Sensitive parameter extraction method for fault diagnosis of dry-type air-core reactor

Abstract

The invention discloses a sensitive parameter extraction method for fault diagnosis of a dry-type air-core reactor, which belongs to the technical field of fault diagnosis of power equipment, and comprises the following steps of: counting parameters for fault diagnosis of the dry-type air-core reactor, and calculating a parameter change rate; calculating a parameter sensitivity basic score; calculating a parameter sensitivity additional coefficient; calculating a parameter sensitivity comprehensive score; sensitive parameters are extracted, and the parameter with the highest parameter sensitivity comprehensive score serves as the sensitive parameter; according to the method, the parameter sensitivity and the parameter monitoring difficulty are comprehensively considered, and different severity degrees of the turn-to-turn insulation breakdown fault of the dry-type air-core shunt reactor are considered, so that the extracted sensitive parameters have the characteristics of high sensitivity and easy monitoring; and the feasibility and accuracy of turn-to-turn insulation breakdown fault identification of the dry-type hollow shunt reactor are improved.

Description

Technical Field

[0001] The invention relates to the technical field of power equipment fault diagnosis, in particular to a sensitive parameter extraction method for dry-type air-core reactor fault diagnosis. Background Art

[0002] The rapid development of power grid construction presents new challenges for long-distance, high-voltage, and large-capacity power transmission. However, reactive power compensation and power-frequency overvoltage issues are becoming increasingly prominent in power grids. Shunt reactors can effectively address these issues. Dry-type air-core shunt reactors, with their simple structure, ease of maintenance, low cost, and lack of oil or iron cores, have become indispensable reactive power compensation devices in substations. However, in actual operation, dry-type air-core reactors are prone to insulation damage due to factors such as localized overheating and moisture in the insulation layer, leading to interturn insulation breakdown and an increasing failure rate.

[0003] Dry-type air-core shunt reactors are classified according to their severity, ranging from mild to severe, into inter-turn short circuits, single-turn short circuits, and multi-turn short circuits. To accurately determine the severity of the fault, sensitive parameters must be extracted as a basis for judgment. However, the sensitivity of these parameters varies with fault severity. If only the variation characteristics of a single parameter at a specific fault severity are considered, without considering the overall variation in parameter sensitivity across different fault severity levels, it will be difficult to accurately identify reactor inter-turn insulation breakdown faults, making it impossible to promptly eliminate the fault, resulting in a high failure rate for dry-type air-core reactors. Summary of the Invention

[0004] In order to solve the problem that existing diagnostic methods are difficult to accurately identify inter-turn insulation breakdown faults of dry-type air-core shunt reactors, the purpose of the present invention is to provide a sensitive parameter extraction method for dry-type air-core reactor fault diagnosis.

[0005] To achieve the above-mentioned purpose, the present invention is implemented through the following technical solutions:

[0006] A sensitive parameter extraction method for dry-type air-core reactor fault diagnosis includes the following steps:

[0007] ① Count the parameters used for dry-type air-core reactor fault diagnosis and calculate the parameter change rate;

[0008] ②Calculate the basic score of parameter sensitivity;

[0009] ③Calculate the additional coefficient of parameter sensitivity;

[0010] ④ Calculate the comprehensive score of parameter sensitivity;

[0011] ⑤ Extract sensitive parameters and take the parameter with the highest comprehensive sensitivity score as the sensitive parameter.

[0012] Preferably, in step ①, the parameters used for dry-type air-core reactor fault diagnosis include total loop current, equivalent resistance, equivalent reactance, equivalent impedance, power factor angle, power factor, active power loss, spatial magnetic induction intensity and force on adjacent turns of the winding fault position.

[0013] Preferably, the parameter change rate in step ① is calculated as follows:

[0014] △Y ij =|Y ij -Y ij0 | / Y ij0 ×100%;

[0015] Among them, △Y ij is the parameter change rate, Y ij Y is the parameter value when the reactor has inter-turn insulation breakdown fault, ij0 It is the parameter value of the reactor during normal operation;

[0016] Preferably, the basic score of parameter sensitivity in step ② is based on: ranking from large to small according to the rate of change, the first place scores x1, the second place scores x2, the third place scores x3, the fourth place scores x4, the fifth place scores x5, the sixth place scores x6, the seventh place scores x7, the eighth place scores x8, and the ninth place scores x9.

