A monitoring method for on-load tap changer transmission mechanism based on sparse filtering

Abstract

The present invention discloses a method for monitoring an on-load tap changer transmission mechanism based on sparse filtering, which relates to the technical field of on-load tap changer transmission mechanism detection. The method involves performing multiple gear switching on the on-load tap changer to collect vibration signals of the transmission mechanism during the on-load tap changer switching process. A sparse filtering network-based vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism is calculated based on the vibration signals. The eigenvalues ​​and eigenvectors of the principal component components of the vibration signal characteristic distribution matrix are calculated. The weighted mean of the elements of the statistic and the corresponding control limits are calculated based on the eigenvalues ​​and eigenvectors of the principal component components of the characteristic distribution matrix. The mechanical state of the transmission mechanism is determined based on the weighted mean of the elements of the statistic and the corresponding control limits. The method of the present invention efficiently and accurately determines the mechanical state of the on-load tap changer transmission mechanism by performing real-time monitoring and computational analysis of the vibration signals of the transmission mechanism during the on-load tap changer switching process.

Description

Technical Field

[0001] The present invention belongs to the technical field of on-load tap changer transmission mechanism detection, and in particular relates to a sparse filtering-based on-load tap changer transmission mechanism monitoring method. Background Art

[0002] The on-load tap-changer is the only movable component of the on-load tap-changing transformer. It is mainly composed of a selector switch, a transfer switch, a transmission mechanism driven by an electric mechanism, etc. It can change the voltage ratio under load conditions, realize the regulation of the power system voltage without power outage, and realize important functions such as compensating for voltage fluctuations, regulating power, improving system performance and improving power quality.

[0003] However, as on-load tapchangers age and the number of voltage regulation operations increases, their failure rate also increases. On-load tapchanger failures primarily include electrical and mechanical failures, with mechanical failure being the primary type, accounting for over 90% of all on-load tapchanger failures. Furthermore, fault inversion analysis of multiple faulty on-load tapchangers indicates that mechanical failure is also a primary cause of some electrical failures. As a crucial component of the on-load tapchanger, the transmission mechanism—mainly the horizontal and vertical drive shafts, the bevel gear box, and the mounting mechanism—is crucial for ensuring the on-load tapchanger's voltage regulation function. As the number of voltage regulation operations and the number of years an on-load tapchanger operates increases, mechanical hazards such as loosening, sticking, and wear are inevitable. Therefore, it is crucial to monitor and analyze changes in the mechanical condition of the on-load tapchanger transmission mechanism in a timely and effective manner to identify early or potential mechanical failures and implement effective maintenance measures. This is crucial for ensuring the safe operation of the on-load tapchanger and transformer.

[0004] During the on-load tapchanger's gear shifting process, mechanical vibrations, such as collisions or friction between components such as the transmission mechanism, are generated. These vibrations are transmitted through the on-load tapchanger's structural components to the tapchanger's head cover or the transformer tank wall, generating mechanical vibration signals. These mechanical vibration signals clearly contain rich information about the on-load tapchanger's mechanical state. Consequently, on-load tapchanger mechanical condition monitoring methods based on vibration analysis have attracted increasing attention from researchers both domestically and internationally. The key advantage of this method is that vibration sensors placed on the tapchanger's head cover or the transformer tank can capture vibration signals during the gear shifting process. Any changes in the mechanical characteristics of the on-load tapchanger during the shifting process can be reflected in the vibration signals, significantly improving the sensitivity of the on-load tapchanger's mechanical condition detection. Furthermore, placing the vibration sensors on the transformer tank wall is a safe and convenient vibration detection method, facilitating online monitoring and improving the operational reliability of the on-load tapchanger.

[0005] However, the mechanical structure of a transformer's on-load tap-changer is complex, and the mechanical vibration signals generated by collisions and friction between its transmission mechanism components exhibit strong time-varying and non-stationary characteristics. Therefore, obtaining vibration signal evaluation indicators for monitoring the mechanical condition of the on-load tap-changer's transmission mechanism has always been a research challenge. Therefore, a sparse filtering-based on-load tap-changer transmission mechanism monitoring method is needed. Summary of the Invention

[0006] The object of the present invention is to provide a method for monitoring an on-load tap changer transmission mechanism based on sparse filtering, thereby overcoming the disadvantage of inaccurate judgment of vibration signals obtained for monitoring the mechanical state of the on-load tap changer transmission mechanism.

