Transformer invisible fault diagnosis method, device, equipment and medium

Abstract

The embodiment of the invention discloses a transformer invisible fault diagnosis method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring voltage and current recording data of a transformer measured on line; performing influence factor elimination operation on the voltage and current recording data, specifically comprising the following steps of: eliminating a higher harmonic component, a direct current component and an attenuation direct current component by adopting a full-wave differential Fourier algorithm to obtain a fundamental wave voltage and a fundamental wave current; when the elliptical orbit is calculated, normalization processing is carried out on different loads; eliminating the influence of abnormal numerical values by adopting a Gaussian regression algorithm; constructing a voltage difference and current elliptic track by using a transformer port relational expression compensated by the exciting current; correcting the elliptical trajectory by using the primary side voltage data, the secondary side voltage data and the primary side current data of the two groups of transformers measured within the sampling time; and after the operation of eliminating the influence factors, realizing the recognition and diagnosis of the invisible fault of the transformer through Lissajous ellipse fitting, characteristic quantity extraction and an artificial intelligence inversion algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer fault monitoring, and in particular to a transformer invisible fault diagnosis method, device, electronic equipment and storage medium. Background Art

[0002] Transformers are the most core power transmission and transformation equipment in the power system, especially large power transformers, which play an important role in the process of electric energy transmission.

[0003] Winding is an important part of the transformer. Winding deformation failure is the main factor affecting the transformer's ability to resist short circuits. Therefore, the state of winding deformation is a direct reflection of the transformer's ability to resist short circuits. Factors such as collision, friction, explosion of dissolved gas in oil, cracking, corrosion and earthquake during the transportation of the transformer may cause the winding to deform to varying degrees. In particular, various fault currents and closing surge currents in the power system generate huge electric forces that act on the windings many times, making the winding deformation have a cumulative effect, resulting in a reduction in the transformer's ability to resist short circuits. With the occurrence of transformer outlet short circuits, winding deformation may develop into a catastrophic failure, which will lead to equipment shutdown and power outages in the power grid, resulting in unpredictable economic losses. In addition, once a transformer fails, it generally takes more than several months to overhaul it, which not only consumes a lot of manpower and material resources, but also seriously affects the reliability of the system. Therefore, it is necessary to strengthen the detection of power transformer winding deformation failures to reduce or even prevent accidents.

[0004] Winding deformation fault is the main factor affecting the short-circuit resistance of the transformer. The degree, type and location of winding deformation will determine the upper and lower limits of the transformer's short-circuit resistance. Therefore, there are many studies on winding deformation fault detection at home and abroad. The most intuitive way to detect transformer winding deformation is to check the cover. As a means of detecting early winding deformation, this method consumes a lot of material and financial resources, and it is difficult to judge the condition of the inner winding. Later, domestic and foreign experts successively developed a variety of non-destructive detection methods for transformer winding deformation. There are mainly short-circuit impedance method, capacitance change method, low-voltage pulse method, frequency response analysis method, vibration analysis method, online frequency response method, etc. However, most of the above methods are offline methods, which need to be implemented after the transformer is shut down, and cannot reflect the transformer winding status in time, which is relatively passive; some other methods of online sensing of transformer winding status have low recognition accuracy due to many influencing factors. Summary of the invention

[0005] The main purpose of the present invention is to provide a transformer invisible fault diagnosis method, device, electronic equipment and storage medium, which can achieve the accuracy of transformer fault voltage-current trajectory method diagnosis, thereby realizing online application.

[0006] To achieve the above objectives, the present application provides a transformer invisible fault diagnosis method, the method comprising:

[0007] Obtain voltage and current recording data of the transformer measured online;

[0008] The operation of eliminating influencing factors on the voltage and current recording data of the transformer specifically includes:

[0009] Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current;

[0010] Based on the fundamental voltage and current, when calculating the elliptical trajectory, normalization processing is performed on different loads;

[0011] Based on the fundamental voltage and current, a Gaussian regression algorithm is used to eliminate the influence of abnormal values;

[0012] Based on the fundamental voltage and current, a transformer port relationship after compensation with the excitation current is used to construct a voltage difference and current elliptical trajectory;

[0013] Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured during sampling time are used to correct the voltage difference and current elliptical trajectory;

[0014] After the operation of eliminating the influencing factors, a processed voltage and current signal is obtained; the processed voltage and current signal is used to realize the identification and diagnosis of transformer invisible faults through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

[0015] 3. Optionally, the normalizing of different loads includes:

[0016] According to the principle of winding equivalent two-port network circuit, different loads are normalized, and the normalized relationship is:

[0017]

[0018] in and The load is Z' L The primary and secondary current values ​​at and According to the circuit principle As a voltage source and A joint response.

