Duan C.G., Karelin V.Y. Abrasive Erosion and Corrosion of Hydraulic machinery

Подождите немного. Документ загружается.

254 Abrasive Erosion and Corrosion of Hydraulic Machinery

Table 5.7 Chemical compositions and physical properties

Specimen

Stainless Steel

(316L)

Hastelloy C

Material A

Material B

com

0.03

0.08

0.05

0.03

Chemical

positions (%)

13.5

53.5

5.5

6.0

17.0

16.5

25.5

28.0

Metallic

structure

Austenite

and ferrite

Austenite

and ferrite

Physical properties

Vickers

hardness

198

210

280

500

Specific

gravity

8.0

8.7

7.7

Slurries were prepared from the particles and the corrosive liquids in

Table 5.8. The liquids have nearly the same chemical compositions as those in

an actual stack gas scrubber. Silica sand particles were used for the reference

slurry. The gypsum particles were the same as that contained in the slurry of

the scrubber.

Table 5.8 Slurries

Liquid or Slurry

Deionized water

sol., pH 4

C1-(C.CI

), 200 ppm

sol., pH 2

Particles

Silica sand

Gypsum

Particle properties

Mean diameter, um

Vickers hardness

1200

20-50

Corrosion tests were conducted in the jet-in-slit apparatus by corrosive

liquids without particles. The slurries of the solid particles and deionized

water were used for the erosion tests, and those of the solid particles and the

corrosive liquids for the erosion-corrosion tests.

Erosion - Resistant Materials 255

5.2.2 Test Results

Some results of the erosion test as well as the erosion-corrosion tests are

shown in Figures 5.8 and 5.9. Fairly good linear relations existed between the

volume loss of specimen and test duration, and constant damage rates

independent of the test duration were obtained from the slopes of the lines.

0.8

0.2

Vibratory Apparatus

3161 Stainless Steel

Silica Sand 60wf/. pH 2

^£*pH4

-^^V

^^^ *

->^^£>

20 30

Time,

min

Figure 5.8 Erosion-corrosion damage versus testing time for

316L stainless steel

(frequency 19.9 kHz; amplitude 20 urn; silica sand

60 weight %; and respective liquids)

1.0

<■>

0.6

E QA

0.2

Jet-In-Slit Apparatus

3I6L Stainless Steel

Silica Sand 6wW, pH 2 /

^ pHA

-"""D.W.

60 90

120 Time, min

Figure 5.9 Erosion-corrosion damage versus testing time for

316L stainless steel

(jet velocity 1.7 m/s; silica sand 6 weight %; and respective liquids)

256

Abrasive Erosion and Corrosion of Hydraulic Machinery

The results of erosion-corrosion tests in the slurries of silica sand

and corrosive liquids are shown in Figure 5.10 (vibratory apparatus)

and Figure 5.11 (jet-in-slit apparatus). Each apparatus has established

the same order in damage rate, namely stainless steel > Hastelloy C >

material A > material B.

-»—■

Vibratory Apparatus

Silica Sand 60wH.

CD D.W.

E33 pH«

ESS pH4*Cl"

ES3 pH2

Figure 5.10 Damage rates of materials in vibratory apparatus

(Frequency 19.9 kHz; amplitude 20 urn; silica

sand 60 weight % and respective liquids)

Jet-In-Slit Apparatus

Silica Sand 6wt7.

CZ3 o.w.

T77X pH U

ESS pH/.vCI"

E3S3 PH2

Hastelloy C

T—

£*

Figure 5.11 Damage rates of materials

jet-in-slit apparatus

(jet velocity 1.7 m/s; silica sand 6 weight % and respective liquids)

Erosion - Resistant Materials

257

Erosion-corrosion tests in the gypsum slurry were conducted only in the

jet-in-slit apparatus (the particle size was so small that cavitation occurred in

the vibratory apparatus). The damage rates shown in Figure 5.12 are

extremely small as compared with those in the silica sand slurry (Figure

5.11). This may be attributed to the size as well as to the hardness of the

particle. As to the order in damage rate, stainless steel and Hastelloy have

exchanged their positions, though material A and B stayed in the same

position. The specimen surfaces showed a matted appearance as well.

7 ut.Tr>-t;iif Anoaratus

Figure 5.12 Damage rates of materials

jet-in-slit apparatus

(jet velocity 1.7 m/s; gypsum 12 weight % and respective liquids)

5.2.3 Damage on Pump Components

The materials A and B had been used for the components of pumps A and B

respectively, which had a common size: bore diameter, 125 mm; impeller

diameter, 265 mm; revolution, 1750 rpm; discharge quantity, 85 m

/h. They

pumped gypsum slurry of sulfuric acidity at the same time under nearly the

same condition: the solid particle concentration has ranged between 20 and 40

wt%, and pH value between 2 and 4.

The average depth of the damage on the blade top of the impellers and the

surface of the casings has been measured at regular intervals. Figure 5.13

258

Abrasive Erosion and Corrosion of Hydraulic Machinery

shows the results. It should be recognized that the damage depth increased

linearly with the duration of the operation just as in the case of the test

apparatuses. In each pump, the casing was damaged more deeply than the

impeller, and the damages of pump B were one-half or one-third smaller than

A which is in good agreement with the test results shown in Figures 5.11 and

5.12.