[0017] Preferably, in step ②, the score value is determined based on:

[0018] |x k+1 -x k | / x k ×100%≥10%;

[0019] Among them, k = 1, 2, 3, 4, 5, 6, 7, 8, 9.

[0020] Preferably, in step ②, when the parameter sensitivity basic scores are tied, the mean parameter change rate under the three fault types is introduced as an auxiliary criterion; the specific method is: calculate the arithmetic mean of the parameter change rate under the three fault situations of short circuit between different turns, short circuit between single turns and short circuit between multiple turns respectively, and re-arrange the parameters that are tied from large to small according to their mean change rate, and the one with a higher mean change rate has a higher sensitivity basic score.

[0021] Preferably, in step ③, the parameter sensitivity additional coefficient is based on: according to the difficulty of parameter monitoring, the scores are m1, m2, m3, m4, m5, m6, m7, m8, and m9 from easiest to most difficult.

[0022] Preferably, in step ③, factors affecting the difficulty of parameter monitoring include environment, cost, difficulty of installing the monitoring sensor and monitoring error.

[0023] Preferably, in step ③, the parameter sensitivity additional coefficient value is determined based on:

[0024] 10%≤|m k+1 -m k | / m k ×100%≤|x k+1 -x k | / x k ×100%;

[0025] Among them, k = 1, 2, 3, 4, 5, 6, 7, 8, 9.

[0026] Preferably, in step ④, the parameter sensitivity comprehensive score is:

[0027]

[0028] Among them, g i is the comprehensive sensitivity score of each parameter, m Yij is the parameter sensitivity additional coefficient corresponding to each parameter, x Yij It is the parameter sensitivity basic score corresponding to each parameter.

[0029] Compared with the prior art, the present invention has the following advantages:

[0030] The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis of the present invention comprehensively considers parameter sensitivity and parameter monitoring difficulty, so that the extracted sensitive parameters have the characteristics of high sensitivity and easy monitoring, thereby improving the feasibility of identifying inter-turn insulation breakdown faults of dry-type air-core shunt reactors.

[0031] The present invention comprehensively considers different severity levels of dry-type hollow shunt reactor interturn insulation breakdown faults, and the extracted sensitive parameters can be applicable to different development stages of the fault, thereby improving the accuracy of dry-type hollow shunt reactor interturn insulation breakdown fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Flowchart of the sensitive parameter extraction method for dry-type air-core reactor fault diagnosis. DETAILED DESCRIPTION

[0033] The object of the present invention is to provide a sensitive parameter extraction method for dry-type air-core reactor fault diagnosis. The present invention will be further described below in conjunction with specific embodiments.

[0034] Example 1

[0035] A sensitive parameter extraction method for dry-type air-core reactor fault diagnosis, such as Figure 1 As shown, the following steps are included:

[0036] Step 1: Calculate the parameter change rate △Y ij

[0037] △Y ij =|Y ij -Y ij0 | / Y ij0 ×100%;

[0038] Among them, △Y ij is the parameter change rate, Y ij Y is the parameter value when the reactor has inter-turn insulation breakdown fault, ij0 It is the parameter value of the reactor during normal operation;

[0039] When j takes different values, it represents different parameters, as follows: j = 1 means the parameter is the total loop current, j = 2 means the parameter is the equivalent resistance, j = 3 means the parameter is the equivalent reactance, j = 4 means the parameter is the equivalent impedance, j = 5 means the parameter is the power factor angle, j = 6 means the parameter is the power factor, j = 7 means the parameter is the active power loss, j = 8 means the parameter is the spatial magnetic induction intensity, and j = 9 means the parameter is the force on the adjacent turns of the group fault position;

[0040] Let i = 1, which represents the change rate of each parameter when the reactor has an inter-turn short-circuit fault; i = 2, which represents the change rate of each parameter when the reactor has a single-turn short-circuit fault; i = 3, which represents the change rate of each parameter when the reactor has a multi-turn short-circuit fault. The change rates of each parameter in this state are calculated, as shown in Table 1.