[0019] Optionally, the using of a Gaussian regression algorithm to eliminate the influence of abnormal values ​​includes:

[0020] Use Gaussian process regression to fit the recorded data points and obtain the confidence interval of the recorded curve;

[0021] When a recorded data point is found to be beyond the confidence interval, it is determined to be an abnormal data point and the point is corrected;

[0022] Repeat the above steps until there are no abnormal data points.

[0023] Optionally, the correcting the point includes:

[0024] Replace the value at a point with the mean of the values ​​before and after the point.

[0025] Optionally, the transformer port relationship after compensation using the excitation current is:

[0026] U1-U2+αU1=(1+α)U1-U2=(R k +jX k )I1,

[0027] Among them, R k and X k are the short-circuit resistance and short-circuit reactance of the transformer respectively, and α is the compensation coefficient of the transformer;

[0028] The method of constructing the voltage difference and current elliptical trajectory comprises:

[0029] The elliptical trajectory is constructed with (1+α)U1-U2 and I1.

[0030] Optionally, the transformer port relationship after compensation using the excitation current is:

[0031] U1-U2+αU1=(1+α)U1-U2=(R k +jX k )I1,

[0032] Among them, R k and X k are the short-circuit resistance and short-circuit reactance of the transformer respectively, and α is the compensation coefficient of the transformer;

[0033] The method of constructing the voltage difference and current elliptical trajectory comprises:

[0034] The elliptical trajectory is constructed with (1+α)U1-U2 and I1.

[0035] Optionally, the voltage and current recording data include steady-state and transient recording data of voltage and current;

[0036] After obtaining the voltage and current waveform data of the transformer measured online, the method further includes:

[0037] Discrete wavelet noise reduction processing is performed based on the voltage and current recording data of the transformer.

[0038] On the other hand, the present application also provides a transformer invisible fault diagnosis device, comprising:

[0039] An acquisition module is used to acquire voltage and current recording data of the transformer measured online;

[0040] The elimination module is used to eliminate the influencing factors of the voltage and current recording data of the transformer, specifically for:

[0041] Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current;

[0042] Based on the fundamental voltage and current, when calculating the elliptical trajectory, normalization processing is performed on different loads;

[0043] Based on the fundamental voltage and current, a Gaussian regression algorithm is used to eliminate the influence of abnormal values;

[0044] Based on the fundamental voltage and current, a transformer port relationship after compensation with the excitation current is used to construct a voltage difference and current elliptical trajectory;

[0045] Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured during sampling time are used to correct the voltage difference and current elliptical trajectory;

[0046] The identification and diagnosis module is used to obtain processed voltage and current signals after the operation of eliminating the influencing factors; use the processed voltage and current signals to realize the identification and diagnosis of transformer invisible faults through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

[0047] On the other hand, the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the first aspect and any possible implementation method thereof.

[0048] On the other hand, the present application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes each step of the method described in the first aspect.

[0049] The present application provides a transformer invisible fault diagnosis method, device, electronic device and storage medium, which obtains the voltage and current recording data of the transformer measured online; and performs operations to eliminate influencing factors on the voltage and current recording data of the transformer, specifically including: based on the voltage and current recording data of the transformer, using a full-wave differential Fourier algorithm to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current; based on the fundamental voltage and current, when calculating the elliptical trajectory, normalizing different loads; based on the fundamental voltage and current, using a Gaussian regression algorithm to eliminate the influence of abnormal values; based on the fundamental voltage and current, using the transformer port relationship after compensation with excitation current to construct a voltage difference and current elliptical trajectory; based on the fundamental voltage and current, using two groups of transformer primary and secondary voltage and primary current data measured within the sampling time, correcting the voltage difference and current elliptical trajectory; after the operation of eliminating the influencing factors, a processed voltage and current signal is obtained; using the processed voltage and current signal, the identification and diagnosis of the transformer's invisible fault is realized through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm; the method in the present application can realize timely and accurate diagnosis of the transformer's invisible fault under live operation, and at the same time realize the evaluation of the short-circuit resistance of the operating transformer. The method avoids the fact that a large number of transformers that already exist in the power grid system and are widely distributed are difficult to realize online detection and evaluation of invisible defect faults such as early winding deformation, or that there are many influencing factors that lead to low online detection sensitivity, and regular power outages for maintenance will waste a lot of manpower and expenses. At the same time, potential abnormalities of reduced short-circuit resistance of the transformer can also be discovered, and replacement strategies can be adopted in time to reduce economic losses for the company and ensure the stability, reliability and safety of the power system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0051] in:

