Anomaly detection method and device for power transmission and transformation equipment

Abstract

The invention provides an anomaly detection method and device for power transmission and transformation equipment. Obtaining target state data of the to-be-detected power transmission and transformation equipment; performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; according to the frequency domain data corresponding to each physical field data in the target state data, performing corresponding order reduction processing on each physical field data in the target state data to obtain the target state data after order reduction; performing reconstruction processing on the target state data after order reduction to obtain reconstructed target state data; using a preset classification prediction model to process the reconstructed target state data, and determining type information of the reconstructed target state data; and determining whether the to-be-detected power transmission and transformation equipment is abnormal or not according to the type information. On the basis of the method, high-precision order reduction processing of the multiple pieces of physical field data related to the power transmission and transformation equipment is rapidly achieved, and the accuracy of abnormal diagnosis of the power transmission and transformation equipment is improved.

Description

Technical Field

[0001] The present invention relates to the field of multi-physics field technology for power equipment, and in particular to a method and device for detecting abnormalities in power transmission and transformation equipment. Background Art

[0002] At present, power equipment, especially power transmission and transformation equipment as an important part of the power system, usually needs to use full-order models to simulate and calculate the physical field data of the power transmission and transformation equipment, judge whether the power transmission and transformation equipment is abnormal based on the calculation results, and judge the operating status of the power transmission and transformation equipment based on the electrical variable measurement results. However, the current full-order model simulation calculation amount is large, and the time consumption is usually measured in days or weeks, resulting in low efficiency and accuracy.

[0003] To address the above problems, no effective solution has been proposed yet. Summary of the invention

[0004] This specification provides a method for detecting abnormalities in power transmission and transformation equipment, which can quickly realize high-precision and lightweight reduction-order processing of multiple physical field data related to the power transmission and transformation equipment, and thus can efficiently and accurately detect and determine whether there are abnormalities in the power transmission and transformation equipment, effectively reducing the amount of calculation and improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment.

[0005] This specification provides a method for detecting abnormality of power transmission and transformation equipment, including: Acquire target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; Performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; According to the frequency domain data corresponding to each physical field data in the target state data, performing corresponding order reduction processing on each physical field data in the target state data to obtain the reduced-order target state data; Reconstructing the reduced-order target state data to obtain reconstructed target state data; Determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; It is determined whether the power transmission and transformation equipment to be detected has any abnormality according to the type information.

[0006] In one embodiment, the step of performing order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field in the target state data to obtain the reduced-order target state data includes: For the frequency domain data corresponding to each physical field data in the target state data, determining a cyclostationary spectral density matrix corresponding to each physical field; Decomposing the cyclostationary spectral density matrix to obtain a modal matrix corresponding to the cyclostationary spectral density matrix; The reduced-order target state data is obtained according to a preset number of dominant modes in the modal matrix.

[0007] In one embodiment, determining the type information of the reconstructed target state data based on the reconstructed target state data according to a preset classification prediction model includes: Obtaining a weight coefficient corresponding to each of the physical field data; According to the weight coefficient, weighted processing is performed on different physical field data in the reconstructed target state data to obtain processed target state data; The preset classification prediction model is used to determine type information of the reconstructed target state data according to the processed target state data.

[0008] In one embodiment, performing frequency domain conversion processing on each physical field data in the target state data to obtain frequency domain data corresponding to the target state data includes: Determining whether the target state data contains abnormal values; When it is determined that the target state data contains abnormal values, the abnormal values ​​in the target state data are assimilated using a preset data assimilation model; wherein the preset data assimilation model is a model constructed according to a data assimilation algorithm; Frequency domain conversion processing is performed on each physical field data in the target state data obtained after the assimilation processing to obtain frequency domain data corresponding to the target state data.

[0009] In one embodiment, the reconstructing the reduced-order target state data to obtain the reconstructed target state data includes: According to a preset energy threshold, the modal data corresponding to each physical field in the reduced-order target state data is screened to obtain the screened target state data; wherein the preset energy threshold is determined according to the energy proportion of the modal data corresponding to each physical field data in the reconstructed target state data; The filtered target state data is reconstructed to obtain reconstructed target state data.

[0010] In one embodiment, the step of performing order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data to obtain the reduced-order target state data includes: By using a preset reduced-order model, each physical field data in the target state data is subjected to reduced-order processing according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced-order target state data.

