Method for detecting an incipient damage in a bearing

Abstract

Provided is a method for detecting an incipient damage in a bearing including receiving from a plurality of sensors of the bearing a plurality of signals corresponding to a plurality of bursts, each burst being emitted each time an indentation is met during the rotation of the bearing at a contact point between a roller of the bearing and a race of the bearing, analyzing the signals for locating on the bearing the indentations corresponding to the bursts, generating a map of points on a digital model of the bearing corresponding to the indentations on the bearing, the map of points being registered and updated throughout at least an interval of the lifetime of the bearing, identifying an incipient damage in the bearing each time the distribution of the points in the map is not random, and generating an incipient damage warning message.

Claims

1. A computer implemented method for detecting an incipient damage in a bearing, the method comprising: receiving from a plurality of sensors of the bearing a plurality of signals corresponding to a plurality of bursts, each burst being emitted each time an indentation is met during a rotation of the bearing at a contact point between a roller of the bearing and a race of the bearing; analyzing the plurality of signals for locating on the bearing a plurality of indentations corresponding to the plurality of bursts; generating a map of points on a digital model of the bearing corresponding to the plurality of indentations on the bearing, the map of points being registered and updated throughout at least an interval of a lifetime of the bearing; identifying the incipient damage in the bearing each time a distribution of the points in the map is not random; and generating an incipient damage warning message.

2. The computer implemented method according to claim 1, wherein the incipient damage in the bearing is identified each time at least a portion of of points in the map are grouped in a cluster.

3. The computer implemented method according to claim 2, wherein the map of points is subdivided in a plurality of areas and wherein the incipient damage in the bearing is identified each time at least a cluster of points is present in at least one of the plurality of areas.

4. The computer implemented method according to claim 1, wherein the locating the plurality of indentations includes measuring a periodicity of the plurality of bursts for determining if each indentation is present on a roller or on an inner race or on an outer race of the bearing.

5. The computer implemented method according to claim 1, wherein the locating the plurality of indentations includes measuring a time delay between at least two signals of the plurality of signals.

6. The computer implemented method according to claim 1, wherein the locating the plurality of indentations on a roller or on an outer race of the bearing includes a Kalman filter.

7. The computer implemented method according to claim 1, further comprising generating a lubrication contamination message when the distribution of the points in the map is random.

8. The computer implemented method according to claim 1, wherein the method is performed periodically.

Description

BRIEF DESCRIPTION

(1) Some of the embodiments will be described in detail, with reference to the following figures, wherein like designations denote like members, wherein:

(2) FIG. 1 shows a partial longitudinal section view of a wind turbine comprising a plurality of bearings, whose failures may be detected according to the method of the present invention.

(3) FIG. 2 shows a bearing to which the method of the present invention is applied,

(4) FIG. 3 shows a digital model of the bearing used in the method of the present invention,

(5) FIG. 4 shows a block diagram, illustrating the method of the present invention.

DETAILED DESCRIPTION

(6) FIG. 1 shows a wind turbine 1. The wind turbine 1 comprises a tower 2, which is mounted on a non-depicted fundament. A nacelle 3 is arranged on top of the tower 2.

(7) The wind turbine 1 further comprises a wind rotor 5 having two, three or more blades 4 (in the perspective of FIG. 1 only two blades 4 are visible). The wind rotor 5 is rotatable around a rotational longitudinal axis Y. The blades 4 extend radially with respect to the rotational axis Y.

(8) The wind turbine 1 comprises a permanent magnet electric generator 11 including a stator and a rotor.

(9) According to other possible embodiments of the present invention (not represented in the attached figures), the present invention may be applied to any other type of permanent magnet machine with either internal or external rotor.

(10) The wind rotor 5 is rotationally coupled with the permanent magnet generator 11 either directly, e.g. direct drive or by means of a rotatable main shaft 9 and through a gear box (not shown in FIG. 1). Two schematically depicted bearings 10 are provided in order to hold in place the main shaft 9 and the rotor 5. The rotatable main shaft 9 extends along the rotational axis Y.

(11) The method for detecting an incipient damage in a bearing may be applied to the bearings 10 of the the main shaft 9 of the wind turbine 1 or to any other bearing, whose monitoring is required.

(12) FIG. 2 shows a schematic representation of a bearing 10, to which the method of the present invention is applied.

(13) The bearing 10 includes a plurality of rollers 20 between an inner race 15 and an outer race 16. The rollers 20 rotate around respective central axis and also around the rotational longitudinal axis Y for allowing a relative rotation between the fixed inner race 15 and the outer race 16, which rotates around the rotational longitudinal axis Y.