[0052] Figure 1 A schematic diagram of a flow chart of a transformer invisible fault diagnosis method provided in an embodiment of the present application;

[0053] Figure 2 A schematic flow chart of a method for eliminating influencing factors in the diagnosis of transformer invisible faults using a voltage-current trajectory method provided in an embodiment of the present application;

[0054] Figure 3A flowchart of a method for detecting and correcting abnormal recorded wave data provided in an embodiment of the present application;

[0055] Figure 4 A schematic diagram of the structure of a transformer invisible fault diagnosis device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0057] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0058] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0059] The embodiments of the present application involve diagnosis of invisible faults of transformers. Invisible faults of transformers generally refer to those faults that are not easily observed or detected directly. The diagnosis of such faults is crucial to ensure safe and stable operation of the power system.

[0060] The voltage-current trajectory method mentioned in the embodiment of the present application is a method for fault analysis and fault location in a power system. It can effectively identify the location and type of the fault by analyzing the changing trajectory of the voltage and current in the power system, and help engineers take corresponding measures in time to repair and restore power supply. The principle of this method is based on the fault propagation characteristics of the power system and the relationship between voltage and current. By collecting voltage and current data at each key node of the power system and analyzing it through mathematical models and algorithms, trajectory diagrams of voltage and current can be obtained. These trajectory diagrams can reflect the propagation path of the current in the power system and the degree of influence of the fault point, so as to determine the location and type of the fault.

[0061] In the embodiment of the present application, in the transformer fault diagnosis, the voltage-current trajectory method can be used to analyze transformer winding deformation and other faults. By installing signal acquisition equipment on the secondary side of the voltage transformer (PT) and the current transformer (CT) at the transformer outlet and measuring the voltage and current signals, a characteristic graph of voltage and current is constructed on this basis, and the online detection of transformer winding deformation can be achieved through the change of the graph. This method is easy to implement and does not affect the safe operation of the transformer itself, and has strong engineering practicality.

[0062] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0063] See also Figure 1 , is a flow chart of a transformer invisible fault diagnosis method provided in an embodiment of the present application. Figure 1 As shown, the method includes:

[0064] 101. Obtain voltage and current recording data of the transformer measured online.

[0065] In the embodiment of the present application, the three-phase voltage and three-phase current of the transformer can be measured online, that is, the voltage and current fault recording can be performed to obtain voltage and current recording data.

[0066] The online measurement of the three-phase voltage and three-phase current of the transformer mentioned in the embodiments of the present application generally involves the following steps:

[0067] Choose the right measuring equipment: According to the search results, you can use professional equipment such as Megger test equipment, CT3500 transformer characteristic parameter tester, GDBR-V transformer capacity and no-load tester for measurement. These devices usually have high-precision sampling and digital signal processing technology, which can meet the testing needs of power transformers.

[0068] Connect the test equipment: Connect the voltage and current probes of the test equipment to the corresponding terminals of the transformer. For a three-phase transformer, connect them to the three phase wires respectively.

[0069] Set test parameters: Set the parameters of the test equipment according to the rated voltage and current of the transformer. Some equipment such as the CT3500 transformer characteristic parameter tester allows the user to enter the specific parameters of the transformer to ensure the accuracy of the test results.

[0070] Take measurements: Start the test equipment and let it start measuring three-phase voltage and current. The equipment will display real-time voltage and current values, and some equipment, such as the GDBR-V transformer capacity and no-load tester, can also record and analyze data.

[0071] Optionally, analysis can be performed based on the obtained data, which may include but is not limited to the following steps:

[0072] Data analysis: After the measurement is completed, analyze the obtained data. For three-phase transformers, it is necessary to ensure the balance of three-phase voltage and current. If an imbalance is found, further inspection and maintenance may be required.

[0073] Error analysis: During the measurement process, measurement errors need to be considered, such as the measurement errors of PT and CT. The random error can be reduced by taking the average value through multiple acquisitions, so that the total error is controlled within an acceptable range.

[0074] Use professional software: Some test equipment, such as the WP4000 variable frequency power analyzer, can be used with special software to automatically perform multiple corrections, such as waveform correction, temperature correction, etc., to improve the accuracy of the test results.

