luxembourg greetings from cavitation methodol… · • korto’s motto: stay with the problem...

Greetings from

Korto Cavitation ServicesKorto Cavitation ServicesLuxembourg

www.korto.com

KortoMultidimensionalTechnology

Korto´s Korto´s basic services and productsbasic services and products

• In situ multidimensional cavitation tests:- Assessment of true prototype turbine

cavitation characteristics- Identification of turbine parts that cause

erosion- Operation optimisation with respect to

cavitation• High performance multidimensional

cavitation monitoring systems

When might you need Korto?When might you need Korto?

• Commissioning a turbine:- Model predictions correct ?- Contract requirements met ?

• Refurbishing or uprating a turbine:- True state before and after

• Routine operation:- Operation optimisation to achieve minimum

erosion- Control of ageing effects and incidents

FurtherFurther services and productsservices and products

• Diagnostic tests of hydropower unitdynamics

• Correcting unit’s dynamic behaviour• General monitoring systems for hydro

units (cavitation, vibration, air gap,magnetic field, temperatures, etc.)

What is Korto?What is Korto?

• Korto’s staff is a highly specialised teamof world-renowned experts from sevencountries

• Our background:Proprietary leading-edge research basedon extensive full-scale experience

• Korto’s motto:Stay with the problem until it is solved

How do we work?How do we work?

If the task is related to cavitation,our first step is the

MULTIDIMENSIONALDIAGNOSTICCAVITATION TESTIN A SPECIFIC PLANT

We bring along up to 250 kg of cargo,depending on the problems we have to deal with.For cavitation only, we need less. We set up our”test headquarters” in a corner of the machinery room.

There we install our equipment for signal and data acquisition and analysis. We mount our sensors and we look for the sources of operation parameters.

Let us remember:

What is cavitation? It´s the generation, development and disappearance of cavities in water; these cavities are filled with water vapour and gas.

What is its origin? A drop in pressure caused by a local increase of flow velocity close to the runner blade and in free vortices.

What are its consequences? Erosion, turbine efficiency drop, turbine instability, vibration, noise, fish mortality.

Turbine cavitation quality is determined by: - head and suction head, - turbine design, - accuracy of the runner finish, - state of the runner surface.

This quality varies during the exploitation: - Initial surface iregularities and the irregularities

resulting from repairs cause cavitation erosion, and it, in turn, intensifies the irregularities. - The incidental passage of hard bodies or sand through the turbine have the same effect.

Our in-plant cavitation tests and cavitation monitoring are performed vibro-acoustically.

Suitable sensors are used to listen to:

•••• hydroacoustic cavitation noise in water

or

•••• structure-born noise caused by cavitation (sound that is spread through the metal structure of a turbine).

We use sensors on the dry side or, occasionally, on the wet side of the turbine: •••• broad-band hydrophones •••• fast pressure transducers •••• fast accelerometers •••• structure-born noise transducers

On the dry side, we install the sensors on:

the draft tube wall, if accessible

the guide-vane shaft

the man-hole wall

the turbine bearing

.

We analyse the noise-power dependence on - degree of loading, - head, and - suction-head.

Normalised turbine power

00,7 0,8 0,9 1,0

.

A simple example follows.

Noise power

00,7 0,8 0,9 1,0

00,7 0,8 0,9 1,0

.

2

3

1

Nor-malisednoisepower


The measurement can reveal different cavitation mechanisms - there are 3 of them in this case.

00,7 0,8 0,9 1,0

.

2

3

0,7 0,8 0,9 1,0

0

0

0

1

1

1

1

It can even yield a quantitative description of the mechanisms.

Nor-malisednoisepower


Such data can be used for permanent monitoring. Hereis the output of a simple cavitation monitoring system:

Cavitation intensity

Power setting

Head water level

Tail water level

9 12 15 18 21 24 3 6 9

Time (hours)

The case shown is an exception: the data in it is easy to interpret.

In a typical case, much more has to be done in order to extract reliable, useful information from the data acquired.

A systematic approach to this is enabled by ...

