steel mesh filter into the right- hand part of the tank, almost free from sand particles. The velocity of the sand-water slurry from the exit nozzle, and the concentration of erodent particles in the shury can be adjusted by controlling one or more of the

334

Fig. 1. Schematic diagram of the slurry erosion test apparatus (not to scale). Details of the ejector are given in Fig. 2.

following: the output of the pumps, which is controlled elec%ricaIly, the bypass fIow rate, the setting of the suction valve and the geomet4 of the ejector. The water-circulating pipeline is made from reinforced polyfvinyl chIoride) tube to avoid corrosion which would be experienced with steel or copper tubii. The conical collector ~ol~ro~ylene) allows sand particles to fall to the bottom of the sand bed, where they are entrained into circulation, ensuring that most of the erodent particles are recirculated. The amount of erodent necessary for a set of tests is thereby rendered relatively small, e.g. between 5 and 6 kg for silica sand.

This test apparatus has some advantages over alternative rigs, It is simple in design, is low in cost and is easy to operate and to remove and replace specimens. Erosion damage develops largely on the test specimen, although some wear of the exit nozzle is inevitable. Only small specimens are required and the erosion co~~tio~ are well controlled and readily mendable. Some environmental effects, such as the corrosiveness of the slurry, are potentially easily adjustable, although this feature is not discussed in the present paper. The exit nozzle is readily ~ter~h~geable when it is necessary either to vary the impact velocity (see below) or to replace a worn nozzle.

2.2. Operation Specimens are tested for shury erosion by cleaning and weighing after

exposure to the slurry jet for a fixed period. By mounting the specimen and the exit nozzle on a single stainless steel beam, the relative position between the exit no&e and the specimen is reproducibly fixed. This ensures that the shrrry strikes the specimen on an identical site during successive cycles of erosion.

336

The impact velocity and tie sand ~u~~~nt~t~~~ are d~te~ed by collecting the slm leaving the exit nozzle for a known period. The impact velocity is derived from the measured cross-sectional area of the exit nozzle and the volume of the slurry collected. It is assumed that the shu-ry is well mixed in the ejector so that the erodent particles have the same velocity as the water, The particle weight concentration within the slurry is calculated by weighing a known volume of slurry.

The performance of the ejector was found to be very sensitive to its structure and geometry, Figure 2 shows the design of the ejector adopted. The driving nozzle and the exit nozzle were made from stainless steel. Different internal diameters for the driving nozzle (o? = 2, 2.5, 3 and 3.7 mm) and the exit nozzle (Cr =4.5, 5, 5.5, 6 and 6.5 mm) were used. The main body of the ejector was formed from a polypropylene T-piece w&h an inner diameter D1 of 6.5 mm. The lengths of the ejector were L1 =28 mm and Z&=35 mm. The performance was opt~zed by adjusting the position of the driving nozzle. It was found that the best performance was obtained with the horizontal distance between the end of the driving nozzle and the axis of the suction tube (distance x in Fig. 2) limited to less than 5 mm. The effect of variation in x on the measured slurry composition is plotted in Fig. 3, where a sharp decay for x > f 5 mm is observed. For all the tests presented below, a value of s= -3 mm was used.

Figure 4 shows the effect of the suction valve on the performance of the ejector. The &ect is recorded in terms of the pressure Pv measured at the inlet of the ejector [marked V in Fig. 1). The graphs illustrate the ejector performance as a function of Pv when other factors are fixed. Figures 4(a) and 4(b) show that the impact velocity and particle concentration both vary

suction

lube

l l * l

8

’ -20 -10 0 10 20

distance

x (mm)

Fig. 2. Details of the ejector for the apparatus shown in Fig. 1. The dimensions are deilned in the text,

Fig. 3. The effect of position of the driving nozzle exit on the erodent concen~a~on (driving nozzle diameter d= 2.6 mm; exit nozzle diameter D- 5.5 mm; water feed pressure P,= 1.72 bar; water feed flow rate Q ~4.3 1 min-‘; erodent, 600-1000 pm silica sand). The distance x is defined as the distance from the central axis of the suction tube, with positive values being closer to the exit nozzle, and negative w&es more distant.

