luxembourg greetings from cavitation methodol… · • korto’s motto: stay with the problem...
TRANSCRIPT
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
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
00,7 0,8 0,9 1,0
.
2
3
1
Nor-malisednoisepower
Normalised turbine power
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
Normalised turbine power
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:
Normalisednoise power
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:
Normalisednoise power
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
70 MW
0 MW
70 MW
0 MW
70 MW
0 MW
70 MW
0 MW
70 MW
0 MW
70 MW
0 MW
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
All cavitationmechanismsincluded
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
All cavitationmechanismsincluded
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
Componentof thecavitationintensityinfluencedby a guidevane(% of thetotal)
Wicket gate cavitation characteristic
Low-powermechanism
Power (MW) Guide vane
Componentof thecavitationintensityinfluencedby a guidevane(% of thetotal)
Wicket gate cavitation characteristic
Basicmechanism
Power (MW)
Componentof thecavitationintensityinfluencedby a guidevane(% of thetotal)
Guide vane
Wicket gate cavitation characteristic
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(%)
All cavitationmechanismsincluded
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
Normalisedcavitationintensity
MW
1
0.5
018 22 26 30 34 38
Anothercavitationmechanism
Normalisedcavitationintensity
MW
1
0.5
018 22 26 30 34 38
Normalisedcavitationintensity
MW
1
0.5
018 22 26 30 34 38
Normalisedcavitationintensity
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.
Normalisedcavitationintensity
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 multi- dimensional 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.
An example of a horizontal-shaft turbine shows the same:
The sensorsin 24circumfer-entiallocations;in each,the depend-ence on theinstan-taneousrunnerposition wasestimated.
Compare with
An example of a horizontal-shaft turbine shows the same:
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