Converter valve side bushing on-line insulation monitoring method based on multi-dimensional time sequence characteristics

Abstract

The application discloses a kind of converter valve side bushing on-line insulation monitoring method based on multidimensional timing characteristics, belong to the on-line monitoring technical field of capacitive equipment.The application obtains the mapping relationship of multiple harmonic resistance and direct current resistance by carrying out off-line test to converter valve side bushing;While synchronously collecting the leakage current and voltage signal of bushing, the mapping relationship is used to calculate timing resistance, multiple harmonic capacitance, multiple harmonic dielectric loss as insulation characteristic parameter, and then the correlation coefficient between each corresponding timing insulation characteristic parameter of the same batch of bushing is calculated;The correlation coefficient between all bushings is used to form sequence and is normalized to obtain evaluation sequence, find out the position of abnormal value and mark, while generating and updating evaluator, finally accurately locate abnormal bushing through the result of evaluator.So, the application realizes high sensitivity on-line monitoring of abnormal state caused by insulation deterioration by using mutual supervision of multiple insulation characteristic parameters between large quantities of bushings.

Description

Technical Field

[0001] This invention belongs to the field of online monitoring technology for capacitive equipment, and more specifically, relates to an online insulation monitoring method for converter transformer valve side bushings based on multi-dimensional time-series characteristics. Background Technology

[0002] Compared to AC transmission systems, ultra-high voltage direct current (UHVDC) transmission is gradually becoming the future direction of power system development due to its advantages of lower cost, lower power consumption, and greater stability. UHVDC transmission is generally carried out in an "AC-DC-AC" manner, where AC and DC systems are interdependent and closely coupled. In the AC-DC conversion process, the converter transformer is the core equipment connecting the converter station and the inverter station. The valve-side bushing of the converter transformer, as a core component, plays a crucial role in conductive connection, insulation isolation, and mechanical connection; its insulation condition directly affects the safety and stability of the entire DC transmission system.

[0003] Due to their special function, the valve-side bushings of converter transformers are located at the connection point between the AC and DC systems, thus bearing both AC and DC voltages simultaneously. This makes the bushing insulation more susceptible to degradation, which is often a slow, cumulative process culminating in an explosive failure. Therefore, only online monitoring and timely detection and repair of anomalies can ensure the stability of the bushing insulation to the greatest extent possible. However, currently, there are few online monitoring methods for the insulation of converter transformer valve-side bushings. The two most mature methods are internal gas pressure monitoring and voltage divider monitoring at the end of the valve-side bushing, and both have low sensitivity to early-stage faults.

[0004] In Chinese invention patent CN 112305319 A, a method was proposed to calculate the dielectric loss value based on measuring the voltage and current amplitude and phase, and to directly evaluate the insulation status of the bushing according to pre-set rules and pre-set thresholds. This method has a certain degree of effectiveness. However, in reality, bushings are usually in a working environment with high DC voltage and multiple harmonic voltages. It is very difficult to obtain the phase information of the voltage and current, and the accuracy of the acquisition is not high. This seriously affects the reliability of the evaluation when the dielectric loss value of the bushing itself is very small. Moreover, simply setting thresholds for comparison can usually only identify insulation degradation that exceeds the acceptable range, and cannot detect problems and carry out repairs in the early stage of degradation. Summary of the Invention

[0005] To address the shortcomings and improvement needs of existing technologies, this invention provides an online insulation monitoring method for converter transformer valve side bushings based on multi-dimensional time-series characteristics, aiming to solve the technical problems of low sensitivity and poor reliability of existing monitoring methods.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for online insulation monitoring of converter transformer valve-side bushings based on multi-dimensional time-series characteristics, comprising:

[0007] S1. Obtain the mapping relationship between the k-th harmonic resistance and DC resistance of the valve-side bushing i of the converter transformer to be monitored through offline testing, i = 1, ..., N, where N is the total number of bushings;

[0008] S2, Measure the leakage current signal I of bushing i. i (t) and the voltage signal U across bushing i i (t), and for each step, the leakage current signal I i (t) and voltage signal U i (t) Perform Fourier analysis to obtain the voltage amplitude U of the kth harmonic within the r-th step of bushing i. (k)(r)(i) and current amplitude I (k)(r)(i) ;

[0009] S3, based on voltage amplitude U (k)(r)(i) Current amplitude I (k)(r)(i) And based on the mapping relationship, calculate the DC resistance R within the r-th step of bushing i. (0)(r)(i) The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) And respectively construct the time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix; where the rows of the three matrices all represent the r-th step size, and the columns all represent the i-th bushing;

[0010] S4. For any casings i and j, calculate the correlation coefficients between the column vectors of the three matrices, and then calculate the correlation coefficient corr between casings i and j by weighting. (i)(j) The correlation coefficients among N sleeves are used to form a sequence P; after performing a normal transformation on the sequence P, the evaluation sequence P1 is obtained, and the upper and lower limits of the evaluation sequence P1 are determined.

