A method for evaluating transformer operation health value

Abstract

The present invention proposes a transformer operation health value evaluation method, which includes the following steps: 1) selecting multiple sound features and determining a first sound feature threshold and a second sound feature threshold for each sound feature; 2) assigning a weight to each sound feature; 3) determining a transformer operation health index according to the sound feature weight and the first sound feature threshold and the second sound feature threshold, wherein the transformer operation health index includes a first health index threshold and a second health index threshold. The present invention can realize a quantitative evaluation of the transformer operation health value, and has the advantages of a simple evaluation method, easy programming implementation, and high reliability.

Description

Technical Field

[0001] The invention relates to a transformer operation health value evaluation method, belonging to the technical field of power safety monitoring. Background Art

[0002] The transformer is one of the most critical equipment in the power system, and its safe operation is of great significance to ensuring the power supply reliability of the power system; the transformer health index is an important parameter for evaluating the operating status of the transformer, which can realize the quantitative evaluation of the health status of the transformer and has objective and practical significance for the safe and stable operation of the transformer.

[0003] Sound is generated as the transformer operates. The amplitude, time domain waveform, and spectrum characteristics of the sound are closely related to its operating voltage, current, mechanical state, excitation state, insulation state, etc., and can promptly reflect changes in the equipment's operating status. Therefore, the sound characteristics can be used to evaluate the health of the transformer.

[0004] At present, the methods of using sound to diagnose transformer faults mainly include manual inspection methods or instrument and equipment detection methods; manual inspection methods are time-consuming and labor-intensive, prone to misjudgment, and the diagnosis results are closely related to the experience of technicians, and it is impossible to quantify the health value of the transformer operation; the existing methods based on instrument and equipment detection can only realize the diagnosis of different types of transformer defects. The technical solutions are complex and it is difficult to quantify the operating status, defect severity or health indicators of the transformer, which is not conducive to arranging equipment maintenance plans. Summary of the invention

[0005] The present invention proposes a transformer operation health value evaluation method, which aims to solve the problem that the prior art cannot perform health evaluation on the operation status of the transformer.

[0006] The technical solution of the present invention is a method for evaluating the operating health value of a transformer, the method comprising the following steps:

[0007] 1) Selecting a plurality of sound features and determining a first sound feature threshold and a second sound feature threshold for each sound feature;

[0008] 2) Assign weights to each sound feature;

[0009] 3) Determine a transformer operation health index according to the sound feature weight and the first sound feature threshold and the second sound feature threshold, wherein the transformer operation health index includes a first health index threshold and a second health index threshold.

[0010] Preferably, the transformer operation health value evaluation method further comprises the following steps:

[0011] 4) Build a sample library for transformer operation health indicators;

[0012] 5) Collect new transformer sound samples and calculate transformer operation health indicators based on sound feature weights and sound features.

[0013] Preferably, the transformer operation health value evaluation method further comprises the following steps:

[0014] 6) Determine the transformer health value based on the proportion of normal samples in the transformer operation health index sample library.

[0015] Preferably, in step 1), the first threshold of the sound characteristic determines the "caution" state of the transformer, and the second threshold of the sound characteristic determines the "fault" state of the transformer.

[0016] Preferably, the multiple sound features include a 100 Hz sound amplitude in a transformer sound signal, a 200 Hz-700 Hz sound amplitude, a 700 Hz-2 kHz sound amplitude, and a 2 kHz-40 kHz sound amplitude in a transformer sound signal.

[0017] Preferably, in the step 2), each sound feature is assigned a weight, specifically comprising: using an artificial neural network algorithm, taking the sound feature as input and the transformer health status as output, and obtaining the weights of the 100 Hz sound amplitude, the 200 Hz to 700 Hz sound amplitude, the 700 Hz to 2 kHz sound amplitude, and the 2 kHz to 40 kHz sound amplitude in the transformer sound signal through sample training.

[0018] Preferably, in step 3), the transformer operation health index is calculated according to formula (1): (1);

[0019] Where: H is the transformer operation health indicator, α i For the i The weight of the sound features, F i For the i The sound characteristics, n is the number of sound features, i =1,2,…… n .