[0007] To achieve the above object, the present invention provides a method for monitoring an on-load tap changer transmission mechanism based on sparse filtering, comprising the following steps:

[0008] Performing multiple gear switching on the on-load tap changer and collecting vibration signals of the transmission mechanism during the on-load tap changer switching process;

[0009] Calculating a vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on a sparse filter network according to the vibration signal;

[0010] Calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix;

[0011] Calculating the element-weighted mean and corresponding control limits of the statistic based on the eigenvalues ​​and eigenvectors of the principal component components of the characteristic distribution matrix;

[0012] The mechanical state of the transmission mechanism is judged according to the weighted mean value of the elements of the statistic and the corresponding control limit, thereby realizing the monitoring of the transmission mechanism of the on-load tap changer.

[0013] Preferably, before calculating the vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on the sparse filter network according to the vibration signal, the vibration signal is segmented.

[0014] Preferably, the segmentation is to segment the vibration signal in each switching process according to the rotation period of the on-load tap changer transmission shaft.

[0015] Preferably, calculating the vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on the sparse filter network according to the vibration signal specifically includes:

[0016] Determine the network structure and parameters of the sparse filtering network;

[0017] Normalize each segmented vibration signal and then construct a sparse filter sample set;

[0018] Randomly initializing the weight matrix and error precision, and constructing the objective function of the sparse filtering network in combination with the sparse filtering sample set;

[0019] Solve the objective function of the sparse filtering network and obtain the required weight matrix;

[0020] The vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism is calculated in combination with the weight matrix.

[0021] Preferably, the sparse filtering network includes an input layer and an output layer, and the input layer is connected to the output layer via a weight matrix.

[0022] Preferably, the objective function of the sparse filtering network is iteratively solved based on the limited memory quasi-Newton method, and the iteration is stopped when the error accuracy is met to obtain the required weight matrix.

[0023] Preferably, calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix specifically includes:

[0024] Performing singular value decomposition on the vibration signal characteristic distribution matrix to obtain singular values ​​and corresponding eigenvectors of the vibration signal characteristic distribution matrix;

[0025] Calculate the cumulative contribution rate of matrix eigenvalues;

[0026] When the cumulative contribution rate of the kth eigenvalue is greater than the threshold, the eigenvectors corresponding to the k eigenvalues ​​before the kth eigenvalue are selected as the principal component eigenvalues ​​and eigenvectors of the characteristic distribution matrix.

[0027] Preferably, the mechanical state of the transmission mechanism is judged according to the weighted mean of the elements of the statistic and the corresponding control limits, thereby realizing monitoring of the on-load tap changer transmission mechanism, which specifically includes:

[0028] When the weighted mean of each element is higher than the corresponding control limit, it is determined that the mechanical state of the on-load tap changer transmission mechanism has changed, and timely maintenance is required; otherwise, the mechanical state of the on-load tap changer transmission mechanism has not changed.

[0029] Compared with the existing technology, the present invention has the following beneficial effects:

[0030] 1. The sparse filtering-based on-load tap changer transmission mechanism monitoring method provided by the present invention collects vibration signals of the transmission mechanism during the on-load tap changer switching process by performing multiple gear switching on the on-load tap changer; calculates the vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on a sparse filtering network based on the vibration signals; calculates the eigenvalues ​​and eigenvectors of the principal component components of the vibration signal characteristic distribution matrix; calculates the element-weighted mean of the statistic and the corresponding control limits based on the principal component components of the characteristic distribution matrix; and discriminates the mechanical state of the transmission mechanism based on the element-weighted mean of the statistic and the corresponding control limits, thereby achieving on-load tap changer transmission mechanism monitoring. The method of the present invention can achieve efficient and accurate judgment of the mechanical state of the on-load tap changer transmission mechanism by real-time monitoring and computational analysis of the vibration signals of the transmission mechanism during the on-load tap changer switching process.

[0031] 2. Before calculating the vibration signal feature distribution matrix of the on-load tap changer transmission mechanism based on the sparse filtering network according to the vibration signal, multiple groups of vibration signals of the transmission mechanism during the gear switching process of the on-load tap changer are segmented and processed. While retaining the main feature information of the vibration signal of the on-load tap changer transmission mechanism to the maximum extent, the problem of extracting the vibration signal features of the transmission mechanism when the training samples are insufficient is effectively solved.

[0032] 3. The method of using sparse filtering network to obtain the characteristic distribution matrix of the transmission mechanism vibration signal can accurately extract the complex characteristic parameters contained in the transmission mechanism vibration signal and has good generalization performance.

[0033] 4. By calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix, the principal component features of the on-load tap changer transmission mechanism vibration signal are extracted, which simplifies the calculation and improves the calculation accuracy.