[0011] S5. Generate and initialize the evaluator COUNT. Traverse each element in the evaluation sequence P1. If an element exceeds the upper or lower limit of the evaluation sequence P1, increment the count of the two sleeves corresponding to that element.

[0012] S6. Determine that the sleeve corresponding to element R in the final evaluator is a normal sleeve, and the rest are abnormal sleeves; where R is the smallest number in the evaluator.

[0013] Furthermore, in S3, based on the voltage amplitude U (k)(r)(i) Current amplitude I (k)(r)(i) And based on the mapping relationship, calculate the DC resistance R within the r-th step of bushing i. (0)(r)(i)The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) Specifically:

[0014]

[0015]

[0016]

[0017] Among them, R (k)(r)(i) =f (i)(k) (R (0)(r)(i) ), f (i)(k) This represents the mapping relationship between the k-th harmonic resistance and the DC resistance of bushing i, ω. k This represents the angular frequency of the kth harmonic.

[0018] Furthermore, in S3, the time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix are respectively expressed as:

[0019] Timing DC resistor matrix:

[0020]

[0021] Timing k-th harmonic capacitor matrix:

[0022]

[0023] Time-series k-th harmonic dielectric loss matrix:

[0024]

[0025] Where s is the total number of steps.

[0026] Furthermore, in S4, the correlation coefficients between the column vectors of the three matrices are as follows:

[0027]

[0028]

[0029]

[0030] Where cR0(i,j) represents matrix R (0) The correlation coefficient between the i-th and j-th columns. Represents matrix R (0) The average value of the elements in the i-th column; cC k (i,j) represents matrix C (k) The correlation coefficient between the i-th and j-th columns. Representing matrix C (k) The average value of the elements in the i-th column; ctanδ k (i,j) represents the matrix tanδ (k) The correlation coefficient between the i-th and j-th columns. Represents the matrix tanδ (k) The average value of the elements in the i-th column.

[0031] Furthermore, in S4, the sequence P is represented as:

[0032] P = [corr] (1)(2) ,corr (1)(3) ,…,corr (1)(N) ,corr (2)(3) ,…,corr (2)(N) ,…,corr (N-1)(N) ].

[0033] Further, in step S4, the evaluation sequence P1 is obtained by performing a normal transformation on sequence P, specifically as follows:

[0034] Perform a Box-Cox transformation on sequence P to obtain the evaluation sequence P1:

[0035]

[0036] Where P1(t) represents the t-th element in sequence P1, P(t) represents the t-th element in sequence P, and the value of λ is taken from [-1,1] depending on the dispersion of the data.

[0037] Furthermore, in S4, the upper limit Max and lower limit Min of the evaluation sequence P1 are respectively:

[0038] Max = Q3 + 1.5(Q3 - Q1)

[0039] Min = Q1 - 1.5(Q3 - Q1)

[0040] Q3 and Q1 are the upper and lower quartiles of the evaluation sequence P1, respectively.

[0041] Furthermore, in S5, the evaluator COUNT = [0,0,0,…,0] is initialized, where COUNT contains N elements and corresponds one-to-one with the N sleeves.

[0042] In a second aspect, the present invention provides an online insulation monitoring system for converter transformer valve side bushing based on multi-dimensional time-series characteristics, comprising: a computer-readable storage medium and a processor;

[0043] The computer-readable storage medium is used to store executable instructions;

[0044] The processor is used to read executable instructions stored in the computer-readable storage medium and execute the online insulation monitoring method for converter transformer side bushing based on multi-dimensional timing characteristics as described in the first aspect.