[0020] Preferably, the first health indicator threshold of the transformer operation health indicator is calculated according to formula (2):

[0021] (2);

[0022] Where: H 1 is the first health indicator threshold of the transformer operation health indicator,V i For the i A sound feature first threshold value of a sound feature.

[0023] Preferably, the second health indicator threshold of the transformer operation health indicator is calculated according to formula (3):

[0024] (3);

[0025] Where: H 2 is the second health indicator threshold of the transformer operation health indicator, T i For the i A second threshold value of a sound feature of a sound feature.

[0026] Preferably, in step 6), the transformer operation health index sample library includes normal samples, attention samples, and abnormal samples, and the transformer health value is determined by the ratio of the number of normal samples to the total samples.

[0027] Beneficial effects of the present invention:

[0028] The present invention selects multiple sound features and determines a first threshold and a second threshold for each sound feature, assigns a weight to each sound feature, determines a first health indicator threshold and a second health indicator threshold of a transformer operation health index according to the sound feature weight and the sound feature first threshold and the sound feature second threshold, builds a sample library for the transformer operation health index, collects new transformer sound samples, calculates the transformer operation health index according to the sound feature weight and the sound feature, determines the transformer health value according to the proportion of normal samples in the transformer operation health index sample library, and can realize quantitative evaluation of the transformer operation health value, and has the advantages of simple evaluation method, easy programming implementation, and high reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Attached Figure 1 Schematic diagram of the basic process of an embodiment of the present invention. DETAILED DESCRIPTION

[0030] A transformer operation health value evaluation method, the method comprising the following steps:

[0031] 1) Selecting a plurality of sound features and determining a first sound feature threshold and a second sound feature threshold for each sound feature;

[0032] 2) Assign weights to each sound feature;

[0033] 3) determining a transformer operation health index according to a sound feature weight, a first sound feature threshold, and a second sound feature threshold, wherein the transformer operation health index includes a first health index threshold and a second health index threshold;

[0034] 4) Build a sample library for transformer operation health indicators;

[0035] 5) Collect new transformer sound samples and calculate transformer operation health indicators based on sound feature weights and sound features;

[0036] 6) Determine the transformer health value based on the proportion of normal samples in the transformer operation health index sample library.

[0037] In step 1), the first threshold of the sound characteristic determines the "caution" state of the transformer, and the second threshold of the sound characteristic determines the "fault" state of the transformer.

[0038] The multiple sound features include a 100 Hz sound amplitude in a transformer sound signal, a 200 Hz to 700 Hz sound amplitude, a 700 Hz to 2 kHz sound amplitude, and a 2 kHz to 40 kHz sound amplitude in a transformer sound signal.

[0039] In the step 2), each sound feature is assigned a weight, specifically comprising: using an artificial neural network algorithm, taking the sound feature as input and the transformer health status as output, and obtaining the weights of the 100 Hz sound amplitude, the 200 Hz to 700 Hz sound amplitude, the 700 Hz to 2 kHz sound amplitude, and the 2 kHz to 40 kHz sound amplitude in the transformer sound signal through sample training.

[0040] In step 3), the transformer operation health index is calculated according to formula (1):

[0041] (1);

[0042] Where: H is the transformer operation health indicator, α i For the i The weight of the sound features, F i For the i The sound characteristics, n is the number of sound features, i =1,2,…… n .

[0043] The first health indicator threshold of the transformer operation health indicator is calculated according to formula (2): (2);

[0044] Where: H1 is the first health indicator threshold of the transformer operation health indicator, V i For the i A sound feature first threshold value of a sound feature.

[0045] The second health indicator threshold of the transformer operation health indicator is calculated according to formula (3): (3);

[0046] Where: H 2 is the second health indicator threshold of the transformer operation health indicator, T i For the i A second threshold value of a sound feature of a sound feature.

[0047] In step 6), the transformer operation health index sample library includes normal samples, attention samples, and abnormal samples, and the transformer health value is determined by the ratio of the number of normal samples to the total samples.

