Method for discriminating transformer fault type based on fluorescence spectrum correction

Abstract

This invention, a transformer fault type discrimination method based on fluorescence spectroscopy correction, belongs to the field of transformer fault diagnosis technology. It addresses the problem of designing a simple, computationally efficient method for rapidly identifying transformer fault types, particularly thermal aging and discharge breakdown faults. The invention establishes a mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbons using a multivariate regression correction method. A multivariate correction filter is designed based on transmittance to calculate the aromatic hydrocarbon concentration in the transformer oil, replacing the fluorescence spectrometer. This not only reduces equipment costs and size but also improves the cost-effectiveness of transformer oil fault detection. Furthermore, based on the time-periodic characteristics of thermal aging and discharge breakdown faults, the rate of change in aromatic hydrocarbon concentration is used to distinguish the transformer fault type, greatly simplifying the calculation process and reducing computational load. This is of significant importance for extending the service life of transformers.

Description

Technical Field

[0001] This invention belongs to the field of transformer fault diagnosis technology, and relates to a method for identifying transformer fault types based on fluorescence spectroscopy correction. Background Technology

[0002] Transformers are the core of energy conversion in the process of power generation and distribution. They are numerous and have a wide impact. Their operating status directly affects the safe and reliable operation of the power system. Once a transformer has an accident, it will not only damage expensive electrical equipment, but also cause large-scale power outages, and even cause casualties, environmental pollution, and huge economic and social losses. Therefore, monitoring the operating status of transformers has become particularly important.

[0003] Transformer oil refers to a type of insulating oil used in oil-filled electrical equipment such as transformers, reactors, instrument transformers, bushings, and oil switches, serving as insulation, cooling, and arc-extinguishing agents. Transformer oil is a fractionation product of petroleum, and its main components are alkanes, cycloalkanes, aromatic unsaturated hydrocarbons, and non-hydrocarbon compounds. Transformer oil can fluoresce under ultraviolet or X-ray irradiation. Fluorescence refers to a photoluminescence phenomenon. When a substance at room temperature is irradiated with incident light of a certain wavelength (usually ultraviolet or X-rays), it absorbs the light energy, enters an excited state, and immediately de-excites, emitting outgoing light with a wavelength longer than the incident light (usually in the visible light range); and once the incident light stops, the luminescence disappears immediately. This property of outgoing light is called fluorescence.

[0004] Fluorescent detection technology (FMS) for transformer operation status analyzes changes in the optical signal of the transformer insulating oil through a fluorescent detection device, thereby achieving the purpose of monitoring the transformer. For example, the Chinese invention patent document "An Online Fluorescent Detection Device for Transformer Insulating Oil" published on July 13, 2021, with application publication number CN113109682A, discloses an online fluorescent detection device for transformer insulating oil that features high sensitivity, short analysis time, immunity to interference from surrounding magnetic and electric fields, and good stability and reproducibility, which can meet the online fault detection requirements of transformers in operation.

[0005] Transformer fault types mainly include thermal aging faults and discharge breakdown faults. Currently, the methods for applying fluorescence analysis to transformer fault diagnosis are all confined to the laboratory and are basically machine learning and deep learning methods. For example, the paper "Research on Transformer Insulation Aging Diagnosis and Fault Prediction Based on Big Data Analysis Method" (Li Jia, Changsha University of Science and Technology, 2017) takes a big data perspective. First, it proposes a data cleaning method with a double-loop structure based on the iterative verification method, processes DGA data, and establishes a data identification model for transformer fault types. Then, it uses the prediction method based on the maximum Lyapunov exponent to predict faults in the DGA time series. Existing technologies have high requirements for hardware, large computational load, and complex models. Summary of the Invention

[0006] The technical problem to be solved by the present invention is how to design a transformer fault type discrimination method based on fluorescence spectroscopy correction with a simple model and low computational cost, so as to realize the rapid discrimination of thermal aging faults and discharge breakdown faults of transformers.