0- 0.85 a.90 0.95 1.00

absolute prc?ssure in ejector %, (be3

Cc>

Fig. 4. The effect of the suetion valve a~~~e~t (recorded as the pressure Pv inside the ejector) on the performance of the test rig (d=2.5 mm; 0=5.14 mm; Pi-l.72 bar; Qm4.3 1 min-I; erodent, WO-lOOO i*m silica sand): (a) impact velocity e, as a function of Pv; (b) erodent concentration C (wt.%), as a function of Pv, (c) water feed rate Q as a fun&an of PU

Iinearly with Pv, Figure 4(c) indicates that the suction valve setting does not Sect the flow rate of the driving water si~c~tly before it enters the ejector. The flow rate & and feed pressure Pt of the driving water remain constant when Pv varies. $ and Pt were measured with the flow gauge and pressure gauge attached to the pipeline as shown in F’ig. 1.

By adjusting the bypass valve the feed pressure Pr casl be altered to change the impaet velocity and particle ~on~entra~on. In Fig. 5 it can be seen that, when the suction valve is fully open, the impact veIocity and particle concentration both increase with increasing feed pressure Pf, for a fixed ejector geometry.

F’igure 6 shows the dependence of the impact velocity and pa&Me concentration on the ratio d/D of the diameter of the driving nozzie to that of the exit nozzle of the ejector. The velocity v of impact increases with increasing d/D while the concentration C of sand particles decreases with increasing d/B as shown in Figs. 6(a) and 6(b) respectively. This leads to an empirically linear refationship between v and C (within the scatter of the data) as indicated in Fig. 6(c). The data points in this graph represent four ejector geometries, achieved by fking the driving nozzle diameter to d=2, 2.5, 3 and 3.7 mm, and altering the exit nozzle diaeter to vary the ratio

0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 water feed pressure water feed pressure

Pf(W PfkW

(a) 01

Fig. 5. The effect of water feed pressure Pr on the performance of the rig (d = 2.5 mm; D = 5.14 mm; erodent, 600-1000 pm silica sand): (a) impact velocity as a function of Pi; (b) erodent concentration as a function of PC

0.0 0.2 0.4 0.6 0.6 1.0 0.0 0.2 0.4 0.6 0.6 1.0 1.2

nozzle diameter ratio (d/D) nozzle diameter ratio (d/D)

(a) @I

40

0 2 4 6 8 1012

impact velocity

v (m s-l)

Cc>

Fig. 6. The effect of ejector structure (d/D) on the performance of the rig, using d=2 mm (0), d-2.5 mm (X), d=3 mm (0) and d=3.5 mm (A) (erodent, 600-1000 q silica sand): (a) impact velocity as a function of d/D; (b) erodent concentration as a function of d/D; (c) effect of impact velocity on concentration of erodent derived from Figs. 6(a) and (b).

d/D. Each set of data gives a linear relationship between v and C. For a particular driving nozzle, C approaches zero when v exceeds a certain value (e.g 4 m s- ’ for d = 2 mm). At this point the ratio d/D is so high that the pressure drop inside the ejector becomes too low for erodent particles to be entrained. From Fig. 6(c) it is clear that, for a fixed ejector body, a given range of particle concentrations C and jet velocities v can be achieved by

338

selecting the approp~ate dimensions for the driving and exit nozzles. This selection can be made using the d&a in Fig. 6, For example, if a particle conce~~ation of 15% is required, then a velocity of between 3.8 and 7 m s-’ (Fig. 6(c)) can be achieved. If a velocity of 5 m s-’ is required, an ejector with the ratio tllr of about 0.5 (Fig. 6(a)) should be adopted.

From the above, it is clear that with this app~atus, using the present geometries, the impact velocity and particle concentration can be varied from 0 to 8 m s-’ and from 0% to more than 30% respectively. Control is achieved by using ejector structures with different nozzle diameter ratios d/D and by adjusting the driving water feed pressure and the vacuum created in the ejector, once the rig is calibrated.

It was also found necessary to monitor the temperature rise in the reservoir tank (see below). Such a temperature rise results from operation of the pumps for extended periods. The data are shown in Fig. 7 and demonstrate a maximum rise of 13.5 “C from an ambient temperature of 20.5 “C. This maximum was achieved after about 15 h, beyond which it remained constant.