[0011] In one embodiment, the reconstructing the reduced-order target state data to obtain the reconstructed target state data includes: The reduced-order target state data is reconstructed using a preset reconstruction model to obtain the reconstructed target state data, wherein the preset reconstruction model is a model constructed according to a preset machine learning algorithm.

[0012] This specification provides an abnormality detection device for power transmission and transformation equipment, including: A data acquisition module, used to acquire target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; A frequency domain conversion module is used to perform frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; A reduction order processing module, used for performing corresponding reduction order processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced order target state data; A reconstruction processing module, used for reconstructing the reduced-order target state data to obtain reconstructed target state data; A classification prediction module, used to determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; The abnormality judgment module is used to determine whether there is an abnormality in the power transmission and transformation equipment to be detected according to the type information.

[0013] This specification also provides an electronic device, including a processor and a memory for storing instructions executable by the processor, wherein when the processor executes the instructions, a method for detecting abnormalities in power transmission and transformation equipment is implemented.

[0014] The present specification also provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed, a method for detecting abnormalities in power transmission and transformation equipment is implemented.

[0015] A method for detecting abnormalities in power transmission and transformation equipment based on the present specification is provided, by acquiring target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes multiple physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; each physical field data in the target state data is subjected to frequency domain conversion processing to obtain corresponding frequency domain data; according to the frequency domain data corresponding to each physical field data in the target state data, each physical field data in the target state data is subjected to corresponding order reduction processing to obtain reduced-order target state data; the reduced-order target state data is subjected to reconstruction processing to obtain reconstructed target state data; a preset classification prediction model is used to process the reconstructed target state data to determine the type information of the reconstructed target state data; according to the type information, it is determined whether the power transmission and transformation equipment to be detected has an abnormality. In this way, the physical field data is reduced in order and reconstructed according to the frequency domain data obtained by conversion, so that the reconstructed target state data can remove the noise data in the original physical field data while retaining the valid information in the original physical field data. Furthermore, through the reconstructed target state data, it is possible to quickly and accurately determine whether the power transmission and transformation equipment to be tested has any abnormality, thereby improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of this specification, the drawings required for use in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flowchart of an abnormality detection method for power transmission and transformation equipment provided by an embodiment of this specification; Figure 2 It is a schematic diagram of the structure of an electronic device provided by an embodiment of this specification; Figure 3 It is a schematic diagram of the structure of an abnormality detection device for power transmission and transformation equipment provided by an embodiment of the present specification; Figure 4 It is a schematic diagram of a data assimilation process of deep learning provided by an embodiment of this specification; Figure 5 It is a schematic diagram of a lightweight reduction algorithm development idea provided by an embodiment of this specification; Figure 6 It is a schematic diagram of another lightweight reduction algorithm development idea provided by an embodiment of this specification; Figure 7is a schematic diagram of an original physical field and a physical field reconstructed after order reduction provided by an embodiment of this specification; Figure 8 It is a schematic diagram of comparing a true value and a predicted result point cloud diagram provided by an embodiment of this specification; Fig. 9 is a schematic diagram of a fourth-order typical mode of magnetic field intensity provided by an embodiment of this specification; Fig.10 It is a schematic diagram of an original physical field of magnetic field intensity and a physical field reconstructed after reduction, provided by an embodiment of this specification; Fig.11 Another schematic diagram of the original physical field of magnetic field strength and the reconstructed physical field after reduction is provided by an embodiment of this specification. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0019] See also Figure 1 As shown, the embodiment of this specification provides a method for detecting abnormalities in power transmission and transformation equipment, wherein the method is specifically applied to a server side. When implemented specifically, the method may include the following contents: S101: Acquire target state data of a power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; S102: performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; S103: performing corresponding order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data, to obtain the reduced-order target state data; S104: reconstructing the reduced-order target state data to obtain reconstructed target state data; S105: Determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; S106: Determine whether the power transmission and transformation equipment to be detected has an abnormality according to the type information.

[0020] The above-mentioned power transmission and transformation equipment may be equipment such as converter valves, transformers, circuit breakers, etc. Specifically, the control transformer may be a part of the above-mentioned power transmission and transformation equipment for stabilizing and regulating voltage.

[0021] Specifically, taking the physical field data as electric field data and the power transmission and transformation equipment as a converter valve as an example, the above-mentioned target state data may include electric variable data such as electric field strength, voltage and current, and the electric variable data may include voltage, current, resistance, capacitance, inductance, etc.