[0045] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0046] (1) This invention obtains the mapping relationship between multiple harmonic resistance and DC resistance by conducting offline tests on the bushings of the converter transformer valve side; while simultaneously acquiring the leakage current and voltage signals of the bushings, the time-series resistance, multiple harmonic capacitance, and multiple harmonic dielectric loss are calculated as insulation characteristic parameters using this mapping relationship, and then the correlation coefficients between the corresponding time-series insulation characteristic parameters of the same batch of bushings are calculated; the correlation coefficients between all bushings are used to form a sequence and after normalization, an evaluation sequence is obtained, the location of outliers is identified and marked, an evaluator is generated and updated, and finally the abnormal bushings are accurately located through the results of the evaluator. Thus, this invention comprehensively considers DC resistance, multiple harmonic capacitance, and multiple harmonic dielectric loss, and utilizes the mutual supervision of multiple insulation characteristic parameters among a large batch of bushings to achieve online mutual monitoring, which greatly improves the sensitivity to changes in insulation characteristic parameters when insulation deteriorates, and enables early detection and maintenance of insulation deterioration.

[0047] (2) The present invention can calculate the insulation characteristic parameters of the valve side bushing without measuring the leakage current phase and the bus voltage phase, which greatly simplifies the process of obtaining the insulation characteristic parameters and the method has higher reliability than the phase calculation. Attached Figure Description

[0048] Figure 1 The flowchart illustrates an online insulation monitoring method for converter transformer valve side bushings based on multi-dimensional time-series characteristics, provided as an embodiment of the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0050] In this invention, the terms "first," "second," etc. (if present) in the invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0051] See Figure 1This invention provides an online insulation monitoring method for converter transformer valve side bushing based on multi-dimensional time-series characteristics, which is divided into an offline stage and an online stage. The offline stage includes operation S1, and the online stage includes operations S2 to S6.

[0052] Operation S1 obtains the mapping relationship between the k-th harmonic resistance and DC resistance of the valve-side bushing i of the converter transformer to be monitored through offline testing, where i = 1, ..., N, and N is the total number of bushings.

[0053] In this embodiment, the valve-side bushings of the converter transformers to be monitored are first classified and numbered according to pre-set rules. These pre-set rules include: same rated voltage, similar or identical voltage waveforms experienced by the bushings during operation, same manufacturer, and similar or identical commissioning times. Bushings conforming to the same category are grouped together and numbered.

[0054] It should be noted that the insulation monitoring method described in this embodiment takes one type of bushing as an example; other types can be operated in the same way, from S1 to S6.

[0055] The mapping relationship between the k-th harmonic resistance and the DC resistance of any bushing i was obtained through offline experiments and can be expressed as R (k)(r)(i) =f (i)(k) (R (0)(r)(i) ).

[0056] Operate S2 to measure the leakage current signal I of bushing i. i (t) and the voltage signal U across bushing i i (t), and for each step, the leakage current signal I i (t) and voltage signal U i (t) Perform Fourier analysis to obtain the voltage amplitude U of the kth harmonic within the r-th step of bushing i. (k)(r)(i) and current amplitude I (k)(r)(i) .

[0057] Specifically, operation S2 includes sub-operations S21 to S23.

[0058] In sub-operation S21, a current sensor is used to measure the leakage current signal I of any bushing i in real time online. i (t), and simultaneously use a voltage transformer to measure the voltage signal U across any bushing i online. i (t).

[0059] In suboperation S22, L is used as the signal frequency band length, and the total number of steps is s; where L depends on the highest order of the observed harmonics, and s is determined according to actual engineering requirements.

[0060] In sub-operation S23, the leakage current signal I within each step size...i (t) and voltage signal U i (t) Perform Fourier analysis to obtain the voltage amplitude U of the kth harmonic within the r-th step of bushing i. (k)(r)(i) and current amplitude I (k)(r)(i) .

[0061] Operation S3, based on voltage amplitude U (k)(r)(i) Current amplitude I (k)(r)(i) And based on the mapping relationship, calculate the DC resistance R within the r-th step of bushing i. (0)(r)(i) The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) The time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix are constructed respectively; where the rows of the three matrices represent the r-th step size and the columns represent the i-th bushing.

[0062] Among them, the DC resistance R in the r-th step of bushing i (0)(r)(i) The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) , respectively represented as:

[0063]

[0064]

[0065]

[0066] In the formula, ω k This represents the angular frequency of the kth harmonic.