[0075] Harmonic analysis: If necessary, harmonic analysis can also be performed to evaluate the quality of the power grid. The WP4000 variable frequency power analyzer can measure the content rate of multiple harmonics and the total distortion rate.

[0076] The technical solution of the embodiment of the present application is mainly to obtain the voltage and current recording data of the primary and secondary sides of the transformer online, and then adopt a series of algorithms and methods based on the recording data to eliminate influencing factors, so as to achieve accurate identification and diagnosis of invisible faults such as transformer winding deformation and inter-turn short circuit.

[0077] In an optional implementation, the voltage and current recording data include steady-state and transient recording data of voltage and current;

[0078] After step 101, the method further includes:

[0079] Discrete wavelet denoising processing is performed based on the voltage and current recording data of the above transformer.

[0080] The discrete wavelet denoising mentioned in the embodiment of the present application is a signal processing technology based on wavelet transform, which is used to remove the noise components in the signal. Its basic principle is to use wavelet transform to decompose the signal into frequency bands of different scales, including low-frequency part and high-frequency part. Since noise usually exists in the high-frequency part, the noise can be suppressed by thresholding the high-frequency wavelet coefficients. The goal of thresholding is to retain the wavelet coefficients mainly controlled by the signal, while removing the wavelet coefficients controlled by the noise. Then, the processed wavelet coefficients are inversely transformed, that is, the denoised signal is reconstructed.

[0081] Through discrete wavelet denoising, equipment noise can be filtered out.

[0082] 102. Eliminate the influencing factors of the voltage and current recording data of the transformer.

[0083] The influences that need to be eliminated in the embodiments of the present application mainly include: load fluctuation influence, on-site noise influence, high-order harmonic component and DC component influence, excitation current influence, transformer measurement error influence, electromagnetic interference abnormal data influence, etc. By eliminating these influencing factors, the accuracy of transformer fault identification and diagnosis can be improved.

[0084] The above step 102 may specifically include:

[0085] Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current;

[0086] Based on the fundamental voltage and current, normalization is performed for different loads when calculating the elliptical trajectory;

[0087] Based on the fundamental voltage and current mentioned above, Gaussian regression algorithm is used to eliminate the influence of abnormal values;

[0088] Based on the fundamental voltage and current, the voltage difference and current elliptical trajectory are constructed using the transformer port relationship after compensation with the excitation current;

[0089] Based on the fundamental voltage and current, the voltage difference and current elliptical trajectory are corrected by using two sets of transformer primary and secondary voltage and primary current data measured within the sampling time.

[0090] Specifically, the elimination operations for the above-mentioned influencing factors may mainly include:

[0091] The influence of load fluctuation is eliminated by load normalization method, the influence of on-site noise is eliminated by discrete wavelet noise reduction method, the influence of high-order harmonic components and DC components is eliminated by full-wave differential Fourier decomposition algorithm, the influence of excitation current is eliminated by excitation current compensation algorithm, the influence of mutual inductor measurement error is eliminated by two sampling value subtraction filtering method, and the influence of abnormal electromagnetic interference data is eliminated by Gaussian regression algorithm fitting and correction method.

[0092] When the influencing factors are subjected to a series of measures such as noise reduction, filtering, compensation, fitting, and correction to obtain available voltage and current signals, step 103 can be executed, and the identification and diagnosis of invisible faults such as transformer winding deformation and inter-turn short circuit can be realized through Lissajous ellipse fitting, feature extraction, and artificial intelligence inversion algorithm.

[0093] In an optional implementation, the data window length used by the full-wave Fourier algorithm is a period of the power frequency, which can completely filter out all harmonics and pure DC components. If there is an attenuated DC component in the signal, it will interfere with the accuracy of the Fourier algorithm. In order to filter out the attenuated DC component in the embodiment of the present application, the full-wave Fourier algorithm is improved, that is, the full-wave differential Fourier algorithm. The amount of data required to be provided by the improved algorithm is increased by two sampling points on the basis of the amount of data required to be provided by the original full-wave Fourier algorithm.

[0094] In an optional implementation manner, the above-mentioned normalization processing of different loads includes:

[0095] According to the principle of winding equivalent two-port network circuit, different loads are normalized. The relationship of the normalization is:

[0096]

[0097] in and The load is Z' L The primary and secondary current values ​​at and According to the circuit principle As a voltage source and A joint response.