(14) According to another possible embodiment of the present invention, the inner race 15 rotates around the rotational longitudinal axis Y and the outer race 16 is fixed.

(15) When a particle is crushed or rolled in the bearing 10, a small indentation 30 is created in the bearing 10 on the rollers 20 and/or on the inner race 15 and/or on the outer race 16. When the bearing is rotating, a strong so-called after effect Acoustic Emission (AE) burst is emitted every time the roller is passing the indentation. Over time the burst will reduce in strength and eventually disappear. The AE burst permits to identify the indentations 30. For detecting the AE bursts, a plurality of sensors (not represented in the attached figures) are mounted on the bearing 10. In FIG. 2 a plurality of indentations 30 on the external race 16 are shown.

(16) With reference to FIG. 3 and FIG. 4, the steps of a computer implemented method 200 for detecting an incipient damage in the bearing 10 are described. The method 200 comprises a first step 210 of receiving from the plurality of sensors of the bearing 10 a plurality of signals corresponding to a plurality of AE bursts. Each burst is emitted when the bearing rotates each time an indentation 30 is met during the rotation of the bearing 10 at a contact point between a roller 20 and the inner race 15 or at a contact point between a roller 20 and the outer race 16.

(17) After the first step 210, the method 200 comprises a second step 220 of analysing the plurality of signals from the sensors for locating on the bearing 10 the plurality of indentations 30 corresponding to the plurality of bursts.

(18) The indentation may be present on one or more rollers 20 and/or on the inner race 15 and/or on the outer race 16. Determining if each indentation is present on a roller 20 or on the inner race 15 or on an outer race 16 can be determining by measuring the periodicity of the plurality of bursts in each signal provided by sensors. The periodicity of the bursts corresponding to the same indentation 110, present on the inner race 15 (which is fixed) or on the rollers 20 or on the outer race 16 (which rotates, but with respective different rotational speeds) is predetermined by the bearing geometrics and is proportional to the rotational speed.

(19) Locating the plurality of indentations 30 includes measuring the time delay between at least two signals of the plurality of signals measured through the sensors. Each signal provided by a respective sensor may include a plurality of peaks corresponding to the plurality of burst caused by the plurality of indentations 30. The analysis of the delay between two peaks on two different signals, the two peaks being both associated to the same burst and the same indentation 30, may be used to identify the distance of the indentation 30 from the sensors, i.e. the position of the indentation on the bearing 10.

(20) Considering that the rollers 20 and the outer race 16 rotate, locating the plurality of indentations 30 on such elements can be efficiently performed by using a Kalman filter.

(21) After the second step 220, the method 200 comprises a third step 230 of generating a map of points 110 on a digital model 100 of the bearing 10, each point 110 corresponding to an indentation 30 on the bearing 10. The map of points 110 is registered and updated throughout at least an interval of the lifetime of the bearing, so that the map includes all historical points 110 identified during such time interval, each point 110 corresponding to an indentation 30 that is or previously has been detected on the bearing 10.

(22) FIG. 3 shows a partial representation of the digital model 100 including only a digital representation of the outer race 16 of FIG. 2, where a plurality of historical detected indentations 30 are present.

(23) The map of points 110 is subdivided in a plurality of areas (four areas 150a-d are shown in FIG. 3)

(24) After the third step 230, the method 200 comprises a fourth step 240 of checking if the distribution of the points 110 in the map is random.

(25) In particular, the fourth step of the method may include checking if the distribution of the points 110 is randomly spread across the areas 150a-d comprising the map of the points 110.

(26) If the distribution of the points 110 in the map is not random, the method continues with a fifth step 250 of identifying an incipient damage in the bearing 10 and generating an incipient damage warning.

(27) Lack of randomness is identified by a cluster 120 of points 110 present in the same area 150a of the map of point 110 shown on FIG. 3.

(28) Alternatively, to the fifth step 250, if the distribution of the points 110 in the map is random, the method continues with a sixth step 260 of generating a lubrication contamination message.

(29) If no indentations are detected or the rate/time delay between the detected indentations is long, the method could further provide a message to make it evident that no critical conditions are being reached.

(30) Although the present invention has been disclosed in the form of preferred embodiments and variations thereon, it will be understood that numerous additional modifications and variations could be made thereto without departing from the scope of the invention.

(31) For the sake of clarity, it is to be understood that the use of “a” or “an” throughout this application does not exclude a plurality, and “comprising” does not exclude other steps or elements.