KORTOMULTIDIMENSIONALTECHNOLOGYThis original approach has been successfullyused on Francis, Kaplan and bulb unitsand has passed reviews in ASME and IAHRpublications and at world hydro conferences.

KORTOMULTIDIMENSIONALTECHNOLOGY

consists of

a proprietary software for the processing of signals and data,

which are acquired by a rather high number of spatially distributed sensors

over a broad range of a unit´s operatingpoints.

In a diagnostic test, which can lastfrom 1 to 7 days per unit - depending on theproblem - a large amount of data is collected.

100-300 GByte per unit is a normal quantity.

The multidimensional diagnostic tests are madein such a way that - once the data has beenacquired and preliminarily checked - any furtherrequired off-line analysis can be performed.

This enables iterative procedures which clarifynew findings without needing to repeat the test.

Analysis duration: 1-12 weeks

An illustration of one step of the analysis,which reveals cavitation mechanisms:

Normalisednoise power

Turbinepower(MW)

Noise frequency (kHz)

Guidevane Instantaneous

runner position (°)

Another step of the analysis - Recognising the role of the guide (and stay) vanes:


090

180270

360

15

1015

200

1

2

3

4

Guidevane

Blade 5Blade 1

Blade 5Instantaneousrunner position (°)

In such a pattern, the contribution of therunner blades can also be identified:


In a raw form, the peaks seen in the previous figureslook like this:

Normalised noise power(radial co-ordinate)

vs.Instantaneous runner position(angular co-ordinate)

The peaks describe variations in thecavitation intensity while a runnerblade is passing through the disturbedflow behind a guide vane.

An illustration of the turbine-power dependence follows.

The patterns vary depending on sensor location andturbine power setting.

70 MW

Turbinepower

0 MW

Cavitationthreshold

70 MW

0 MW

70 MW

0 MW

Below the cavitation threshold, the patternsare almost circular since flow noise and othersources of background noise do not dependon the instantaneous runner postition.

70 MW

0 MW

Once again: Below the threshold

70 MW

0 MW

At the threshold

70 MW

0 MW

High above the threshold

0° 90° 180° 270° 360°

38

35

30

25

20

one runner blade one guide vane

Turbinepower

(MW)

Instantaneous runner position

Review of such results recorded in one sensor location:

By analysing 100-300 Gbyte of datarecorded at 20-30 power settingsin 8-30 sensor locations

(depending on the case), and processing this datain the manner as partially illustrated above,

one achieves a set of

TURBINE CAVITATION CHARACTERISTICS

as follows ...

1 24

68

101 2

1 416 17

38

40

4 2

44

46

4 8

500

0. 1

0. 2

0. 3

0. 4

0. 5

0. 6

0. 7

0. 8

Ne t h ead 1 16.5 +-0 .2 mTailwater 1 23.7 3 +-0.06 mHe adwater 244. 73 +-0. 07 mPlan t powe r 243 +-6 MW

Runner blade

Unit power (MW)

1, 48 MW% = To tal in tens ity

Power (MW)

Componentof thecavitationintensityon a runnerblade,influencedby a guidevane(% of thetotal)

Runner blade

Detailed cavitation characteristicFor each guide vane - one characteristic

Guide vane 1

All cavitationmechanismsincluded

1 24

68

101 2

1 416 17

38

40

4 2

44

46

4 8

500

1

2

3

4

5

6

7

8

9

10

Ne t h ead 1 16.5 +-0 .2 mTailwater 1 23.7 3 +-0.06 mHe adwater 244. 73 +-0. 07 mPlan t powe r 243 +-6 MW

Runner blade

Unit power (MW)

1, 48 MW% = To tal in tens ity

Power (MW)

Componentof thecavitationintensity ona runnerblade(% of thetotal)

Runner blade

Runner cavitation characteristic


1 2 3 4 5 6 7 8 9 1 0 11 1 2 1 3 14 15 16 17 18 19 20

3839

4041

4 243

444 5

4647

4 84 9

500

1

2

3

4

5

6

7

8

9

10

Net h ead 11 6.5 +-0. 2 mTailwater 12 3.73 +-0 .06 mHea dwa ter 244. 73 +-0.0 7 mPlan t p ower 243 +-6 MWUnit po wer + 0.7 MW