[0098] Specifically, based on the fundamental voltage and current obtained above, when calculating the elliptical trajectory, the change in the secondary side load size (load fluctuation) will cause the major axis and minor axis of the elliptical trajectory to increase or decrease proportionally, thereby affecting the diagnosis of the transformer's invisible fault. To this end, in the embodiment of the present application, different loads can be normalized according to the principle of the winding equivalent two-port network circuit, and a method for modifying the Lissajous figure is proposed by considering the impact of load changes in real time, thereby avoiding the impact of load changes. The normalization relationship is:

[0099]

[0100] in and The load is Z' L The primary and secondary current values ​​at and According to the circuit principle As a voltage source and A joint response.

[0101] In order to eliminate the influence of load changes on the elliptical trajectory in the embodiment of the present application, a load normalization method can be used, which involves normalizing the measured voltage and current data to standard load conditions, so that the trajectory diagram is not affected by the actual load size.

[0102] In an optional implementation, the Gaussian regression algorithm is used to eliminate the influence of abnormal values, including:

[0103] Use Gaussian process regression to fit the recorded data points and obtain the confidence interval of the recorded curve;

[0104] When a recorded data point is found to be beyond the above confidence interval, it is determined to be an abnormal data point and the point is corrected;

[0105] Repeat the above steps until there are no abnormal data points.

[0106] Optionally, the above correction to this point includes:

[0107] Replace the value at a point with the mean of the values ​​before and after the point.

[0108] In the embodiment of the present application, the above correction method can be adjusted as needed. The value of the point can be replaced by the mean of the two points before and after the abnormal data point, or the mean of multiple points before and after can be used. A specific formula (including weight parameters, etc.) can be used or set to calculate and adjust the abnormal data point, and the embodiment of the present application does not limit this.

[0109] Among them, the above method steps of using Gaussian regression algorithm to eliminate the influence of abnormal values ​​can be referred to Figure 3 Detailed description of the illustrated embodiments.

[0110] In an optional implementation manner, the transformer port relationship after the above compensation using the excitation current is:

[0111] U1-U2+αU1=(1+α)U1-U2=(R k +jX k )I1,

[0112] Among them, R k and X k are the short-circuit resistance and short-circuit reactance of the above transformer respectively, and α is the compensation coefficient of the above transformer;

[0113] The above construction of voltage difference and current elliptical trajectory includes:

[0114] The above elliptical trajectory is constructed using (1+α)U1-U2 and I1.

[0115] Specifically, based on the fundamental voltage and current obtained above, a ΔU-I1 trajectory can be constructed based on the transformer port relationship after compensation of the excitation current to eliminate the influence of the excitation current. Where ΔU=U1-U2, the transformer port relationship after compensation is U1-U2+αU1=(1+α)U1-U2=(R k +jX k )I1, where R k and X k are the short-circuit resistance and short-circuit reactance of the transformer, α=X2 / X m is the compensation coefficient, which can be obtained based on the nameplate data of the transformer when it leaves the factory. Thus, (1+α)U1-U2 and I1 can construct an elliptical trajectory to eliminate the influence of the excitation current.

[0116] In an embodiment of the present application, the processed data can be used to construct an elliptical trajectory based on the relationship between voltage and current. This usually involves representing the phasor diagram of voltage and current, where the phasor endpoints of voltage and current describe an elliptical trajectory on the complex plane. Based on the constructed elliptical trajectory, feature analysis can be performed: analyzing the features of the elliptical trajectory, such as the major axis, minor axis, and tilt angle, etc., these features can reflect the type and location of faults in the power system.

[0117] In an optional implementation, based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured within the sampling time are used to correct the voltage difference and current elliptical trajectory, including:

[0118] Based on the fundamental voltage and current, the two sets of transformer primary and secondary voltage and primary current data measured within the sampling time are used to correct the elliptical trajectory constructed with ΔU and I1 or the elliptical trajectory constructed with (1+α)U1-U2 and I1, where ΔU=U1-U2.

[0119] Specifically, based on the fundamental voltage and current obtained above, two sets of transformer primary and secondary side voltage and primary side current data measured within a shorter sampling time can be used to correct the ΔU-I1 elliptical trajectory or the elliptical trajectory of (1+α)U1-U2 and I1, thereby reducing or eliminating the transformer measurement error.