[0007] The present invention solves the above-mentioned technical problems through the following technical solutions:

[0008] A transformer fault type discrimination method based on fluorescence spectroscopy correction includes the following steps:

[0009] S1. The mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbon compounds is established by using the multiple regression correction method, thereby obtaining the multiple regression correction coefficient;

[0010] S2. The positive and negative coefficients after standardization of the multivariate regression correction coefficients are used as the transmittance of the filter. Based on the transmittance, positive multivariate correction filters and negative multivariate correction filters are designed to calculate the aromatic hydrocarbon concentration of transformer oil.

[0011] S3. Based on the rate of change of transformer oil aromatic hydrocarbon concentration between two consecutive measurements, the thermal aging fault and discharge breakdown fault of the transformer are identified.

[0012] Furthermore, the mapping relationship described in step S1 is as follows:

[0013] The fluorescence spectrum of transformer oil is directly related to the concentration of aromatic hydrocarbons. The formula for calculating the concentration of aromatic hydrocarbons is derived using multiple linear regression:

[0014]

[0015] Where c is the concentration of aromatic hydrocarbons. The multivariate regression correction coefficients for the fluorescence spectra obtained in bands 1 to n are used. Here, represents the fluorescence spectrum of the first to nth bands, and b is the bias coefficient.

[0016] The vector form of the formula for calculating the concentration of aromatic hydrocarbon compounds using multiple linear regression correction is as follows:

[0017]

[0018] in, This is the vector of regression correction coefficients; = , T , is the fluorescence spectrum vector.

[0019] Furthermore, the calculation process for transmittance in step S2 is as follows:

[0020] Define half coefficient and , and They are respectively The positive and negative parts, that is:

[0021] ;

[0022] Find the maximum value of the half coefficient:

[0023]

[0024] Therefore, the transmittances of the positive and negative filters are obtained as follows:

[0025] ;

[0026] Where i = 1, 2, ..., n, and n is a natural number. The regression correction coefficient vector The i-th element.

[0027] Furthermore, the process for calculating the aromatic hydrocarbon concentration in transformer oil described in step S2 is as follows:

[0028] Assuming the current target fluorescence spectrum is Then, the energy received by the detector after passing through the positive multivariate correction filter and the negative multivariate correction filter is expressed as:

[0029] ;

[0030] According to the detector measurements and Calculate the aromatic hydrocarbon concentration in transformer oil under the current conditions. as follows:

[0031]

[0032] in, ; i = 1, 2, ..., n, where n is a natural number. The regression correction coefficient vector The i-th element.

[0033] Furthermore, the method for distinguishing between thermal aging faults and discharge breakdown faults of transformers based on the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements described in step S3 is as follows: Based on the time period characteristics of thermal aging faults and discharge breakdown faults, the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements under thermal aging faults and discharge breakdown faults is calculated respectively. The one with a slower rate of change of transformer oil aromatic hydrocarbon concentration is a thermal aging fault, and the one with a faster rate of change of transformer oil aromatic hydrocarbon concentration is a discharge breakdown fault.

[0034] The formula for calculating the rate of change of aromatic hydrocarbon concentration in transformer oil between two consecutive measurements is as follows:

[0035]

[0036] in, This represents the change in aromatic hydrocarbon concentration in transformer oil between two consecutive measurements. The time interval between two consecutive measurements. This represents the rate of change in aromatic hydrocarbon concentration in transformer oil.

[0037] The advantages of this invention are:

[0038] This invention establishes a mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbons using a multivariate regression correction method. Based on transmittance, a multivariate correction filter is designed to calculate the aromatic hydrocarbon concentration of transformer oil, replacing the fluorescence spectrometer. This not only reduces equipment costs and size but also improves the cost-effectiveness of transformer oil fault detection. Furthermore, based on the time periodic characteristics of thermal aging and discharge breakdown faults in transformer oil, the rate of change in aromatic hydrocarbon concentration is used to determine the type of transformer fault, greatly simplifying the calculation process and reducing the amount of computation. This is of great significance for extending the service life of transformers. Attached Figure Description

[0039] Figure 1 This is a flowchart of a transformer fault type discrimination method based on fluorescence spectroscopy correction according to an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram illustrating the transmittance calculation principle of the multi-element correction filter according to an embodiment of the present invention;

[0041] Figure 3This is a flowchart illustrating the design of the film system structure of the multi-element correction filter according to an embodiment of the present invention;

[0042] Figure 4 This is a fluorescence spectrum data diagram of the same transformer oil under thermal aging fault and discharge breakdown fault according to an embodiment of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0045] Example 1

[0046] like Figure 1 As shown, the transformer fault type discrimination method based on fluorescence spectroscopy correction includes the following steps:

[0047] Step 1: Establish the mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbon compounds using the multiple regression correction method, thereby obtaining the multiple regression correction coefficients.