Guide v ane

Unit power (MW)

4 8 MW= Total intens ity

Power (MW)

Componentof thecavitationintensityinfluencedby a guidevane(% of thetotal)

Guide vane

Wicket gate cavitation characteristic


Cavitation mechanisms

In most cases, several different cavitation types appear ina turbine (leading-edge, trailing-edge, surface, etc.), and thesame type can be found in different places within the turbine.These cavitation occurances are referred to as cavitationmechanisms. A cavitation mechanism can be erosive or non-erosive.Each cavitation characteristic can be expressed for the totalcavitation (as was the case above) or for a single mechanism.In the case shown, three cavitation mechanisms were found.They are illustrated in the following.

Power (MW) Guide vane



Low-powermechanism

Power (MW) Guide vane



Basicmechanism

Power (MW)


Guide vane


High-powermechanism

38 40 42 44 46 48 50-20

0

20

40

60

80

100

120

140

Normalisat ionreference:Unit 5, 48 MW

Unit power (MW)

Tota

l cav

itatio

n in

tens

ity in

the

turb

ine

(nor

mal

ised

)

Meas ured Total Eros iveUnit 1Unit 5

Burfe ll Unit 1 (13-30 S ep tembe r 2003) & Un it 5 (21-30 Se pte mber

Unit power +-0 .15 MWPla nt power 243 +-6 MWHea dwater 244.73 +-0.07 mTailwate r 123.73 +-0.06 mNet he ad 116.5 +-0 .2 m

Power (MW)

Totalcavitationintensity(%)


Erosivemechanism

Global cavitation characteristic

Net head (m)

Discharge(m3/s)

Data quantity(descriptionquality) incommonmodel testsand inprototypemultidimen-sional vibro-acousticcavitation tests

Model tests vs. In-plant tests or monitoring

Net head (m)

Operatingrange

Discharge(m3/s)

Net head (m)

Model:only 2 pointsuseful inpractice

Prototype:detaileddescription

Discharge(m3/s)

Model tests vs. In-plant tests or monitoring

In a typical model cavitation test, much less useful data for practicaloperation of the prototype is obtained than can be obtained by meansof an in-plant multidimensional vibro-acoustic monitor or a test.

In some cases, not all types of cavitation can be seen in a modeltest. All can be heard and assessed in a good, multidimensionalin-plant vibro-acoustic test.

There are strong scale effects in incipient cavitation modelling.Thus, cavitation should be checked on the prototype.

Turbine cavitation performance varies in time, making continuouscontrol necessary.

Spatial resolution

Quite useful in making cavitation diagnosis is thedata on the spatial distribution of cavitation withina turbine. Such data is delivered by the multidimen-sional technology.

In what follows, a simple case of such a descriptionis illustrated: the distribution of the cavitationintensity over the angular segments in a large-diameter turbine with a horizontal-axis. Here,pressure differences in the upper and lower positionscause high variations in the cavitation intensity.

18 22 26 30 34 38MW

1

0.5

0

Angular segmentsaround the axis of ahorizontal-axis turbine

Normalisedcavitationintensity

18 22 26 30 34 38MW

1

0.5

0

Combined analysis:

Angular segments &Cavitation mechanisms

A cavitationmechanism


MW

1

0.5

018 22 26 30 34 38

Anothercavitationmechanism


MW

1

0.5

018 22 26 30 34 38


MW

1

0.5

018 22 26 30 34 38

Therefore, the spatialdistribution and the power-dependence of each of the4 cavitation mechanisms isshown here.


To conclude:

When applied to cavitation, Korto´s multidimensional technology

•••• identifies cavitation mechanisms, •••• assesses the role of turbine parts in cavitation, •••• yields data on the spatial distribution of cavitation in a turbine, and •••• delivers detailed turbine cavitation characteristics.

Further...

Even for simple quantities, such as the total cavitation intensity in a turbine, the multidimensional approach is needed. An estimate of the total intensity is derived through a spatial averaging.

Simpler monitoring algorithms use one or only a few sensors and deliver arbitrary data.

This is illustrated in the following.