Casing A

tmpeKer A^^^

Casing B

6^--

Impelter B

_i——***—"*

•

i5xio

Time, h

Figure 5.13 Erosion-corrosion damage versus operation time

for pump components in actual service

5.2.4 Requisites for Laboratory Tests

The test results obtained by the jet-in-slit apparatus agreed well with the

actual performance in the pumps not only on the order but also on the relative

magnitude of the damage rate. In other words, a quantitative relationship has

been established between the test result and the actual damage of the pump.

By using the relationship, it is possible for the apparatus to assess

quantitatively the damage rate of any material, which is to be used as pump

components under the same operating condition.

It is, however, indispensable to know the effects of the test parameters on

the test results for estimating the damage rate to any material which is to be

used under some other conditions. With this object in mind, a quantitative

comparison of the damage rate between the test results and the field

Erosion - Resistant Materials 259

performance of the material B is given in Table 5.9, in which the damage rate

in mm/h of the pump components was divided by the bulk particle flow rate to

obtain the damage rate per unit mass of the particle i.e. mm/kg. It arouses

interest that the damage rate in mm/h of the specimen in the gypsum slurry is

of nearly the same magnitude as those of the pump components, though the

rate in mm/kg is by 10

times larger than those of the components. On the

contrary, the damage rates of specimen in the silica sand slurry are much

larger than any damage rate of the actual pump.

Table 5.9 Comparison between test results and

actual performance of material B

Jet-in-slit test

Silica sand slurry

Gypsum slurry

Actual performance in

Impeller

Casing

Damage rate (mm/h)

pump

6.2X10"

2.3X10"

5.5X10"

1.4X10"

Damage rate (mm/kg)

8.6X10"

1.6X10"

3.2X10"

8.2X10"

<1> Slurry particles:

A point of advantage in using the silica sand slurry is the shortening of

the test duration. For example, it takes, at a minimum 20 h, to

determine the damage rate of Hastelloy C in the gypsum slurry, but

only 2 h is enough for precise measurements in the silica sand slurry.

As to the test results, however, stainless steel and Hastelloy C have

exchanged their positions in the order of damage rate when the silica

sand slurry was used for the gypsum slurry (comp. Figure 5.11 with

5.12). This suggests that the actual slurry in question should be used

even though the purpose of the test might be a rough selection of

materials. In the tests for estimating the damage rate of a material in

the actual plant, it is absolutely necessary to use the actual slurry.

<2> Slurry liquid:

Tests on a material in the vibratory apparatus by silica sand slurries

have resulted in nearly the same magnitude of damage rates

260

Abrasive Erosion and Corrosion of Hydraulic Machinery

independent of the slurry liquid (Figure 5.10). A similar tendency was

recognized in the test results of the jet-in-slit apparatus where silica

sand particles were used (Figure 5.11). In contrast to these, the damage

rates of the material varied largely depending on the slurry liquid when

gypsum particles were used (Figure 5.12). This is apparently because

the erosiveness of gypsum particles is less intense than that of silica

sand, and correspondingly, the effect of corrosion has appeared more

distinctly on the test results. Thus, in this aspect as well, it is necessary

to use the actual slurry in laboratory tests.

<3> Particle impact parameters:

The average velocity of slurry flow over the impeller surface is on the

order of

m/s, which is the volumetric flow rate divided by the area of

the flow cross section. Over the casing surface, it is on the order of 10

m/s on the assumption that the slurry is circulating at the same velocity

as the impeller. It is generally recognized that the erosion rate by

particles impingement is proportional to the 2.3th (ductile material) ~

6th (brittle material) power of the impact velocity. The difference in the

damage depth between the impeller and the casing is less than expected

from the slurry flow velocities. This discrepancy may be attributed to

the differences in the particle flow rate and their impact angle. These

parameters are characteristic of each components and depend decisively

on the operating conditions.

The particle's impact velocity in the jet-in-slit apparatus is 1.7 m/s as

shown in Table 5.6, which is naturally not coincident with those in the pump.

It is impossible and not necessary for them to coincide with each other,

because the quantification of the impact velocity effect is possible, and hence

the damage rate under any impact velocity may be easily estimated based on

test results under known conditions.

The impact angle is the most important parameter that gives decisive

influence not only on the extent of damage but also on its nature. In the jet-in-

slit test, a slurry jet impacted on the disk-shaped specimen, and then the slurry

run radially over the specimen surface to the periphery. Thus, the impact of

particles was at a right angle in the central part of the specimen and, is

oblique at its periphery. It was found, however, that the oblique impact of the

particles barely contributed to the weight loss of the specimen because the

damage (decrease in the thickness) was concentrated on the central part of it.