3.1. Experimental procedure Rounded silica sand partic’fes with diameters in the two ranges from

425 to 600 pm and from 600 to 1000 pm were used for calibrating the rig and conducting erosion tests. Rectangular specimens of mourn, copper, mild steel (containing 0.15% C) and alumina were tested. The specimens, about 3-5 mm thick, were cut to 35 mmx 30 mm and were ah given the same surface finish (1200 grit silicon carbide abrasive paper). The ~~~ and copper were of commercial purity and all materials were used in the as-received condition with no further thermal treatment. The alumina was a glass-bonded sintered material, with noes 97.5% alumina content (Deranox, Morgan Mate&& Technology plc). Specimens were tested one at a time by mounting onto the specimen holder and eroding for a fixed period. They were then dismounted for weight measurement. The cleaning procedure

running time (W

Fig. 7. Slurry temperature as a function of the running time of the rig for an ambient temperature of 20.5 “C.

339

involved washing with tap water, rinsing in acetone and drying in warm air. After weighing, the specimens were remounted onto the specimen holder for continued erosion. The Vickers hardness of the test materials was measured at a load of 0.98 N (100 gf) prior to erosion, with the same surface Cnish as that described above. Experiments were performed at ambient temperature, although some warming of the slurry occurred during their progress as shown in Fig. 7.

3.2. Results Initial experiments were aimed at dete rmining the effects of the degra-

dation of erodent particles arising from their recirculation; such degradation usualIy leads to a decrease in erosion rate because the particles become smaller and more rounded. These tests were conducted using 600 to 1000 pm sand particles eroding ahuninium at a velocity of about 5 m s-’ and a particle concentration of about 20% by weight to estimate the erodent service life. Figure 8(a) shows the specimen mass loss, and Fig. S(b) the differential erosion rate E ’ (mass loss of target material divided by the erosion time), each plotted against total erosion time. In this experiment a new specimen was used for each incremental erosion period to minimize possible effects of changing surface topography on erosion rate as erosion proceeds. The results therefore reflect only changes in the erosivity of the system as time progresses. To a good approximation the data points in Fig. 8(a) lie on a straight line although some curvature towards a higher t is seen after about 900 min. The total time represented in Fig. 8 would have involved more than 600 cycles for the erodent particles. The erosion rate (Fig. 8(b)) shows a small rise over the first 200 min. It then remains constant up to a time of about 900 min. A slow decline, by a maximum of about 20% in erosion rate, is then observed; this last decline is thought to be due to degradation of the erodent particles. The maximum decline is reached after about 2300 min, involving more than 450 impact cycles of erosion for each sand particle.

2000

1000 2000

erosion time (min)

3000 0 1000 2000 3000 time t (mln)

64 0) Fig. 8. Effects of erodent particle degradation due to recirculation of silica sand (600-1000 pm) slurry impacting onto ahuninhun (v=4.9 m s-l; C= 19.2%; impact angle, 40”): (a) accumulated mass loss as a function of t; (b) differential erosion rate E' as a function of t. Each data point was achieved by erosion of a freshly polished specimen for periods between 30 and 240 min. The abscissa represents the integrated erosion time.

340

The initial rise in erosion rate E ’ (to 200 min in Fig. 8(b)) is believed to be due to the slow rise in temperature observed during the early running of the equipment; this temperature rise is quantified in Fig. 7. The rise of 13.5 “C shown in Fig. 7 represents a decrease in the viscosity of pure water from 0.990 to 0.734 cP; although the water used for the present work was not purSed, such a decrease in viscosity is expected to result in a corresponding increase in impact velocity. For Fig. 8, the MO effects of temperature rise and sand degradation render the erosion rate approximately constant for the total time of the experiment, with a maximum variation of only & 15% from the mean. The service life of sand particles can therefore be determined by setting a Emit to the maximum difference in erosion rate between the highest and the lowest values.

The ability to vary the impact velocity continuously allows the effect of impact velocity on erosion to be investigated. Figure 9 shows the relationship between erosion rate and impact velocity measured for WI-lOOO pm silica sand slurry impinging onto aluminium. This establishes empirically that the erosion rate, expressed as material mass loss per unit mass of erodent, increases with increasing impact velocity according to a power law, having an exponent a(log IQ/&Jog v) = 2.7. This compares with values reported in the literature of between 2.3 and 2,5 for ductile materials [ZO] and between 2.0 and 3.4 for metals eroded by air-borne particles [2 1 I.

The reproducibility and stability of the test conditions allow the ranking of materials according to their resistance to shrry erosion. Figure 10 shows the relationship between the target material mass loss and the mass of erodent used (the latter of which is proportional to the erosion tune), for the four different materials ~~~, copper, mild steel and ahunina, all at an oblique impact angle of 30”. All materials showed a linear increase in mass loss with increasing time. Alumina showed the lowest erosion rate, and ahuninium the highest. Figure 11 illustrates the relationship between erosion rate and Vickers hardness Hv of the material (with the hardness having been

‘“I l

.