[0067] Furthermore, the time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix are respectively expressed as:

[0068] Timing DC resistor matrix:

[0069]

[0070] Timing k-th harmonic capacitor matrix:

[0071]

[0072] Time-series k-th harmonic dielectric loss matrix:

[0073]

[0074] In the formula, s is the total number of steps.

[0075] Operation S4: For any casing i and j, calculate the correlation coefficients between the column vectors of the three matrices, and then calculate the correlation coefficient corr between casing i and j by weighting. (i)(j) The correlation coefficients among N casings are used to form a sequence P; after performing a normal transformation on the sequence P, the evaluation sequence P1 is obtained, and the upper and lower limits of the evaluation sequence P1 are determined.

[0076] Specifically, operation S4 includes sub-operations S41 to S45.

[0077] In suboperation S41, the correlation coefficients between the column vectors of the three matrices are calculated as follows:

[0078]

[0079]

[0080]

[0081] In the formula, cR0(i,j) represents matrix R (0) The correlation coefficient between the i-th and j-th columns. Represents matrix R (0) The average value of the elements in the i-th column; cC k (i,j) represents matrix C (k) The correlation coefficient between the i-th and j-th columns. Representing matrix C (k) The average value of the elements in the i-th column; ctanδ k (i,j) represents the matrix tanδ (k) The correlation coefficient between the i-th and j-th columns. Represents the matrix tanδ (k) The average value of the elements in the i-th column.

[0082] In suboperation S42, the correlation coefficient corr between casings i and j is calculated using a weighted average. (i)(j) :

[0083]

[0084] In the formula, w0, w Ck w tanδk K represents the set of values ​​for k, where k is the weighting coefficient.

[0085] In sub-operation S43, the correlation coefficient matrix corr is formed using the correlation coefficients between the various bushings; the lower triangular matrix of corr is taken, the data on the diagonal are removed, and the matrix is ​​straightened to form the sequence P = [corr]. (1)(2) ,corr (1)(3) ,…,corr (1)(N) ,corr(2)(3) ,…,corr (2)(N) ,…,corr (N-1)(N) ].

[0086] In suboperation S44, the Box-Cox transformation is performed on sequence P to obtain the evaluation sequence P1:

[0087]

[0088] In the formula, P1(t) represents the t-th element in sequence P1, P(t) represents the t-th element in sequence P, and the value of λ is taken from [-1,1] depending on the dispersion of the data.

[0089] In suboperation S45, the upper limit Max and lower limit Min of the evaluation sequence P1 are calculated as follows:

[0090] Max = Q3 + 1.5(Q3 - Q1)

[0091] Min = Q1 - 1.5(Q3 - Q1)

[0092] In the formula, Q3 and Q1 are the upper and lower quartiles of the evaluation sequence P1, respectively.

[0093] Operation S5 generates and initializes the evaluator COUNT. Iterates through each element in the evaluation sequence P1. If an element exceeds the upper or lower limit of the evaluation sequence P1, the count of the two sleeves corresponding to that element is incremented by one.

[0094] Specifically, initialize the evaluator COUNT = [0,0,0,…,0], where COUNT contains N elements, each corresponding to one of the N sleeves. Traverse the evaluation sequence P1. If the t-th element exceeds the upper and lower limits, mark the t-th element in the corresponding sequence P and determine the two corresponding sleeves, thereby incrementing the values ​​of the two corresponding elements in the evaluator COUNT by one.

[0095] Operation S6 determines that the sleeve corresponding to element R in the final evaluator is a normal sleeve, and the rest are abnormal sleeves; where R is the smallest number in the evaluator.

[0096] Specifically, the first step is to find the smallest number R in the evaluator. The sleeve corresponding to the element R in the final evaluator is determined to be a normal sleeve, and the rest are abnormal sleeves, which should be inspected offline as soon as possible.

[0097] The invention will be further illustrated below with reference to examples. The specific steps of a method for online insulation monitoring of converter transformer valve side bushings based on multi-dimensional time-series characteristics are as follows:

[0098] Step 1: Based on the information on the valve side bushing nameplate, group the 10 800kV voltage level converter transformer valve side bushings, manufactured by ABB, and all of which have been in use for about 3 years, into a group numbered 1 to 10.