Erosion - Resistant Materials

261

The damage of the specimen was consequently due to the particle impact at a

right angle. In contrast with this, the particle impact angle on the impeller and

casing surface was apparently oblique except for the small central portion of

the impeller. Then a question arises why such an excellent coincidence could

be obtained for the erosion behavior of the test specimen with that of pump

impeller? An answer to this question is given by Figure 5.14 which was

obtained for commercially pure iron using the slurry of silica sand and

deionized water in fixed-bed type vibratory testing apparatus in which particle

impact angle was precisely regulated (See Oka, 1986). You can see that

erosion rate at the impact velocity of 1.4 m/sec depends greatly on the impact

angle but it does not at all when the impact velocity was decreased to 0.4

m/sec. This behavior is explained as follows. The difference in the damage

mechanism (deformation and cutting) originates itself in the shape of craters

which were created on the metal surface by particle impact: the oblique

impact of a particle causes a creator of slender shape due to some sliding of

the particle following the impact and cutting is predominant on the crater

surface. Whereas a particle impact at a right angle brings about a round

crater where deformation is predominant. When a particle was impacted at a

low velocity (0.4 m/sec for example), the particle possibly can not slide over

the surface even though the impact angle was oblique. Or it can slide but

without leaving any scratch on the surface. In other words, the damage

mechanism will possibly be the same with that of a right angle impact

irrespective of the impact angle of the particle provided that the impact

velocity is lower than a certain critical velocity. In the case of Figure 5.14 the

critical velocity lies between 1.4 and 0.4 m/sec. In the case of the slurry

pump, the critical impact velocity is surely higher than this because a softer

particle (gypsum) strikes a harder metal surface as compared to the

experiment of Figure 5.14.

In conclusion, the point of reasons why the test results of the jet-in-slit

apparatus have agreed with the actual performance of the material might be

the low damage rate (mm/h) in the apparatus, which was inevitable in using

the actual slurry in the field. The specimen with a large testing surface has

achieved the measurement of such a low damage rate and yet of small

weight, as well as by the stability of the apparatus which could bear the

operation of such a long duration.

262

Abrasive Erosion and Corrosion of Hydraulic Machinery

E 1.0

Jo.5

'(/)

Figure 5.14 Effects of impact angle on erosion rate of iron specimen

in fixed-bed facility (frequency 19.9 kHz; amplitude 24 urn

and 8 urn; silica sand and deionized water, 40°C)

5.3 Organic Polymer Linings

M. Matsumura

Excellent ability to resist erosion is inherent in some organic polymers as is

described in section 5.1 However, since they have certain weak points

regarding mechanical strength and heat resistance, they are usually applied as

lining to parts of machines and equipment used in ambient temperatures. For

example, steel pipe lined with polyurethane was applied by a clay factory in

England to a kaolin tailing pipeline in 1970. Currently it is being used

successfully in a dam deposit pipeline in Japan. Its applications to mining and

dredging operations are also being evaluated. For further applications and

wider use, however, the following problems have to be resolved.

1> The manufacturing cost of polyurethane lined pipe is much higher than

conventional steel pipe.

2> The flange joint and mechanical joint of the lined pipe are very expensive

and troublesome. A welding joint has to be developed.

3> Experience in using polyurethane lined pipe is not extensive. Its long-

term durability is as yet unknown.

30 60 90

Impact Angle

(deg)

Erosion - Resistant Materials

263

To solve these problems, Kawasaki Steel Corporation and Mitsubishi

Chemical Industries Co. Ltd. of Japan developed a system for room

temperature curing of polyurethane (RTV) and established an economical

manufacturing technique for RTV lined pipe (See Owada, 1986). In the

following, new technical developments are discussed and compared with the

conventional technique.

5.3.1 Conventional Polyurethane Lined Pipe

Conventional thermosetting polyurethane uses a polyether-polyurethane

system which consists of Toluenediissocyanate (IDI) - polytetramethylene

ether glycol (PTMG) as prepolymer and 4,4"-methylenebisorthochloroaniline

(MOCA®, Dupont Company) as curing agent. Typical molding conditions

and mechanical properties of thermosetting polyurethane are shown in Table

5.10. Heating treatment is indispensable for the mixing, casting and curing

processes, for promoting chemical reaction and for obtaining the appropriate

properties.

The manufacturing process of thermosetting polyurethane lined pipe is

shown in Figure 5.15. The steel pipe is blasted and degreased to get a proper

surface for the following primer application. The purpose of the primer

application is to protect the pipe from corrosion and to bond polyurethane

elastomer to the steel surface. Phenol-Polyvinyl (butyl alcohol) primer is

commonly used. This primer has to be baked at nearly 100°C, and the

optimum coating thickness is usually 20 to 50 um when it is dried.

The prepolymer and the curing agent are separately melted and degassed

at 80°C and 120°C respectively. Then, they are mixed in quantities of 12

(prepolymer) to 1 (curing agent).

Mixed polyurethane is poured into the steel pipe, which is preheated to

between 100° and 110°C. After leveling the liquid polyurethane, centrifugal

casting is performed for straight piping. For bend piping and other fittings,

mold casting is used. The lining is cured at 100° and 110°C for 3 to 4 hours.

Lastly, polyurethane lined pipes are usually aged for 1 to 2 weeks at room

temperature before shipping.

There are the following problems with this conventional manufacturing

technique.

1> Heating processes result in extensive energy consumption which together

with heating equipment increases the cost of production.