[0120] The total measurement error of the transformer is mainly composed of random errors composed of free noise, system errors, and measurement deviations caused by abnormal operation of components. Since the error of the transformer is accumulated during a long period of operation, it is mainly affected by factors such as temperature, humidity, and electromagnetic environment. In a short sampling time, these external environments change very little. Therefore, if the two sets of primary and secondary voltage and primary current data measured in a short sampling time are and and and Then the true values ​​of the two differences of the primary and secondary voltage differences and the two differences of the primary current can be expressed as Where Z k The value is calculated based on the measured value of the transformer, namely:

[0121]

[0122] To a certain extent, the interference of measurement errors on calculation results is reduced.

[0123] 103. After the above operation of eliminating the influencing factors, the processed voltage and current signals are obtained; the processed voltage and current signals are used to realize the identification and diagnosis of the transformer invisible fault through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

[0124] The Lissajous ellipse fitting mentioned in the embodiments of the present application is a commonly used technique in the field of optical measurement and signal processing, especially in phase measurement and vibration analysis. This method is implemented by analyzing the trajectory synthesized by two orthogonal sinusoidal signals that may have different frequencies and phase differences. Mathematically, the Lissajous curve is defined by parametric equations that describe the synthetic trajectory of two sinusoidal vibrations in perpendicular directions.

[0125] The general steps of Lissajous ellipse fitting may include:

[0126] Data acquisition: Obtain the interferogram or vibration signal containing the desired information.

[0127] Create Lissajous Figures: Create Lissajous figures by analyzing the orthogonal components of a signal.

[0128] Ellipse fitting: Use an ellipse fitting algorithm (such as the least squares method) to determine the parameters of the ellipse, such as the center, major axis, minor axis, and tilt angle.

[0129] Parameter extraction: Extract phase and other relevant information from the fitted ellipse.

[0130] Error Analysis: Evaluate the accuracy of the fit and make necessary corrections.

[0131] The advantage of this method is that it can provide high-precision phase extraction and has good tolerance for local changes in the signal. In addition, through the improved algorithm, the demodulation capability of weak signals can be improved, thereby improving the resolution of the measurement system.

[0132] The feature extraction and artificial intelligence inversion algorithms in the embodiments of the present application can be set or selected as needed and are not limited here.

[0133] After eliminating the influencing factors based on a series of algorithms and methods such as noise reduction, filtering, compensation, fitting, and correction in the above steps, the available voltage and current signals are obtained. Then, through Lissajous ellipse fitting, feature extraction, and artificial intelligence inversion algorithm, the identification and diagnosis of invisible faults such as transformer winding deformation and turn-to-turn short circuit can be realized.

[0134] The offline method currently used needs to be implemented after the transformer is shut down, and cannot reflect the transformer winding status in time, which is relatively passive. Among them, although the vibration analysis method and the online frequency response method can be applied online, they have low sensitivity for detecting early latent faults such as slight deformation of transformer windings, inter-turn short circuits, inter-layer short circuits, etc., and their anti-interference ability is insufficient, and there are safety hazards in installation.

[0135] Invisible faults in transformer windings can cause changes in the primary and secondary voltage and current of the transformer, or more precisely, changes in the trajectory of the voltage-current characteristic spectrum. Therefore, the existing PT, CT and other components of the transformer are used to actively record the transformer port data, fit the Lissajous ellipse diagram using the port voltage and current data, extract the characteristic parameters of the ellipse diagram, and sense the transformer winding status online, so that transformer-affecting faults can be identified and diagnosed online, and the ability of the transformer winding to withstand short-circuit current can be evaluated, thereby adjusting the transformer replacement strategy. However, this technology has not been widely used today, mainly because many influencing factors lead to low recognition accuracy.

[0136] The method in the embodiment of the present application can provide a method for eliminating multiple influencing factors through online measurement of the three-phase voltage and three-phase current of the transformer, and inversion of invisible faults such as transformer winding deformation and turn-to-turn short circuit through the characteristic parameters of the Lissajous ellipse diagram (voltage-current trajectory method), thereby achieving the accuracy of transformer fault voltage-current trajectory method diagnosis, thereby realizing the online application of this method.