[0099] The mapping relationship between its resistance and DC resistance under power frequency and third harmonic conditions was measured offline, where:

[0100] R (1)(i) =0.9R (0)(i)

[0101] R (3)(i) =0.85R (0)(i)

[0102] Step 2: Measure the leakage current amplitude and voltage amplitude across the bushings of valves 1-10 online. The sensor sampling frequency is 10kHz, with a step size of 10k points and a total step size of 10. Perform Fourier analysis to extract the amplitude of the kth harmonic component and record it in a 10*10 matrix. The column coordinates represent the i-th bushing, and the row coordinates represent the r-th step size.

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109] Step 3: Using the data from Step 2 and the mapping relationship from Step 1, calculate the DC resistance, fundamental and third harmonic capacitance, and dielectric loss factor of any i-th bushing within the r-th step size, and record them in a 10*10 matrix, where the column coordinates represent the i-th bushing and the row coordinates represent the r-th step size.

[0110]

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117] Step S4: Calculate the correlation coefficients of various insulation characteristic parameters for the 10 bushings:

[0118]

[0119]

[0120]

[0121]

[0122]

[0123] The correlation coefficient corr between casings i and j was calculated using a weighted average. (i)(j) :

[0124]

[0125] In this embodiment, w0 = 0.3. The correlation coefficient matrix corr is:

[0126]

[0127] Take the lower half of the correlation coefficient matrix corr, remove the diagonal, and add it sequentially to form sequence P:

[0128] P=[0.99110,0.99251,0.98906,…0.99384]

[0129] Taking λ = 0.8, perform a Box-Cox transformation on the sequence P according to the following formula to normalize it:

[0130]

[0131] A new evaluation sequence is obtained:

[0132] P1=[-0.20019,-0.20016,-0.20021,…-0.20013]

[0133] The upper and lower quartiles of the evaluation sequence P1 were calculated to be: Q3 = -0.20014, Q1 = -0.20026, and the upper and lower limits were further determined to be:

[0134] Max=Q3+1.5(Q3-Q1)=-0.199958

[0135] Min = Q1 - 1.5(Q3 - Q1) = -0.200454

[0136] Step 5: Initialize the evaluator COUNT = [0, 0, 0, …, 0], where COUNT contains N elements and corresponds to N bushings one by one. Traverse the evaluation sequence P1. If the t-th element exceeds the upper and lower limits, mark the t-th element in the corresponding sequence P, and determine the corresponding two bushings, so as to increment the values of the corresponding two elements in the evaluator COUNT by one.

[0137] For example, P1(6) = -0.217549 < Min, which corresponds to P(6) and corr(1, 7). Therefore, the evaluator is updated to:

[0138] COUNT = [1, 0, 0, 0, 0, 0, 1, 0, …, 0]

[0139] And so on. After directly traversing completely, the final evaluator is:

[0140] COUNT = [1, 1, 1, 1, 1, 1, 9, 1, 1, 1]

[0141] Step 6: After evaluation, the smallest number in the evaluator is 1. There are 9 bushings whose corresponding elements in the evaluator are 1. These 9 bushings are normal bushings, and the result of the bushing numbered 7 in the evaluator is 9. It is an abnormal bushing and should be repaired immediately.

[0142] To sum up, through the mapping relationship between the k-th harmonic resistance and the DC insulation resistance of the valve side bushings of the converter transformer measured offline, the present invention can calculate the k-th harmonic capacitance and the dielectric loss factor without measuring the bus voltage phase and the bushing leakage current phase. Moreover, through the mutual correlation and mutual supervision of a large number of bushings, high-sensitivity online monitoring of abnormal states caused by insulation deterioration is achieved.