05. bb

. . . j

0, L*-...J 1 10

impact velocity

” (m s-1)

Fig. 9. Erosion &e E as a function of impact veloei~ v for silh eand QD3-lOOO WI] SIU.W erosion of aiumini~ at an impact angle of 30".

F'ig. 10. hfass of target matefi eroded as a function of mass of erodent used for 6C@-l~o gm silica sand slurry eroding aluminlum (0), copper (+), mild steel (0) and hmina (@I Se an impact v&&y Qf 7.7 m s-1, with C-4.5% and m impact angle of 30”.

341

10 100 1000 0 102030405060708090

Vickers hardness (H,) impact angle (degrees)

Fig. 11. Erosion rate as a function of Vickers hardness Hv of the target for silica sand (600-1000 pm) slung erosion at 7.7 m .3-l impacting at angles of 30” (0) and 90” (0) (C=4.5%). The target materials in increasing order of hardness are aluminium, copper and mild steel. a(log E>/a(log Hv) = - 0.8.

Fig. 12. Erosion rate as a function of impact angle for ahunini~ (0), copper (O), mild steel (Cl) and alumina (B) eroded by silica sand (600-1000 pm) slurry at v =5.3 m s-l and C= 17%.

measured before erosion of the specimen surfaces) for the three ductile metals at both 30” and 90” impact angles. The erosion rate is shown to decrease as the target hardness increases. Although data for only three materials are presented, they are plotted according to a power law with an exponent a(log E)/a(log Hv) = - 0.8 as shown in Fig. 11. This exponent lies

within the two observed values of - 0.58 (achieved by determining the target hardness after erosion) and - 1 .O (achieved by determining the target hardness of the annealed material before erosion) described by Sheldon [22] from a study of the effects of target hardness on erosion rate using air-borne particles.

The variation in erosion rate with impact angle is shown in Fig. 12 for the four target materials. The erosion rate of alumina, a brittle material, increases monotonically with impact angle, showing a maximum at near- normal incidence. The erosion rates of the metals in contrast increase with increasing impact angle only at small angles of incidence, with peak erosion occurring at about 40”, and slowly decline at larger impact angles. The impact angle is clearly an important factor in ranking materials according to their resistance to erosion. From the results reported here, alumina shows the maximum resistance to erosion. Additionally, however, the largest dif- ference between the erosion resistance of alumina and that of, for example, mild steel occurs at a fairly small impact angle of near 30”-40”, where the metals all have their maximum erosion rates. This advantage of alumina in terms of resistance to shury erosion diminishes as the impact angle increases towards 90”.

4. Discussion

The data described above demonstrate the successful application of the new rig for examination of slurry erosion. The rig is very simple to construct and to operate. It also advantageously allows control of the target impact angle and velocity to close tolerance. The construction of the conical collector

342

enables the sand batch used in a particular shrrry to be recycled almost completely, and the data shown in Fig. 8 demonstrate that this can be performed through many cycles (at least for the aluminium target) without significant degradation of the erodent particles. The tendency for the effect of a rise in temperature in the early stages of a run and the effect of degradation of the erodent particles to annihilate each other (Fig. 8) is perhaps fo~tous; separation of these effects could be achieved by running the apparatus to operating temperature prior to insertion of the specimen with associated measurement of the change in impact velocity. Nevertheless, the qu~ti~tive effects are rather small and are unlikely to alter the trends in the data significantly. Currently, the slurries investigated have been water based but, because of the closed-loop design of the equipment, there is no reason in principle why alternative liquids may not be used. Because of the filtration and sedimentation procedure involved in recycling the water compo- nent of the slurry, only some of the fines produced by degradation of the sand enter the pump stage of the rig, and wear of the pump components is consequently minimal.

In its present construction, the apparatus is capable of a continuous range of impact velocities up to about 8 m s- ‘; this could doubtless be increased by increasing the pump capacity and making appropriate modi- fication to the ejector nozzle geometry. As currently set up, however, the impact velocity and the erodent concentration in the slurry (0 d CG 25%) can be varied independently of each other only within certain ranges, as shown in F’ig. 6(c). The upper Limit for the slurry eon~en~~on is due chiefly to the nature of the ejector. For a given concentration the impact velocity can be preset by selecting the appropriate ejector nozzle diameter, once c~b~tion has been achieved.