[0048] Example 1

[0049] like Figure 1 As shown, the steps of the transformer operation health value evaluation method based on sound feature weights in this embodiment include:

[0050] 1) Selecting a plurality of sound features and determining a first sound feature threshold and a second sound feature threshold for each sound feature;

[0051] 2) Assign weights to each sound feature;

[0052] 3) determining a first health indicator threshold and a second health indicator threshold of the transformer operation health indicator according to the sound feature weight and the sound feature first threshold and the sound feature second threshold of the sound feature;

[0053] 4) Build a sample library for transformer operation health indicators;

[0054] 5) Collect new transformer sound samples and calculate the transformer operation health index based on the sound feature weight and sound feature; the purpose of this step is to increase the number of samples of transformer operation health index in the sample library so as to calculate the proportion of normal samples, attention samples and abnormal samples;

[0055] 6) Determine the transformer health value based on the proportion of normal samples in the transformer operation health index sample library.

[0056] In this embodiment, four transformer sound characteristics are selected, namely, 100Hz sound amplitude, 200Hz-700Hz sound amplitude, 700Hz-2kHz sound amplitude, 2kHz-40kHz sound amplitude in the transformer sound signal; the 100Hz sound amplitude is the sound characteristic F 1 , sound characteristics F 1 The first threshold of the sound feature V 1 is 0.6, sound characteristics F 1 The second threshold of the sound feature T 1 is 0.8; the 200Hz~700Hz sound amplitude is the sound characteristic F 2 , sound characteristics F 2 The first threshold of the sound feature V 2 is 0.3, sound characteristics F 2 The second threshold of the sound feature T 2 is 0.5; the 700Hz~2kHz sound amplitude is the sound characteristic F 3 , sound characteristics F 3 The first threshold of the sound feature V 3 is 0.5, sound characteristics F 3 The second threshold of the sound feature T 3 is 0.8;

[0057] The 2kHz to 40kHz sound amplitude is the sound characteristic F 4 , sound characteristics F 4 The first threshold of the sound feature V 4 is 0.2, sound characteristics F 4 The second threshold of the sound feature T 4 is 0.75.

[0058] In this embodiment, the first threshold of the sound characteristic in step 1) determines the "caution" state of the transformer, and the second threshold of the sound characteristic determines the "fault" state of the transformer.

[0059] In the step 2), each sound feature is assigned a weight, specifically comprising: using an artificial neural network algorithm, taking the sound feature as input and the transformer health status as output, obtaining the weights of the 100Hz sound amplitude, the 200Hz-700Hz sound amplitude, the 700Hz-2kHz sound amplitude, and the 2kHz-40kHz sound amplitude in the transformer sound signal through sample training; the sound feature in this embodiment F 1 Weight α 1 is 0.4; sound characteristics F 2 Weight α 2 is 0.4; sound characteristics F 3 Weight α 3 is 0.1; sound characteristics F 4 Weight α 4 is 0.1; the sum of the weights of each sound feature is 1.

[0060] In this embodiment, the transformer operation health index in step 3) is calculated according to formula (1): (1);

[0061] Where: H is the transformer operation health indicator, α i For the i The weight of the sound features, F i For the i The sound characteristics, n is the number of sound features, i =1,2,…… n ;

[0062] Correspondingly, the first health indicator threshold of the transformer operation health indicator is calculated according to formula (2):

[0063] (2);

[0064] Where: H 1 is the first health indicator threshold of the transformer operation health indicator, V i For the i The first threshold of the sound feature of the sound feature, i =1, 2, 3, 4;

[0065] Correspondingly, the second health indicator threshold of the transformer operation health indicator is calculated according to formula (3):

[0066] (3);

[0067] Where: H 2 is the second health indicator threshold of the transformer operation health indicator, T i For the i The second threshold value of the sound feature of the sound feature, i =1, 2, 3, 4.

[0068] According to the above calculation method, the first health indicator threshold of the transformer operation health index is determined to be 0.43 and the second health indicator threshold is 0.675 according to the sound feature weight and the first sound feature threshold and the second sound feature threshold of the sound feature.

[0069] In step 4) of this embodiment, a sample library is constructed for the transformer operation health index, which specifically includes: obtaining a transformer operation health index sample library by continuously detecting the sound of a transformer and calculating the health index; in the transformer operation health index sample library, a health index lower than a first health index threshold of the transformer operation health index is a normal sample, a health index greater than or equal to the first health index threshold of the transformer operation health index and less than a second health index threshold of the transformer operation health index is a caution sample, and a health index greater than or equal to the second health index threshold of the transformer operation health index is an abnormal sample; in this embodiment, the transformer operation health index sample library contains 99 sample data, including 60 normal sample data, 30 caution sample data, and 9 abnormal sample data.