[0041] Table 1 Change rate of each parameter when the inter-turn insulation breakdown fault of the reactor occurs

[0042]

[0043] Step 2: Calculate the parameter sensitivity base score;

[0044] The basic scoring basis of parameter sensitivity is as follows: according to the change rate from large to small, the first place is given x1 points, the second place is given x2 points, the third place is given x3 points, the fourth place is given x4 points, the fifth place is given x5 points, the sixth place is given x6 points, the seventh place is given x7 points, the eighth place is given x8 points, and the ninth place is given x9 points;

[0045] The score is determined based on:

[0046] |x k+1 -x k | / x k ×100%≥10%;

[0047] Where k = 1, 2, 3, 4, 5, 6, 7, 8, 9;

[0048] Tied Parameter Sorting: If a parameter's sensitivity base score is tied, the average parameter change rate for three fault types can be used as an auxiliary criterion. Specifically, the arithmetic mean of the parameter's change rate for three fault scenarios, namely, inter-turn short circuit, single-turn short circuit, and multi-turn short circuit, is calculated. Tied parameters are reordered from highest to lowest based on their average change rate. Parameters with higher average change rate have higher sensitivity base scores.

[0049] The specific ranking and parameter sensitivity basic score calculation process is as follows:

[0050] Assuming i = 1, that is, the change rate of each parameter when the short circuit fault occurs between different turns of the reactor, the ranking from the first to the ninth and the obtained parameter sensitivity basic scores are shown in Table 2.

[0051] Table 2 Ranking of parameters and basic parameter sensitivity scores when i=1

[0052]

[0053]

[0054] Let i = 2, which represents the change rate of each parameter when a single-turn short-circuit fault occurs in the reactor. The ranking from the first to the ninth position and the obtained parameter sensitivity basic scores are shown in Table 3.

[0055] Table 3 Ranking of parameters and basic parameter sensitivity scores when i=2

[0056]

[0057] Let i = 3, which represents the change rate of each parameter when the reactor has a multi-turn short-circuit fault. The ranking from the first to the ninth position and the obtained parameter sensitivity basic scores are shown in Table 4.

[0058] Table 4 Ranking of parameters and basic parameter sensitivity scores when i=3

[0059]

[0060]

[0061] Step 3: Calculate the parameter sensitivity additional coefficient

[0062] The value of the parameter sensitivity additional coefficient is determined based on:

[0063] 10%≤|m k+1 -m k | / m k×100%≤|x k+1 -x k | / x k ×100%;

[0064] According to the difficulty of parameter monitoring, the scores from easiest to most difficult are m1, m2, m3, m4, m5, m6, m7, m8, and m9; the additional sensitivity coefficients of parameters at different positions are as follows: m1 = 0.9, m2 = 0.8, m3 = 0.7, m4 = 0.6, m5 = 0.5, m6 = 0.4, m7 = 0.3, m8 = 0.2, and m9 = 0.1:

[0065] Factors that affect the difficulty of parameter monitoring include the environment, cost, the difficulty of installing the monitoring sensor, and monitoring errors. Parameters directly acquired by sensors are the easiest to set. For example, the total circuit current and spatial magnetic induction intensity can be directly acquired. However, the total circuit current is an electrical parameter, and the spatial magnetic induction intensity is a magnetic parameter. Electrical parameters are easier to acquire than magnetic parameters, so the total circuit current ranks first. The power factor angle can be obtained based on the phase difference between the measured voltage and current waveforms, which involves an additional step, so it ranks third. The directly acquired parameters are then ranked based on the complexity of the calculation. The calculation of power factor only involves one parameter, the power factor angle, while the calculation of equivalent impedance involves two parameters. The calculation of equivalent resistance, equivalent reactance, and active power loss involves three parameters, so the above parameters are ranked in order. Accurately measuring the stress on adjacent turns of a winding fault is difficult with sensors, so it is set to be the most challenging. Therefore, the first to ninth digits are, in order, total loop current, spatial magnetic induction intensity, power factor angle, power factor, equivalent impedance, active power loss, equivalent resistance, equivalent reactance, and stress on adjacent turns of the winding fault. The parameter sensitivity coefficients are 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, and 0.1, respectively.

[0066] If there is a tie according to the above criteria, the parameters are sorted by their rate of change, with the parameter with the larger rate of change ranking higher. If the parameter change rates are tied, the average of the parameter change rates under the three fault types is introduced as an auxiliary judgment criterion; specifically, the arithmetic mean of the parameter change rate under the three fault conditions of inter-turn short circuit, single-turn short circuit, and multi-turn short circuit is calculated. For the tied parameters, they are re-sorted from large to small according to their average rate of change. The parameter with the higher average rate of change has a higher sensitivity score.