The slurry erosion behaviour of the materials described above is largely in accord with previous work. The dependence of erosion rate on impact angle has been discussed 123-261. For air-borne particles, maxima in the erosion rate have been observed at impact angles of 20” and 90” respectively for ductile and brittle materials [23], although the angle corresponding to max&num erosion also depends on the size and shape of the erodent. Truscott [24] has shown the maximum erosion rate in pipelines to occur at impact angles of 50”-60”. The present data are thus in approximate agreement with these values; in particular, it is to be noted from Fig. 12 that the one brittle material examined shows maximum erosion at 90” whereas that for the three ductile metals be at about 40”. The data are not however, in complete accord with those of Levy and coworkers [25, 261, who examined the slurry erosion of aluminium and copper using a jet impingement technique. Their results show either a monotonic increase in erosion with impact angle or a maximum at 40” (similar to Fig. 12) followed by a minimum at 60” and a second maximum at normal incidence. It should be noted from Fig. 12 that the three ductile metals show a small decline in erosion rate for angles greater than about 40”, tailing to a plateau value towards 90”. The origins of this difference in behaviour are not yet understood.

343

6. Concluding remarks

An apparatus for shnry erosion testing has been described, which provides a useful alternative to existing methods for laboratory-scale experimentation in low velocity slurry erosion. Once calibrated, the equipment is capable of a continuous range of impact velocities up to a detiable maximum, and of target impact angles. Test results over extensive periods of continuous experimentation have proven the rig to be reliable for ranking the resistance of materials to shury erosion under different test conditions, and consequently for investigation of the mechanisms of slurry erosion.

Acknowledgments

We are grateful to the Chinese Petroleum Industry Minktry for flnancial support of J.-B. Z. and to Professor D. Hull for provision of laboratory facilities.

References

1 A. G. Bain, Getting working data on coal pipelines, Engineering, 196 (1963) 322-323. 2 V. R. Ml&y, Wear of slurry system components due to abrasion, in L. G&tins (ed.), Wear

in Slurry Pipelines, BHRA Irlfonnation Series Number I, BHRA Fluid Eng., Cranfleld, Bedfordshire, 1980, pp. 69-108.

3 E. Debreczeni, T. Meggyes and I. Tarjan, Measurement methods in an experimental rig for hydraulic transport, Proc. 5th Int. Corlf: on the Hydraulic !Fmnsport sf Solids in Pipes, Hannover, May 8-11, 1978, BRHA Fluid Erg., Cranileld, Bedfordshire, 1978, Paper Gl.

4 A. J. Karabelas, An experimental study of pipe erosion by turbulent slurry flow, Proc. 5th Int. Corlf: on the Hydraulic !Pransport of Solids in Pipes, Hanvwver, May 8-11, 1978, BRHA Fluid Eng., Cranileld, Bedfordshire, 1978, Paper E2.

5 S. Murakami, T. Kawsshima, S. Tsukahara and T. Okada, Wear test of pipe Iinings for hydraulic transport of dam deposit, Proc. 7th In.& Coqf on the Hydraulic Tmnsport of Solids in Pipes, Send&, September k6, 1980, BRHA Fluid Eng., CranfIeld, Bedfordshire, 1980, Paper H2.

6 H. Hocke and H. N. Wilkinson, Testing abrasion resistance of slurry pipeline materials, l’ribol. Int., II (1978) 289-294.

7 M. C. Roco, P. Nalr and G. R. Addie, Test approach for dense slurry erosion, in J. E. Miller and F. E. Schmidt, Jr. (eds.), Slum Erosion: Uses, Applications, and Test Methods, ASTMSpec. Tech. PubL 946, 1987, pp. 185-210.

8 B. E. A. Jacobs and J. Boothroyde, Pipe wear testing 1976-1977, in L. G&tins (ed.), Wear in Slurry Pipelines, BHRA Iqfkmution Series, Number 1, BHRA Fluid Eng., Crantleld, Bedfordshire, 1980, pp. 50-87.

9 Y.-H. Lee and H. MCI. Clark, The relative erosivity of coal-oil, coal-water, and petroleum coke-oil slurries, ln J. E. Miller and F. E. Schmidt, Jr. (eds.), Sluny Ems&x Uses, Applications, and Test Methods, ASTM Spec. Tech. Fubl. 946, 1987, pp. 47-01.

