Method of detecting and locating a source of partial discharge in an electrical apparatus

1. A method of detecting and localizing a source of partial discharge in an electrical apparatus placed in an enclosure including an acoustically-conductive fluid, the method including:

positioning against the outside of the enclosure N acoustic detectors configured to detect an acoustic wave emitted by the source of partial discharge, and acquiring positions of the N acoustic detectors;

defining P groups of acoustic detectors from the N acoustic detectors with P=Z!/3!(Z−3)!, each of the P groups comprising at least three detectors, and Z being equal to N or to N−1;

performing a first check to verify whether each group is free from any triangle having vertices embodied by three acoustic detectors in which at least one angle lies outside a range of angles, and otherwise eliminating the group;

first calculating an approximate position for the source of partial discharge, for each group that is not eliminated after the performing the first check, from the positions of the acoustic detectors of the group and the instants the acoustic wave emitted by the source of partial discharge is received by each of the detectors of the group;

performing a second check to verify whether that each group that is not eliminated after the performing the first check is free from any triangle having vertices embodied by two of the detectors of the group and by the approximate position of the source of partial discharge with at least one angle lying outside the range, and otherwise eliminating the group; and

second calculating the looked-for position of the source of partial discharge by averaging the approximate positions calculated for the at least one group that are not eliminated after the second check.

2. The method according to claim 1 , wherein the range of angles extends from 30° to 151° (end values included).

3. The method according to claim 1 or claim 2 , the method further comprising:

acquiring the dimensions of the enclosure; and

performing a third check to verify whether the approximate position of the source of partial discharge, as calculated for the at least one group that is not eliminated after the second check, lies inside the enclosure of the apparatus, and otherwise eliminating the group.

4. The method according to claim 3 , wherein the performing the third check is performed with a tolerance.

5. The method according to claim 3 , wherein an average is computed using the approximate position calculated for at least one group that is not eliminated after the performing the third check.

6. The method according to claim 1 , wherein the method further includes:

performing a fourth check to eliminate at least one acoustic detector that has led to at least one approximate position considered deviant.

7. The method according to claim 6 , wherein the fourth check includes:

first identifying at least one extreme approximate position from the approximate positions for the source of partial discharge as calculated for at least one group that is not eliminated after the fourth check;

second identifying at least one acoustic detector at the origin of the at least one extreme approximate position;

eliminating any group of acoustic detectors that include the identified acoustic detector; and

repeating said first and second identifying so long as a spherical error probability (SEP) parameter calculated from the at least one approximate position for the source of partial discharge from the non-eliminated groups remains greater than a threshold value.

8. The method according to claim 7 , wherein the SEP parameter is defined by SEP=0.513 (σ x +σ y +σ z ), and

σ x is the standard deviation along the x-axis for the approximate positions of the source of partial discharge,

σ y is the standard deviation along the y-axis of the approximate positions for the source of partial discharge, and

σ z is the standard deviation along the z-axis of the approximate positions for the source of partial discharge.

9. The method according to claim 6 , wherein the average is computed using the approximate positions calculated for all groups that are not eliminated after the fourth check.

10. The method according to claim 1 , wherein the method further includes:

applying M acoustic detectors to the outside wall of the enclosure of an electrical apparatus, and

selecting amongst the M acoustic detectors, N acoustic detectors suitable for detecting an acoustic wave emitted by the source of partial discharge.

11. The method according to claim 1 , wherein when Z=N, a group is formed from a triplet of acoustic detectors, with the approximate positions being calculated by triangulation, by solving the following equations:

( x Gi −xi ) 2 +( y Gi −yi ) 2 +( Z Gi −zi ) 2 −v 2 ( ti−t 0) 2 =0

( x Gi −xj ) 2 +( y Gi −yj ) 2 +( z Gi −zj ) 2 −v 2 ( tj−t 0) 2 =0

( x Gi −xk ) 2 +( y Gi −yk ) 2+( z Gi −zk ) 2 −v 2 ( tk−t 0) 2 =0, wherein

X Gi , y Gi , Z Gi represent the approximate position of the source of partial discharge,

(xi, yi, zi), (xj, yj, zj), (xk, yk, zk) represent the positions of the acoustic detectors of the triplet,

t 0 represents an instant at which a partial discharge generated by the source is emitted, and

ti, tj, tk represent the instants at which the acoustic wave emitted by the partial discharge is picked up by each of the acoustic detectors of the triplet, and v represents the speed of the acoustic wave in the fluid.

12. The method according to claim 11 , wherein the instants of the emission of the partial discharge is measured by electrical means which co-operates with the enclosure.

13. The method according to claim 1 , wherein Z=N−1, a group is formed by a quadruplet of acoustic detectors including one acoustic detector that acts as a time reference, the approximate positions being calculated by triangulation by solving the following equations:

√{square root over (( x Gj −xj ) 2 +( y Gj −yj ) 2 +( z Gj −zj ) 2 )}{square root over (( x Gj −xj ) 2 +( y Gj −yj ) 2 +( z Gj −zj ) 2 )}{square root over (( x Gj −xj ) 2 +( y Gj −yj ) 2 +( z Gj −zj ) 2 )} −√{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )} − v ( tj−ti )=0

√{square root over (( x Gj −xk ) 2 +( y Gj −yk ) 2 +( z Gj −zk ) 2 )}{square root over (( x Gj −xk ) 2 +( y Gj −yk ) 2 +( z Gj −zk ) 2 )}{square root over (( x Gj −xk ) 2 +( y Gj −yk ) 2 +( z Gj −zk ) 2 )} −√{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )} − v ( tk−ti )=0

√{square root over (( x Gj −xl ) 2 +( y Gj −yl ) 2 +( z Gj −zl ) 2 )}{square root over (( x Gj −xl ) 2 +( y Gj −yl ) 2 +( z Gj −zl ) 2 )}{square root over (( x Gj −xl ) 2 +( y Gj −yl ) 2 +( z Gj −zl ) 2 )} −√{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )}{square root over (( x Gj −xi ) 2 +( y Gj −yi ) 2 +( z Gj −zi ) 2 )} − v ( tl−ti )=0, wherein

x Gj , y Gj , z Gj represent the approximate position of the source of partial discharge,

(xi, yi, zi), (xj, yj, zj), (xk, yk, zk), (xl, yl, zl) represent the positions of the four acoustic detectors of the quadruplets,

ti, tj, tk, tl represent the instants at which the acoustic wave emitted by the partial discharge is received by each of the acoustic detectors of the quadruplet, and

v is the speed of the acoustic wave in the fluid.

14. The method according to claim 11 , wherein the electrical means is an antenna detector.