[0048] The fluorescence spectrum of transformer oil is directly related to the concentration of aromatic hydrocarbons. The formula for calculating the concentration of aromatic hydrocarbons is derived using multiple linear regression:

[0049]

[0050] Where c is the concentration of aromatic hydrocarbons. The regression correction coefficients for the fluorescence spectra obtained in bands 1 to n are given. Here, represents the fluorescence spectrum of the first to nth bands, and b is the bias coefficient.

[0051] The vector form of the formula for calculating the concentration of aromatic hydrocarbon compounds using multiple linear regression correction is as follows:

[0052]

[0053] in, This is the vector of regression correction coefficients; = , T , is the fluorescence spectrum vector.

[0054] Step 2: Standardize the positive and negative coefficients of the multivariate regression correction coefficients and use them as the transmittance of the filter. Based on the transmittance, design positive and negative multivariate correction filters to calculate the aromatic hydrocarbon concentration in transformer oil.

[0055] like Figure 2 As shown, the regression correction coefficient vector The positive and negative parts are used to create two correction filters, namely the regression correction coefficient vector. The standardized positive and negative coefficients are used as the transmittance of the filter;

[0056] Define half coefficient and , and They are respectively The positive and negative parts, that is:

[0057] ;

[0058] Find the maximum value of the half coefficient:

[0059]

[0060] Therefore, the transmittances of the positive and negative filters are obtained as follows:

[0061] ;

[0062] Assuming the current target fluorescence spectrum is Then, the energy received by the detector after passing through the positive multivariate correction filter and the negative multivariate correction filter is expressed as:

[0063] ;

[0064] According to the detector measurements and Calculate the aromatic hydrocarbon concentration in transformer oil. as follows:

[0065]

[0066] in, ; i = 1, 2, ..., n, where n is a natural number. The regression correction coefficient vector The i-th element, For vectors The i-th element, For vectors The i-th element.

[0067] like Figure 3 As shown, the design process of the membrane structure is as follows: The transmittance vector t is divided into two parts according to positive and negative values, forming a positive vector A and a negative vector B. The absolute value of the negative vector B is taken to obtain vector B2. The positive and negative filters are designed with vector A and vector B2 as targets respectively. First, the initial membrane structure of the filter is selected. Then, the thickness of each membrane layer is changed, and it is judged whether the similarity between the transmittance of the filter and vector A and vector B2 reaches the threshold. If it does, the design of the positive and negative filters is completed. If not, it is judged whether the iteration has reached the predetermined number of times. If the iteration has not reached the predetermined number of times, the thickness of each membrane layer is changed again. If the iteration has reached the predetermined number of times, the number of membrane layers is increased and the initial membrane structure of the filter is selected again.

[0068] Step 3: Based on the rate of change of transformer oil aromatic hydrocarbon concentration between two consecutive measurements, determine whether the transformer is experiencing thermal aging or discharge breakdown faults.

[0069] like Figure 4 The image shows the fluorescence spectral data of the same transformer oil under thermal aging fault and discharge breakdown fault. It can be seen that the characteristic peak positions of the fluorescence spectrum are consistent under different faults, and the fluorescence intensity overlaps. Therefore, it is impossible to directly distinguish the transformer fault type by fluorescence intensity and the position of the characteristic peak of the fluorescence spectrum.