10 W. Tsai, J. A. C. Humphrey, I. Cornet and A. V. Levy, Experimental messurement of accelerated erosion in a slurry pot tester, Wear, 68 (1981) 289303.

11 S. E. M. De Bree, W. F. Rosenbrand and A. W. J. de Gee, On the erosion resistance ln water-sand mixtures of steels for application in slurry pipelines, Proc. 8th Iti. CM on the Hydraulic Transport of Solids in Pipes, Johannesburg, August 25-27, 1982, BRHA Fluid Eng., Cranfield, Bedford&he, 1982, Paper C3.

344

12 W. J. Schumacher, Ball mill and hub test methods for slurry erosion evaluation of materials, in J. E. Miller and F. E. Schmidt, Jr. (eds.), Slum Erosion; Uses, Applications, and Test Methods* ASTM Spec. Tech. Pub1 946, 1987, pp. 5-18.

13 B. W. Madsen, A study of parameters using a new constant-wear-rate shrrry test, in K C. Ludema (ed.), I%oc. Int. Co?& on th% Wear of Matsria-ls, American Society of Mechanical Engineers, New York, 1985, pp. 345-854.

14 J. Zahavi and H. J. Wagner, The role of protective flhn removal and regrowth on the rate of erosion-corrosion in metals, &WC. 5th In&. Cc@ on Erosion bg L&t&3 and Solid Impact, Caged ~a~esh~~, Sept-ember 3-6, 1979, University of Cambridge, Cambridge, 1979, Paper 50.

15 A. Elkholy, Prediction of abrasion wear for slurry pump materials, Wear, 84 (1983) 39-49. 16 I. G. Wright, D. K. Shetty and A H. Clauer, Shrrry erosion of WC-Co cermets and its

relationship to materials properties, Froc. 6th Int. Co@ on Emsim bg4 L+.qzd an& Solid Impact, Cambridge, Cambri.dgesh&q September S-S, 2983, University of Cambridge, Cambridge, 1983, Paper 63.

17 A. A. Sagiies, D. K. Spencer, V. K. Sethl and G. A. Sargent, Shury erosion and abrasion of metal-ceramic coatings, ln J. E. Miller and F. E. Schmidt, Jr. (eds.), Slurry Eroskm Uses, A~l~~~, and Test Methods, ASTM Spec. Tech. Arbl. 946, 1987, pp. 19-44.

18 Y. Oka and M. Matsumura, Erosive wear testing apparatus-slmulatlon of erosion caused by slurry of low impingement velocity, in K. C. Ludema (ed.), Proc. Iti. Cqf on WeaT of Materials, American Society for Mechanical Enghteers, New York, 1983, pp. 360-366.

19 M. Matsumura, Y. Oka and M. Yamawaki, Slurry erosion+orrosion of commercially pure iron in fountain-jet testing facet-mech~m of erosion (corrosion under comparable intensity of erosion and corrosion), Proc. 7th Int. Cm& m Erosim bg Liquid ad Solid Impact, Cambridge, Cambridgeshire, September 7-10, 1987, University of Cambridge, Cambridge, 1987, Paper 40.

20 1. M. Hut~gs, Recent advances in the ideal of solid particle erosion, &f&z, Mar&., Ekctr., 365-366 (1980) 185-192.

21 C. M. Preece and N. H. Macmillan, Erosion, Annu, Rev. Ma&r. Sci., 7 (1977) 95-121. 22 G. L. Sheldon, Effect of surface hardness and other material properties on erosive wear

of metals by solid particles, J. Eng. Mater. Technd, PP (1977) 133-137. 23 A. W. Rug and S. M. Wlederhorn, Erosion by solid particle impact, Trec&i.@3 Ma&z. S&.

TechmL, 16 (1979) 93. 24 G. F. Truscott, A literature survey on wear ln pipelines, in L. Gittlns (ed.), Wear in Slunz/

fipeliws, BhXA &jbnmtian Series Number 1, BHRA Fluid Eng., Cranfleld, Bedford&he, 1980, pp. 23-50.

25 A. V. Levy, N. Jee and P. Yau, Erosion of steels in coal-solven~ slurries, I&p. LBL17238, 1984 (Lawrence Berkeley Laboratory).

26 S. Ka-Keung Li, J. A. C. Humphrey and k V. Levy, Erosive wear of ductile metals by a particle-laden high velocity liquid-jet, I&p. I&&-11959, 1980 (Lawrence Berkeley Laboratory).

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具