[0034] 5. The present invention also determines the mechanical state of the transmission mechanism through the weighted mean of the statistical elements and the corresponding control limits, and provides a quantitative evaluation standard for the mechanical state monitoring of the on-load tap changer transmission mechanism, which provides an important basis for the inspection and maintenance of the on-load tap changer. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is a flow chart of a method for monitoring an on-load tap changer transmission mechanism based on sparse filtering according to the present invention;

[0037] Figure 2 It is a vibration signal used to monitor the mechanical state of the on-load tap changer transmission mechanism in this embodiment. DETAILED DESCRIPTION

[0038] The following is a clear and complete description of the technical solutions of the present invention in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts are within the scope of protection of the present invention.

[0039] like Figure 1 As shown, the on-load tap changer transmission mechanism monitoring method based on sparse filtering provided by the present invention includes the following steps:

[0040] S1. Switch the on-load tap changer multiple times and collect the vibration signal v(t) of the transmission mechanism during the on-load tap changer switching process. The length of the vibration signal collected during one switching process is N0, and the sampling frequency is f s .

[0041] S2. Segment the vibration signal. The segmentation is to segment the vibration signal during each switching process according to the rotation period of the on-load tap changer transmission shaft to obtain M segments of vibration signals x(t). The rotation period of the on-load tap changer transmission shaft is T0, and the length of each segment of the vibration signal x(t) after segmentation is 2T0f. s .

[0042] S3, calculating the vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on the sparse filter network according to the vibration signal; specifically comprising the following steps:

[0043] S31. Determine the network structure and parameters of a sparse filtering network, wherein the sparse filtering network includes an input layer and an output layer, wherein the input layer is connected to the output layer via a weight matrix, and the parameters of the sparse filtering network include: the number of nodes M1 in the input layer, the number of nodes l in the output layer, and a weight matrix W, wherein the dimension of the weight matrix W is M1×l;

[0044] S32, perform L2 norm normalization on each segmented vibration signal x(t), and then construct a sparse filter sample set Here, M1 is the number of sample sets;

[0045] S33, randomly initialize the weight matrix W and the error precision ε, and construct the objective function of the sparse filtering network in combination with the sparse filtering sample set. The objective function of the sparse filtering network can be expressed as:

[0046]

[0047]

[0048]

[0049] In the above formula, represents the jth characteristic component of the i-th sample; w j represents the j-th column vector of the weight matrix W; ||·||1 and ||·||2 represent the 1-norm and 2-norm respectively; ε=10 -8 ;

[0050] S34, iteratively solving the objective function of the sparse filtering network based on the limited memory quasi-Newton method, stopping the iteration when the error precision ε is satisfied, and obtaining the required weight matrix W;

[0051] S35. Calculate the vibration signal characteristic distribution matrix F of the on-load tap changer transmission mechanism in combination with the weight matrix W. The vibration signal characteristic distribution matrix F is expressed as:

[0052]

[0053]

[0054] In the above formula, g(·) is the soft absolute value activation function, T represents the transpose, represents the jth feature component of the i-th sample, f l m represents the jth feature component of the i-th sample.

[0055] S4, calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix F; specifically comprising the following steps:

[0056] S41, performing singular value decomposition on the vibration signal characteristic distribution matrix F, and obtaining C singular values ​​λ1, λ2, ..., λ C and the corresponding eigenvectors ω1, ω2, …, ω C , and λ1≥λ2≥…≥λ C ;

[0057] S42, calculate the cumulative contribution rate of the matrix eigenvalues, where the cumulative contribution rate η of the kth eigenvalue is k The calculation formula is:

[0058]

[0059] S43, when the cumulative contribution rate of the kth eigenvalue is greater than the threshold, that is, η k>δ, select the eigenvectors v1, v2, …, v corresponding to the first k eigenvalues ​​of the kth eigenvalue k The eigenvalues ​​and eigenvectors of the principal components of the characteristic distribution matrix are denoted as P C =[ω1,ω2,…,ω C ] and Λ C =diag{λ1,λ2,…,λ C}.

[0060] S5, based on the principal component eigenvalue P of the characteristic distribution matrix F C and eigenvector Λ C Calculate the statistic T 2 The element-weighted mean of and the corresponding control limits τ; where,

[0061] The statistics are expressed as:

[0062] T 2 =P C (Λ C ) -1 (P C ) T (7)

[0063]

[0064]

[0065]

[0066] In the above formula, is a chi-square distribution with α degrees of freedom.

[0067] S6. Discerning the mechanical state of the transmission mechanism according to the weighted mean of the elements of the statistic and the corresponding control limits, thereby monitoring the transmission mechanism of the on-load tap changer, specifically comprising:

[0068] When the weighted mean of the elements When it is higher than the corresponding control limit τ, it is determined that the mechanical state of the on-load tap changer transmission mechanism has changed, and timely maintenance is required; otherwise, the mechanical state of the on-load tap changer transmission mechanism has not changed.