[0070] This embodiment determines the transformer fault type by calculating the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements under thermal aging faults and discharge breakdown faults (i.e., the time period of thermal aging faults is relatively long, generally in the form of weeks or months, while the time period of discharge breakdown faults is very short, generally in the form of seconds or minutes). The transformer fault type is determined by calculating the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements under thermal aging faults and discharge breakdown faults. The transformer fault type is determined by the rate of change of transformer oil aromatic hydrocarbon concentration when the rate of change is slow, and the transformer fault type is determined by the rate of change of transformer oil aromatic hydrocarbon concentration when the rate of change is fast.

[0071] The formula for calculating the rate of change of aromatic hydrocarbon concentration in transformer oil between two consecutive measurements is as follows:

[0072]

[0073] in, This represents the change in aromatic hydrocarbon concentration in transformer oil between two consecutive measurements. The time interval between two consecutive measurements. This represents the rate of change in aromatic hydrocarbon concentration in transformer oil.

[0074] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying transformer fault types based on fluorescence spectroscopy correction, characterized in that, Includes the following steps: S1. The mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbon compounds is established by using the multiple regression correction method, thereby obtaining the multiple regression correction coefficient; S2. The positive and negative coefficients after standardization of the multivariate regression correction coefficients are used as the transmittance of the filter. Based on the transmittance, positive multivariate correction filters and negative multivariate correction filters are designed to calculate the aromatic hydrocarbon concentration of transformer oil. S3. Based on the rate of change of transformer oil aromatic hydrocarbon concentration between two consecutive measurements, the thermal aging fault and discharge breakdown fault of the transformer are identified.

2. The transformer fault type discrimination method based on fluorescence spectroscopy correction according to claim 1, characterized in that, The mapping relationship described in step S1 is as follows: The fluorescence spectrum of transformer oil is directly related to the concentration of aromatic hydrocarbons. The formula for calculating the concentration of aromatic hydrocarbons is derived using multiple linear regression: Where c is the concentration of aromatic hydrocarbons. The multivariate regression correction coefficients for the fluorescence spectra obtained in bands 1 to n are used. Here, represents the fluorescence spectrum of the first to nth bands, and b is the bias coefficient. The vector form of the formula for calculating the concentration of aromatic hydrocarbon compounds using multiple linear regression correction is as follows: in, This is the vector of regression correction coefficients; = , T , is the fluorescence spectrum vector.

3. The transformer fault type discrimination method based on fluorescence spectroscopy correction according to claim 2, characterized in that, The calculation process for transmittance in step S2 is as follows: Define half coefficient and , and They are respectively The positive and negative parts, that is: ； Find the maximum value of the half coefficient: Therefore, the transmittances of the positive and negative filters are obtained as follows: ； Where i = 1, 2, ..., n, and n is a natural number. The regression correction coefficient vector The i-th element.

4. The transformer fault type discrimination method based on fluorescence spectroscopy correction according to claim 3, characterized in that, The process for calculating the aromatic hydrocarbon concentration in transformer oil described in step S2 is as follows: Assuming the current target fluorescence spectrum is Then, the energy received by the detector after passing through the positive multivariate correction filter and the negative multivariate correction filter is expressed as: ； According to the detector measurements and Calculate the aromatic hydrocarbon concentration in transformer oil under the current conditions. as follows: in, ; i = 1, 2, ..., n, where n is a natural number. The regression correction coefficient vector The i-th element.

5. The transformer fault type discrimination method based on fluorescence spectroscopy correction according to claim 1, characterized in that, The method for distinguishing between thermal aging faults and discharge breakdown faults of transformers based on the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements described in step S3 is as follows: Based on the time period characteristics of thermal aging faults and discharge breakdown faults, the rate of change of transformer oil aromatic hydrocarbon concentration between two adjacent measurements under thermal aging faults and discharge breakdown faults is calculated respectively. The one with a slower rate of change of transformer oil aromatic hydrocarbon concentration is thermal aging fault, and the one with a faster rate of change of transformer oil aromatic hydrocarbon concentration is discharge breakdown fault. The formula for calculating the rate of change of aromatic hydrocarbon concentration in transformer oil between two consecutive measurements is as follows: in, This represents the change in aromatic hydrocarbon concentration in transformer oil between two consecutive measurements. The time interval between two consecutive measurements. This represents the rate of change in aromatic hydrocarbon concentration in transformer oil.

Images