[0070] In this embodiment, step 5) collects new transformer sound samples, specifically including: calculating the transformer operation health index according to the sound feature weight and the sound feature, and calculating the new transformer operation health index as 0.45 according to the sound feature weight and the sound feature, which is greater than the first health index threshold of the transformer operation health index and less than the second health index threshold of the transformer operation health index, corresponding to the attention sample; after the new transformer operation health index is added to the transformer operation health index sample library, the number of samples in the transformer operation health index sample library reaches 100; among them, there are 60 normal sample data, 31 attention sample data, and 9 abnormal sample data.

[0071] In this embodiment, step 6) determines the transformer health value according to the proportion of normal samples in the transformer operation health index sample library, specifically including step 6) that the transformer operation health index sample library contains 60 normal samples, 31 attention samples, and 9 abnormal samples, and the proportion of normal samples to the total number of samples is 0.6, therefore, the transformer health value is determined to be 0.6; generally speaking, a transformer health value exceeding 0.9 is considered to be in a healthy state, a value below 0.7 is considered to be in an abnormal state, and the rest are in a sub-healthy state; in this embodiment, the transformer is in an abnormal state, and it is necessary to strengthen the monitoring of the transformer. If the health value continues to decline, it is necessary to conduct an on-site inspection of the transformer to determine the cause of the fault and propose maintenance measures.

[0072] The above is only a preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A transformer operation health value evaluation method, characterized in that The following steps are involved: Step 1) selecting a plurality of sound features and determining a first sound feature threshold and a second sound feature threshold for each sound feature; Step 2) assign weights to each sound feature; Step 3) determining a transformer operation health index according to the sound feature weight and the first sound feature threshold and the second sound feature threshold, wherein the transformer operation health index includes a first health index threshold and a second health index threshold; Step 4) Build a sample library for transformer operation health indicators; Step 5) Collect new transformer sound samples and calculate transformer operation health indicators based on sound feature weights and sound features; Step 6) determining the transformer health value according to the proportion of normal samples in the transformer operation health index sample library; The multiple sound features include a 100 Hz sound amplitude in a transformer sound signal, a 200 Hz to 700 Hz sound amplitude, a 700 Hz to 2 kHz sound amplitude, and a 2 kHz to 40 kHz sound amplitude in a transformer sound signal; In step 3), the transformer operation health index is calculated according to formula (1): （1）； Where: H is the transformer operation health indicator, α i For the i The weight of the sound features, F i For the i The sound characteristics, n is the number of sound features, i =1,2,…… n ; The first health indicator threshold is calculated according to formula (2): （2）； Where: H 1 is the first health indicator threshold of the transformer operation health indicator, V i For the i The first threshold of the sound feature of the sound feature, n is the number of sound features, i =1,2,…… n ; The second health indicator threshold is calculated according to formula (3): （3）； Where: H 2 is the second health indicator threshold of the transformer operation health indicator, T i For the i The second threshold value of the sound feature of the sound feature, n is the number of sound features, i =1,2,…… n。 2. A transformer operation health value evaluation method according to claim 1, characterized in that The first threshold of the sound characteristic in step 1) determines the "attention" state of the transformer. The second threshold of the sound signature determines the "fault" state of the transformer.

3. The transformer operation health value evaluation method according to claim 1 is characterized in that In the step 2), each sound feature is assigned a weight, specifically comprising: using an artificial neural network algorithm, taking the sound feature as input and the transformer health status as output, and obtaining the weights of the 100 Hz sound amplitude, the 200 Hz to 700 Hz sound amplitude, the 700 Hz to 2 kHz sound amplitude, and the 2 kHz to 40 kHz sound amplitude in the transformer sound signal through sample training.

4. A transformer operation health value evaluation method according to claim 1, characterized in that In step 6), the transformer operation health index sample library includes normal samples, attention samples, and abnormal samples, and the transformer health value is determined by the ratio of the number of normal samples to the total samples.

Images