How do the estimates of cavitation depend on the sensors´ location?

How many sensors are needed?

An example of a vertical-shaft turbine: the sensors on the casing, in 12 positions around the runner

spiralcasing

sensors´locations

Total cavitationintensity recordedby means of thesensor in a givenlocation,presented in apolar diagram

The estimate of the mean total cavitation intensity strongly depends on the position of the sensor with respect to the spiral casing.

The same is true inrespect to the formof the dependenceon the instantaneousrunner position.

A comparisonof the resultsobtained intwo positions,

and shows how greatthe differences inthe estimates ofthe mean intensitywould be if onlyone sensor in oneor anotherlocation were tobe used.

and acomparisonof

shows howdifferent theconclusionson the roleof the runnerblades andguide vaneswould be.

A comparisonof the resultsobtained intwo positions,

and shows how highthe differences inthe estimates ofthe mean intensitywould be if onlyone sensor in oneor anotherlocation would beused,

and

An example of a horizontal-shaft turbine shows the same:

The sensorsin 24circumfer-entiallocations;in each,the depend-ence on theinstan-taneousrunnerposition wasestimated.


The sensorsin 24circumfer-entiallocations;in each,the depend-ence on theinstan-taneousrunnerposition wasestimated.

Compare with


The sensorsin 24circumfer-entiallocations;in each,the depend-ence on theinstant-aneousrunnerposition wasestimated.

Compare withand with

In order to ensure a reliable cavitation sampling, - a rather higher number of sensors suitably distributed over the turbine, and - a suitable multidimensional algorithm for processing the data they deliver, are necessary.

The diagnosis or monitoring based on only one or only a few sensors can yield - erroneous estimates of cavitation intensity, and - false judgments of the role of turbine parts.

Our practice:- In the diagnostic tests, we use a high number of

sensors (possibly as many as 20-30 for cavitation).- For permanent monitoring, we reduce the

number to a minimum, based on the test results (typically 8, 6, or 4 for cavitation).

An example of a sensor set used on a bulb unit, in ageneral diagnostic test that included cavitation, ispresented on the next slide. For permanentcavitation monitoring, 8 were kept mounted.

Cavitation

How is the cavitation intensity calibrated?

What about estimates of the erosion rate in kilograms of the metal lost per unit time?

The answer follows in four steps.

Not all cavitation mechanisms are erosive.

The first step - to recognise different mechanisms - is made by the multidimensional vibro-acoustical method.

The second step - to recognise which of the mechanisms are erosive. Here, additional information outside the method is needed (model tests, repair experience).

The third step - to assess the relative erosion rate - is made by the method: For erosive mechanisms, the cavitation intensity estimates it yields are proportional to the erosion rate.

The fourth step - to calibrate these estimates into the absolute erosion rate - is most often not needed. For the majority of applications, such as operation optimisation, repair programming, etc., the relative estimates suffice.

Attempts were made to make this absolute calibration a priori; they are not well tested.

The most reliable is, however, an a posteriori approach: The monitor logs the accumulated cavitation intensity between two overhauls and compares it to the metal loss found.

Some others working in this field claimed to be able to recognise cavitation erosiveness vibro- acoustically. That proved to be false.

38 40 42 44 46 48 500

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Accumulated erosionP owe r s etting s tatis ticsEros ion rate density

Vibr

o-ac

oust

ical

est

imat

e of

kg

of e

lect

rode

s in

100

00 h

ours

Expe

cted

kg

of e

lect

rode

s in

100

00 h

ours

acc

ordn

ing

to w

hat w

as f

ound

in 1

3469

hou

rs

Power (MW)

Erosionratedensity(kg in10,000hoursper0.5 MWinterval)

1/4 xRelativetimespent in a0.5 MWinterval(%)

Powerstatistics

Erosion calibration: An example

Erosion rate density(with error bounds)

Accumulatederosionkg in 10,000 hours

Found

Predicted

Thank you for your attention.For further information please visitwww.korto.comor contact

[email protected]

[email protected]

[email protected]

[email protected]

luxembourg greetings from cavitation methodol… · • korto’s motto: stay with the problem...

Documents