[0067] If the equivalent resistance, equivalent reactance, and active power loss are listed in a tie, they will be sorted according to the parameter change rate, with active power loss, equivalent resistance, and equivalent reactance ranked from sixth to eighth.

[0068] Step 4: Calculate the parameter sensitivity composite score

[0069] according to The calculated comprehensive sensitivity scores of each parameter are as follows:

[0070] Taking the total circuit current score as an example, g = 0.9 × (2 + 2 + 2) = 5.4; the scores of the remaining parameters are as follows:

[0071] The equivalent resistance score is 5.7, the equivalent reactance score is 2, the equivalent impedance score is 2.5, the power factor angle score is 9.1, the power factor score is 11.4, the active power loss score is 8.8, the spatial magnetic induction intensity score is 11.2, and the force score of the adjacent turns of the winding fault position is 2.7.

[0072] Step 5: Extract sensitive parameters

[0073] According to the calculation results of step 4, the parameter with the highest comprehensive parameter sensitivity score is the power factor, so the power factor is extracted as the sensitive parameter.

Claims

1. A sensitive parameter extraction method for dry-type air-core reactor fault diagnosis, characterized by: The following steps are involved: ① Count the parameters used for dry-type air-core reactor fault diagnosis and calculate the parameter change rate; ②Calculate the basic score of parameter sensitivity; ③Calculate the additional coefficient of parameter sensitivity; ④ Calculate the comprehensive score of parameter sensitivity; ⑤ Extract sensitive parameters and take the parameter with the highest comprehensive sensitivity score as the sensitive parameter.

2. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ①, the parameters used for dry-type air-core reactor fault diagnosis include total loop current, equivalent resistance, equivalent reactance, equivalent impedance, power factor angle, power factor, active power loss, spatial magnetic induction intensity and the force on adjacent turns of the winding fault position.

3. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1 is characterized in that: In step ①, the parameter change rate is calculated as follows: △And ij =|And ij -AND ij0 | / And ij0 ×100%. Among them, △Y ij is the parameter change rate, Y ij Y is the parameter value when the reactor has inter-turn insulation breakdown fault, ij0 These are the parameter values ​​for the reactor during normal operation.

4. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: The basis for calculating the basic score of parameter sensitivity in step ② is: rank from large to small according to the rate of change, the first place is given x1 points, the second place is given x2 points, the third place is given x3 points, the fourth place is given x4 points, the fifth place is given x5 points, the sixth place is given x6 points, the seventh place is given x7 points, the eighth place is given x8 points, and the ninth place is given x9 points.

5. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ②, the score value is determined based on: |x k+1 -x k | / x k ×100%≥10%; Among them, k = 1, 2, 3, 4, 5, 6, 7, 8, 9.

6. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ②, when the parameter sensitivity basic scores are tied, the mean parameter change rate under the three fault types is introduced as an auxiliary criterion; the specific method is: calculate the arithmetic mean of the parameter change rate under the three fault situations of short circuit between different turns, short circuit between single turns, and short circuit between multiple turns, respectively. For the tied parameters, re-sort them from large to small according to their mean change rate. The one with a higher mean change rate has a higher sensitivity basic score.

7. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ③, the parameter sensitivity additional coefficient is based on the following: according to the difficulty of parameter monitoring, the scores are m1, m2, m3, m4, m5, m6, m7, m8, and m9 from the easiest to the most difficult.

8. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ③, factors that affect the difficulty of parameter monitoring include environment, cost, difficulty of installing monitoring sensors, and monitoring error.

9. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ③, the value of the parameter sensitivity additional coefficient is determined based on: 10%≤|m k+1 -m k | / m k ×100%≤|x k+1 -x k | / x k ×100%; Among them, k = 1, 2, 3, 4, 5, 6, 7, 8, 9.

10. The sensitive parameter extraction method for dry-type air-core reactor fault diagnosis according to claim 1, characterized in that: In step ④, the comprehensive parameter sensitivity score is: Among them, gi is the comprehensive sensitivity score of each parameter, m Yij is the parameter sensitivity additional coefficient corresponding to each parameter, x Yij It is the parameter sensitivity basic score corresponding to each parameter.

Images