[0069] Taking a 110kV transformer on-load tap changer as the test object, the vibration signal during its switching process is tested. Based on this, the mechanical condition monitoring method of the on-load tap changer transmission mechanism is described to enable those skilled in the art to better understand the present invention:

[0070] S1. At the substation site, a vibration sensor is placed on top of the transformer on-load tap-changer. The vibration sensor is connected to the vibration signal acquisition system via a connecting line. The vibration signal v(t) of the transmission mechanism during the on-load tap-changer gear switching process is collected. Its length is N0 and the sampling frequency is f s ,like Figure 2 As shown, here, f s =50kHz; in step S3, ε=10 -8 ; In step S4, δ = 0.9; In step S5, α=3;τ=0.9205;According to the on-load tap changer transmission mechanism monitoring method based on sparse filtering, the element weighted mean of the statistic of the characteristic distribution matrix of the vibration signal of the on-load tap changer transmission mechanism is finally calculated. The value was 1.328, which was higher than the control limit of 0.9205, indicating that the mechanical condition of the on-load tapchanger transmission mechanism was abnormal. On-site maintenance revealed that the on-load tapchanger bevel gearbox was lacking oil, which effectively verified the effectiveness and accuracy of this method.

[0071] In summary, the present invention adopts the above-mentioned technical solution to accurately monitor the mechanical state of the transmission mechanism through the vibration signal of the transmission mechanism during the gear switching process of the on-load tap changer, thereby effectively identifying the hidden dangers of early faults of the on-load tap changer and then adopting effective operation and maintenance strategies to avoid major faults.

[0072] The above disclosure is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, and they should all be covered by the scope of protection of the present invention.

Claims

1. A method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering, characterized in that: The following steps are involved: Performing multiple gear switching on the on-load tap changer and collecting vibration signals of the transmission mechanism during the on-load tap changer switching process; Calculating a vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on a sparse filter network according to the vibration signal; Calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix; Calculating the element-weighted mean and corresponding control limits of the statistic based on the eigenvalues ​​and eigenvectors of the principal component components of the characteristic distribution matrix; The mechanical state of the transmission mechanism is judged according to the weighted mean of the elements of the statistic and the corresponding control limits, thereby realizing monitoring of the transmission mechanism of the on-load tap changer; Calculating the principal component eigenvalues ​​and eigenvectors of the vibration signal characteristic distribution matrix specifically includes: Performing singular value decomposition on the vibration signal characteristic distribution matrix to obtain singular values ​​and corresponding eigenvectors of the vibration signal characteristic distribution matrix; Calculate the cumulative contribution rate of matrix eigenvalues; When the first When the cumulative contribution rate of the eigenvalues ​​is greater than the threshold, select Before the eigenvalue The eigenvalues ​​corresponding to the eigenvectors are used as the principal component eigenvalues ​​and eigenvectors of the characteristic distribution matrix.

2. The on-load tap changer transmission mechanism monitoring method based on sparse filtering according to claim 1, characterized in that: Before calculating the vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on the sparse filter network according to the vibration signal, the vibration signal is segmented.

3. The method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering according to claim 2, characterized in that: The segmentation is to segment the vibration signal in each switching process according to the rotation period of the on-load tap changer transmission shaft.

4. The method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering according to claim 1, characterized in that: Calculating a vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism based on a sparse filter network according to the vibration signal specifically includes: Determine the network structure and parameters of the sparse filtering network; Normalize each segmented vibration signal and then construct a sparse filter sample set; Randomly initializing the weight matrix and error precision, and constructing the objective function of the sparse filtering network in combination with the sparse filtering sample set; Solve the objective function of the sparse filtering network and obtain the required weight matrix; The vibration signal characteristic distribution matrix of the on-load tap changer transmission mechanism is calculated in combination with the weight matrix.

5. The method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering according to claim 4, characterized in that: The sparse filtering network includes an input layer and an output layer, and the input layer is connected to the output layer through a weight matrix.

6. The method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering according to claim 4, characterized in that: The objective function of the sparse filtering network is iteratively solved based on the finite memory quasi-Newton method. When the error accuracy is met, the iteration is stopped and the required weight matrix is ​​obtained.

7. The method for monitoring the transmission mechanism of an on-load tap changer based on sparse filtering according to claim 1, characterized in that: The mechanical state of the transmission mechanism is judged according to the weighted mean of the elements of the statistic and the corresponding control limits, thereby realizing monitoring of the on-load tap changer transmission mechanism, specifically including: When the weighted mean of each element is higher than the corresponding control limit, it is determined that the mechanical state of the on-load tap changer transmission mechanism has changed, and timely maintenance is required; otherwise, the mechanical state of the on-load tap changer transmission mechanism has not changed.

Images