Korto Cavitation Services




➔ Home Introduction Tests Monitoring Models Information Archive (Propellers) Contact
➔ PlantManagers TurbineManufacturers MonitoringManufacturers


Introduction to Vibro-Acoustical Diagnostics of Cavitation




• An introductory example

The relationship between the form of cavitation and its vibro-acoustical signature was tested on a model turbine in a test stand equipped with a vibro-acoustical sensor and a camera synchronised with the rotation. The cavitation intensity was described by the power of the noise received by the sensor.













In this particular case, three different cavitation mechanisms, active in different load ranges, were found. As shown below, they were vibro-acoustically distinguished and quantified.



• Consequences of the rotor-stator interaction
The wake in the flow behind the guide vanes (or combinations of the stay vanes and guide vanes, also influenced by other up-stream bodies in a turbine), causes modulation of the cavitation noise generated by the runner blades as received by the vibro-acoustical sensors. Analysis of these effects enables an assessment to be made of the intensity of the cavitation components related to each static/rotating pair. The graphs below illustrate this on a Kaplan turbine with four runner blades and 24 guide vanes, seven of which actively contribute to the result.


• Spatial and temporal distribution of cavitation intensity

The estimates of cavitation intensity derived from the signals acquired by the sensors in different angular positions around the runner differ substantially. They do so both in the mean value and in the form of temporal dependence, i.e. dependence on the runner´s instantaneous angular position. This is illustrated here on a bulb unit with the sensors on every other of the 24 guide vanes:


• Multidimensional method

Our multidimensional measurement scheme applied in on-site vibro-acoustical diagnostic tests of turbine cavitation consists of vibro-acoustical sensors on the trunnions of all the guide vanes and a sensor on the turbine shaft, and processing the signals they deliver in a combined manner, involving simultaneous analysis over three dimensions:

• spatial (sensors distributed circumferentially around the runner);
• temporal (following the signals in all runner´s instantaneous angular positions within a revolution);
• operational (doing so at a dense set of operation points where cavitation is expected and at some other points for assessing the non-cavitation background).

The measurements are synchronised with the turbine rotation, the results are averaged over revolutions, and the estimates of the cavitation intensity are expressed by a deterministic quantity additive for different cavitation mechanisms.

The sensor on the turbine shaft reacts with equal sensitivity to all cavitation segments which act on the runner and in its vicinity, while the sensors on the guide vanes allow for spatial resolution among the segments. Parallel use of both sensor locations results in a method possessing both these qualities.


The above graph illustrates the results of a multidimensional test on a 650-MW Francis unit. It shows, in grey scale, the cavitation intensity as acquired by the sensor on guide vane 13 within a time interval covering the passage of four guide vanes. The notation 13/2, 11/3, etc. follows a v/b-scheme presenting the cavitation components caused on runner blade b behind guide vane v. Due to the indivisibility of the numbers of guide vanes and runner blades, 32:13, and the fast reaction of the sensors and electronics, traces of cavitation stemming from all the active v/b-combinations are revealed and, as shown, the load-dependence of the related cavitation intensity is assessed. The three straight lines drawn above the 13/4-trace present the three groups of cavitation mechanisms acting in different load intervals in this particular turbine. Thus, here, there are up to 3×32×13 cavitation processes, which prove to be independent and thus their intensities assessed vibro-acoustically can be combined as additive. This is used in the three basic forms of the turbine cavitation characteristics assessed by the multidimensional method:

• I(Ο) mean global cavitation intensity;
• I(Ο,Θ) circumferential spatial distribution of the mean local cavitation intensity;
• I(Ο,Θ,Φ) time dependent quasi-instantaneous local circumferential spatial distribution, where:

Ο stands for a set of quantities which specify the operation,
Θ is the spatial variable, and
Φ is the runner´s instantaneous angular position which represents the time variable.
Both Θ and Φ are defined between 0° and 360°; Θ takes only the values corresponding to the guide-vane locations.

In addition to the cavitation-intensity functions as defined above, which are derived by means of the sensors on the guide vanes, the following two functions are derived using the sensor on the turbine shaft:

• J(Ο) mean global cavitation intensity, and
• J(Ο,Φ) time dependent quasi-instantaneous global cavitation intensity.

J(Ο) is related to I(Ο), but can differ in typical values and in the dependence on Ο. J(Ο,Φ) is related to the sum of I(Ο,Θ,Φ) over Θ.

The above cavitation descriptions are obtained by the multidimensional analysis/synthesis procedure in which all cavitation segments are equally taken into account. This guarantees the unbiased assessment of all of them. Rather popular simple cavitation assessment methods are not satisfactory. If they are based on one or only a few sensors, they may overestimate some cavitation segments and underestimate others (see "Spatial and temporal..." above) and thus deliver erroneous global assessments. If the sensors are not suitably selected and/or the signal processing in them is not well organised, they cannot deliver time-resolved assessments such as I(Ο,Θ,Φ).