Method for recognizing partial discharges emitted inside or outside an electrical apparatus

1. A progressive method of recognizing partial discharges emitted inside or outside electrical equipment, wherein this recognition of partial discharges is shape recognition of the partial discharge patterns wherein after recording known partial discharge patterns in a data base with each pattern being associated with the type of fault that generated it, and after detecting at least one partial discharge activity, said method comprises the following steps:

acquiring a color image by measuring a series of partial discharge patterns in phase and in amplitude, the series of partial discharge patterns resulting from the emitted partial discharges at the equipment

processing and transposing the color image into a black and white image;

acquiring a new partial discharge pattern from the black and white image;

sub-dividing the new partial discharge pattern into four equal portions by splitting the new partial discharge pattern in two along a horizontal axis, and in two along a vertical axis;

analyzing the four equal portions to allocate a category to the new partial discharge pattern;

comparing, as a function of the category, said new partial discharge pattern with partial discharge patterns that are already recorded in the data base;

diagnosing a source of said new partial discharge pattern and obtaining a matching fault in the equipment; and

recording the diagnosing and matching result obtained in the data base.

2. A method according to claim 1 , including an automatic correction step for adjusting parameters of the processing of the color image.

3. A method according to claim 1 , including a step of making the surface area of the new partial discharge pattern uniform.

4. A method according to claim 1 , wherein the step of comparing the new partial discharge pattern with partial discharge patterns comprises computing a match percentage between the new partial discharge pattern and the patterns that are already recorded in the data base, and classifying matches by decreasing percentage.

5. A method according to claim 1 , including the following steps:

entering the color image corresponding to the new pattern of said partial discharges;

processing and transposing the color image into the black and white image;

testing whether an image remains;

if no images remain:

adjusting parameters of the processing of the color image and returning to the start of the first step;

if an image remains:

checking to find out whether a size of the partial discharge pattern is greater than a predetermined threshold;

if the size is lower than the predetermined threshold:

expanding the partial discharge pattern along the ordinate axis;

if the size is greater than or equal to the predetermined threshold:

sub-dividing said partial discharge pattern into four zones of interest;

analyzing said zones of interest in order to classify the partial discharge

comparing the partial discharge pattern with corresponding partial discharge patterns in the data base;

listing the matching partial discharge patterns;

diagnosing the source of said partial discharge;

if diagnosis indicates an inside source of discharge then examining the equipment in order to repair said equipment and correlating the diagnosis; and

recording the results thus obtained in the data base.

6. A progressive method of recognizing partial discharges emitted inside or outside an electrical equipment, the method comprising recording known partial discharge patterns in a data base, each known partial discharge pattern being associated with a type of fault that generated it,

the method further comprising, upon detection of a partial discharge activity:

generating a color image by processing a series of partial discharge patterns in phase and in amplitude, the partial discharge patterns being measured with a measurement probe connected to the electrical equipment;

transposing the color image into a black and white image, the black and white image forming a new partial discharge pattern representative of the partial discharge activity;

sub-dividing the new partial discharge pattern into four equal portions, by splitting the new partial discharge pattern in two along a horizontal axis, and in two along a vertical axis;

analyzing the four equal portions to allocate a category to the new partial discharge pattern;

selecting, in the data base, known partial discharge patterns as a function of the category allocated to the new partial discharge pattern;

comparing the new partial discharge pattern with the selected known partial discharge patterns, thereby obtaining a result;

as a function of the result, retrieving, from the data base, the type of fault generating the partial discharge activity; and

recording the result in the data base.

7. The method according to claim 6 , further comprising automatically adjusting parameters for generating the color image.

8. The method according to claim 6 , including making a surface area of the new partial discharge pattern uniform.

9. The method according to claim 6 , wherein comparing the new partial discharge pattern with the selected known partial discharge patterns comprises computing a match percentage between the new partial discharge pattern and the selected known partial discharge patterns, and classifying matches by decreasing percentage.

10. The method according to claim 6 , further comprising, after transposing the color image into the black and white image, and before sub-dividing the new partial discharge pattern:

testing the black and white image to assess whether an image remains; and

if no image remains, adjusting parameters of the processing of the color image, then performing again generating the color image and transposing the color image into a black and white image to obtain the new partial discharge pattern;

if an image remains, comparing a size of the new partial discharge pattern to a predetermined threshold, and

if the size is lower than the predetermined threshold, expanding the new partial discharge pattern along an ordinate axis;

if the size is greater than or equal to the predetermined threshold, not expanding the new partial discharge pattern.

11. A progressive method of recognizing partial discharges emitted inside or outside an electrical equipment, the method comprising recording known partial discharge patterns in a data base, each known partial discharge pattern being associated with a type of fault that generated it,

the method further comprising, upon detection of a partial discharge activity, the steps of:

generating a color image by processing a series of partial discharge patterns in phase and in amplitude, the partial discharge patterns being measured with a measurement probe connected to the electrical equipment;

transposing the color image into a black and white image, the black and white image forming a new partial discharge pattern representative of the partial discharge activity;

testing the black and white image to assess whether an image remains; and

if no image remains, adjusting parameters of the processing of the color image, then performing again the steps of generating the color image and transposing the color image into a black and white image to obtain the new partial discharge pattern;

if an image remains, comparing a size of the new partial discharge pattern to a predetermined threshold, and

if the size is lower than the predetermined threshold, expanding the new partial discharge pattern along an ordinate axis;

if the size is greater than or equal to the predetermined threshold, not expanding the new partial discharge pattern;

sub-dividing the new partial discharge pattern into four equal portions, by splitting the new partial discharge pattern in two along a horizontal axis, and in two along a vertical axis;

analyzing the four equal portions to allocate a category to the new partial discharge pattern;

selecting, in the data base, known partial discharge patterns as a function of the category allocated to the new partial discharge pattern;

comparing the new partial discharge pattern with the selected known partial discharge patterns, thereby obtaining a result;

as a function of the result, retrieving, from the data base, the type of fault generating the partial discharge activity; and

recording the result in the data base.