[0137] You can refer to Figure 2 The flowchart of a method for eliminating influencing factors in the diagnosis of transformer invisible fault voltage-current trajectory method is shown in FIG. Figure 2 As shown, the embodiment of the present application provides a method for eliminating influencing factors such as noise, electromagnetic interference, higher harmonics in voltage and current signals, load fluctuations, excitation current, and measurement errors of transformers under actual working conditions. First, the voltage and current recording data of the primary and secondary sides of the transformer are obtained online, and then a series of algorithms and methods are adopted based on the recording data to eliminate the influencing factors, wherein the load fluctuation influence is eliminated by the load normalization method, the field noise influence is eliminated by the discrete wavelet noise reduction method, the higher harmonic component and the DC component influence are eliminated by the full-wave differential Fourier decomposition algorithm, the excitation current influence is eliminated by the excitation current compensation algorithm, the mutual inductor measurement error influence is eliminated by the two sampling value subtraction filtering method, and the electromagnetic interference abnormal data influence is eliminated by the Gaussian regression algorithm fitting and correction method. When the influencing factors are subjected to a series of measures such as noise reduction, filtering, compensation, fitting, and correction, the available voltage and current signals are obtained, and then the transformer winding deformation, inter-turn short circuit and other invisible faults are recognized and diagnosed through Lissajous ellipse fitting, feature extraction, and artificial intelligence inversion algorithm.

[0138] Compared with the existing transformer invisible fault diagnosis technology based on the voltage-current trajectory method, the method in the embodiment of the present application can realize the timely and accurate diagnosis of the transformer invisible fault under live operation, and at the same time realize the evaluation of the short-circuit resistance of the transformer in operation. Therefore, the method in the embodiment of the present application avoids the problem that it is difficult to realize the online detection and evaluation of invisible defect faults such as early winding deformation, or the numerous influencing factors lead to low online detection sensitivity, and regular power outages and maintenance will waste a lot of manpower and expenses. At the same time, it can also find the potential abnormality of the transformer's reduced short-circuit resistance, and take replacement strategies in time to reduce economic losses for the company and ensure the stability, reliability and safety of the power system operation.

[0139] See also Figure 3 , is a flow chart of a method for detecting and correcting abnormal recorded wave data provided in an embodiment of the present application, such as Figure 3 As shown, the method includes:

[0140] 301. Read voltage and current recording data;

[0141] 302. Waveform recording curve regression fitting (Gaussian regression algorithm);

[0142] 303. Obtain the confidence interval of the recording curve;

[0143] 304. Determine whether any data point exceeds the limit according to the confidence interval; if so, execute step 305; if not, execute step 306;

[0144] 305. Record the data point that exceeds the limit, and execute step 306;

[0145] 306. Determine whether there is an out-of-limit data point mark; if so, execute step 307; if not, end;

[0146] 307. Replace the out-of-limit data points with the mean before and after.

[0147] In the embodiment of the present application, a Gaussian regression algorithm can be used to eliminate the influence of abnormal values ​​such as on-site interference. The core idea of ​​this method is to use Gaussian process regression to fit the recorded data points and obtain the confidence interval of the recorded curve. When a recorded data point is found to exceed the interval range, it is determined to be an abnormal data point and the mean before and after the point is used for correction. Thereafter, the above process is repeated until there is no limit point.

[0148] Based on the description of the aforementioned method embodiment, the embodiment of the present application further discloses a transformer invisible fault diagnosis device.

[0149] Figure 4 A schematic diagram of the structure of a transformer invisible fault diagnosis device provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the transformer invisible fault diagnosis device 400 may include:

[0150] An acquisition module 410 is used to acquire voltage and current recording data of the transformer measured online;

[0151] The elimination module 420 is used to eliminate the influencing factors of the voltage and current recording data of the transformer, and is specifically used to:

[0152] Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current;

[0153] Based on the fundamental voltage and current, when calculating the elliptical trajectory, normalization processing is performed on different loads;

[0154] Based on the fundamental voltage and current, a Gaussian regression algorithm is used to eliminate the influence of abnormal values;

[0155] Based on the fundamental voltage and current, a transformer port relationship after compensation with the excitation current is used to construct a voltage difference and current elliptical trajectory;

[0156] Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured during sampling time are used to correct the voltage difference and current elliptical trajectory;

[0157] The identification and diagnosis module 430 is used to obtain processed voltage and current signals after the operation of eliminating the influencing factors; use the processed voltage and current signals to realize the identification and diagnosis of transformer invisible faults through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

[0158] It can be understood that the relevant contents of each module in the above device have been described in detail in the aforementioned method embodiment, and the details can be referred to the contents in the method embodiment; that is, a transformer invisible fault diagnosis device provided in the present application can be executed as follows Figure 1 , Figure 2 or Figure 3 Any steps in the illustrated embodiment will not be described in detail here.