[0143] It is easy for those skilled in the art to understand that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for online insulation monitoring of converter transformer valve-side bushings based on multi-dimensional time-series characteristics, characterized in that, include: S1. Obtain the mapping relationship between the k-th harmonic resistance and DC resistance of the valve-side bushing i of the converter transformer to be monitored through offline testing, i = 1, ..., N, where N is the total number of bushings; S2, Measure the leakage current signal I of bushing i. i (t) and the voltage signal U across bushing i i (t), and for each step, the leakage current signal I i (t) and voltage signal U i (t) Perform Fourier analysis to obtain the voltage amplitude U of the kth harmonic within the r-th step of bushing i. (k)(r)(i) and current amplitude I (k)(r)(i) ; S3, based on voltage amplitude U (k)(r)(i) Current amplitude I (k)(r)(i) And based on the mapping relationship, calculate the DC resistance R within the r-th step of bushing i. (0)(r)(i) The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) And respectively construct the time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix; where the rows of the three matrices all represent the r-th step size, and the columns all represent the i-th bushing; S4. For any casings i and j, calculate the correlation coefficients between the column vectors of the three matrices, and then calculate the correlation coefficient corr between casings i and j by weighting. (i)(j) The correlation coefficients among N sleeves are used to form a sequence P; after performing a normal transformation on the sequence P, the evaluation sequence P1 is obtained, and the upper and lower limits of the evaluation sequence P1 are determined. S5. Generate and initialize the evaluator COUNT. Traverse each element in the evaluation sequence P1. If an element exceeds the upper or lower limit of the evaluation sequence P1, increment the count of the two sleeves corresponding to that element. S6. Determine that the sleeve corresponding to element R in the final evaluator is a normal sleeve, and the rest are abnormal sleeves; where R is the smallest number in the evaluator.

2. The online insulation monitoring method for converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 1, characterized in that, In S3, based on the voltage amplitude U (k)(r)(i) Current amplitude I (k)(r)(i) And based on the mapping relationship, calculate the DC resistance R within the r-th step of bushing i. (0)(r)(i) The equivalent capacitance C under the kth harmonic (k)(r)(i) and the dielectric loss factor tanδ under the kth harmonic. (k)(r)(i) Specifically: Among them, R (k)(r)(i) =f (i)(k) (R (0)(r)(i) ), f (i)(k) This represents the mapping relationship between the k-th harmonic resistance and the DC resistance of bushing i, ω. k This represents the angular frequency of the kth harmonic.

3. The online insulation monitoring method for converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 2, characterized in that, In S3, the time-series DC resistance matrix, the time-series k-th harmonic capacitance matrix, and the time-series k-th harmonic dielectric loss matrix are respectively represented as: Timing DC resistor matrix: Timing k-th harmonic capacitor matrix: Time-series k-th harmonic dielectric loss matrix: Where s is the total number of steps.

4. The online insulation monitoring method for converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 3, characterized in that, In S4, the correlation coefficients between the column vectors of the three matrices are as follows: Where cR0(i,j) represents matrix R (0) The correlation coefficient between the i-th and j-th columns. Represents matrix R (0) The average value of the elements in the i-th column; cC k (i,j) represents matrix C (k) The correlation coefficient between the i-th and j-th columns. Representing matrix C (k) The average value of the elements in the i-th column; ctanδ k (i,j) represents the matrix tanδ (k) The correlation coefficient between the i-th and j-th columns. Represents the matrix tanδ (k) The average value of the elements in the i-th column.

5. The method for online insulation monitoring of converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 1, characterized in that, In S4, sequence P is represented as: P=[corr (1)(2) ,corr (1)(3) ,…,corr (1)(N) ,corr (2)(3) ,…,corr (2)(N) ,…,corr (N-1)(N) ]。 6. The online insulation monitoring method for converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 5, characterized in that, In step S4, the evaluation sequence P1 is obtained by performing a normal transformation on sequence P, specifically as follows: Perform a Box-Cox transformation on sequence P to obtain the evaluation sequence P1: Where P1(t) represents the t-th element in sequence P1, P(t) represents the t-th element in sequence P, and the value of λ is taken from [-1,1] depending on the dispersion of the data.

7. The method for online insulation monitoring of converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 6, characterized in that, In S4, the upper limit Max and lower limit Min of the evaluation sequence P1 are respectively: Max = Q3 + 1.5(Q3 - Q1) Min = Q1 - 1.5(Q3 - Q1) Q3 and Q1 are the upper and lower quartiles of the evaluation sequence P1, respectively.

8. The method for online insulation monitoring of converter transformer valve-side bushings based on multi-dimensional time-series characteristics according to claim 1, characterized in that, In step S5, the evaluator COUNT = [0,0,0,…,0] is initialized, where COUNT contains N elements and corresponds one-to-one with N sleeves.

9. An online insulation monitoring system for converter transformer valve-side bushings based on multi-dimensional time-series characteristics, characterized in that, include: Computer-readable storage media and processors; The computer-readable storage medium is used to store executable instructions; The processor is used to read executable instructions stored in the computer-readable storage medium and execute the online insulation monitoring method for converter transformer side bushing based on multi-dimensional timing characteristics as described in any one of claims 1-8.

Images