[0159] In one embodiment of the present application, an electronic device is also provided. The electronic device may include a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the execution Figure 1 , Figure 2 or Figure 3 Any step in the method embodiment shown can execute the relevant construction of the model and can also execute the application calculation of the model. The electronic device can also include an input / output device, etc. In a specific implementation, the electronic device can be a server, a terminal device, etc.

[0160] In one embodiment, a computer-readable storage medium is further provided. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor executes any step in the above method embodiment.

[0161] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0162] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0163] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A transformer invisible fault diagnosis method, characterized in that: The method comprises: Obtain voltage and current recording data of the transformer measured online; The operation of eliminating influencing factors on the voltage and current recording data of the transformer specifically includes: Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current; Based on the fundamental voltage and current, when calculating the elliptical trajectory, normalization processing is performed on different loads; Based on the fundamental voltage and current, a Gaussian regression algorithm is used to eliminate the influence of abnormal values; Based on the fundamental voltage and current, a transformer port relationship after compensation with the excitation current is used to construct a voltage difference and current elliptical trajectory; Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured during sampling time are used to correct the voltage difference and current elliptical trajectory; After the operation of eliminating the influencing factors, a processed voltage and current signal is obtained; the processed voltage and current signal is used to realize the identification and diagnosis of transformer invisible faults through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

2. The transformer invisible fault diagnosis method according to claim 1 is characterized in that: The normalizing of different loads includes: According to the principle of winding equivalent two-port network circuit, different loads are normalized, and the normalized relationship is: in and The load is Z' L The primary and secondary current values ​​at and According to the circuit principle As a voltage source and A joint response.

3. The transformer invisible fault diagnosis method according to claim 2 is characterized in that: The Gaussian regression algorithm is used to eliminate the influence of abnormal values, including: Use Gaussian process regression to fit the recorded data points and obtain the confidence interval of the recorded curve; When a recorded data point is found to be beyond the confidence interval, it is determined to be an abnormal data point and the point is corrected; Repeat the above steps until there are no abnormal data points.

4. The transformer invisible fault diagnosis method according to claim 3 is characterized in that: The correction of this point includes: Replace the value at a point with the mean of the values ​​before and after the point.

5. The transformer invisible fault diagnosis method according to claim 1, characterized in that: The transformer port relationship after compensation using the excitation current is: U1-U2+αU1=(1+α)U1-U2=(R k +jX k )I1, Among them, R k and X k are the short-circuit resistance and short-circuit reactance of the transformer respectively, and α is the compensation coefficient of the transformer; The method of constructing the voltage difference and current elliptical trajectory comprises: The elliptical trajectory is constructed with (1+α)U1-U2 and I1.

6. The transformer invisible fault diagnosis method according to claim 5, characterized in that: Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured within the sampling time are used to correct the voltage difference and current elliptical trajectory, including: Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured within the sampling time are used to correct the elliptical trajectory constructed with ΔU and I1 or the elliptical trajectory constructed with (1+α)U1-U2 and I1, where ΔU=U1-U2.

7. The transformer invisible fault diagnosis method according to claim 1, characterized in that: The voltage and current recording data include steady-state and transient recording data of voltage and current; After obtaining the voltage and current waveform data of the transformer measured online, the method further includes: Discrete wavelet noise reduction processing is performed based on the voltage and current recording data of the transformer.

8. A transformer invisible fault diagnosis device, characterized in that: include: An acquisition module is used to acquire voltage and current recording data of the transformer measured online; The elimination module is used to eliminate the influencing factors of the voltage and current recording data of the transformer, specifically for: Based on the voltage and current recording data of the transformer, a full-wave differential Fourier algorithm is used to eliminate high-order harmonic components, DC components and attenuated DC components to obtain fundamental voltage and current; Based on the fundamental voltage and current, when calculating the elliptical trajectory, normalization processing is performed on different loads; Based on the fundamental voltage and current, a Gaussian regression algorithm is used to eliminate the influence of abnormal values; Based on the fundamental voltage and current, a transformer port relationship after compensation with the excitation current is used to construct a voltage difference and current elliptical trajectory; Based on the fundamental voltage and current, two sets of transformer primary and secondary voltage and primary current data measured during sampling time are used to correct the voltage difference and current elliptical trajectory; The identification and diagnosis module is used to obtain processed voltage and current signals after the operation of eliminating the influencing factors; use the processed voltage and current signals to realize the identification and diagnosis of transformer invisible faults through Lissajous ellipse fitting, feature extraction and artificial intelligence inversion algorithm.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.

Images