[0022] In addition, taking the physical field data as magnetic field data as an example, the above-mentioned target state data may include magnetic field data such as magnetic induction intensity, magnetic field intensity, magnetic flux and magnetic flux density; taking the physical field data as temperature field data as an example, the above-mentioned target state data may include temperature field data such as heat flux, hot melt and thermal conductivity.

[0023] In some embodiments, the frequency domain conversion processing is performed on each physical field data in the target state data to obtain corresponding frequency domain data. When specifically implemented, it may include: Using a preset frequency domain conversion algorithm, each physical field data in the target state data is subjected to frequency domain conversion processing to obtain corresponding frequency domain data.

[0024] The preset frequency domain conversion algorithm may be a Fourier transform algorithm, a wavelet transform algorithm or the like, which is used to convert time domain data into frequency domain data.

[0025] Specifically, taking the Fourier transform algorithm as an example, the time domain signal corresponding to each physical field data in the target state data is converted into a frequency domain signal, and then the corresponding frequency domain data is obtained. The converted frequency domain data contains an amplitude spectrum and a phase spectrum, which are used to display the intensity and phase information of each frequency component. For example, the high-frequency component of the electric field data signal can reflect the electromagnetic interference problem.

[0026] In some embodiments, taking the physical field data as electric field data as an example, since the electric field intensity in the electric field data focuses on the spatial distribution and instantaneous changes, the voltage in the electric field data focuses on the amplitude and phase characteristics, and the current in the electric field data focuses on the load characteristics and periodic changes, therefore, each physical field data in the target state data is subjected to corresponding order reduction processing.

[0027] In some embodiments, the reconstructing the reduced-order target state data to obtain the reconstructed target state data may include: S1: extracting modal data whose modal data energy ratio is higher than a preset energy ratio from the frequency domain data corresponding to each physical field data in the target state data as the main modal data; S2: extracting modal coefficients of the main modal data to obtain modal coefficients of the main modal data; S3: Reconstruct the main modal data using a preset reconstruction algorithm and the modal coefficients to obtain reconstructed target state data.

[0028] Among them, the above-mentioned preset classification prediction model can be a classification model constructed based on a preset machine learning algorithm.

[0029] For example, taking the preset classification prediction model as a classification model built based on the random forest algorithm, historical sample data can be obtained to train the preset classification prediction model to obtain the trained preset classification prediction model. Since the multi-decision tree structure of the random forest algorithm can provide the final classification result through a voting mechanism, the high accuracy and robustness of the trained preset classification prediction model can be ensured.

[0030] The above type information may include a normal type, an abnormal type or a specific fault type. The abnormal type may refer to any insulation failure risk of the power transmission and transformation equipment, and the specific fault type may be, for example, a fault caused by overheating, short circuit, harmonic interference or electromagnetic interference.

[0031] For instance, taking the physical field data as electric field data, since the electric field strength is an important indicator for monitoring and maintaining electrical equipment, the change amplitude of the electric field strength in the physical field data can be used as the main factor, while the change amplitude of the voltage and current can be used as an auxiliary factor to comprehensively determine the type information of the reconstructed target state data.

[0032] Specifically, if the change amplitude of the electric field strength is greater than a first preset threshold, the type information may be an abnormal type or a fault type; if the change amplitude of the electric field strength is not greater than the first preset threshold, and the change amplitude of the voltage or current is greater than the second preset threshold, the type information is a specific fault type; if the change amplitude of the electric field strength is not greater than the first preset threshold, and the change amplitudes of the voltage and current are not greater than the second preset threshold, the type information is a normal type.

[0033] In some embodiments, after determining whether the power transmission and transformation equipment to be detected is abnormal according to the type information, the method may further include: When it is determined that the power transmission and transformation equipment to be detected has an abnormality, an alarm of the type information is sent to relevant operation and maintenance personnel according to the type information.

[0034] In this way, the relevant operation and maintenance personnel can take different abnormal handling measures according to different types of information. At the same time, the alarm can be accompanied by relevant physical field data, monitoring time, equipment number and abnormality degree, etc., to help the operation and maintenance personnel quickly locate the problem. With clear alarm information, the operation and maintenance personnel can give priority to handling the most urgent faults to ensure the normal operation of the equipment, while also achieving the effect of energy saving and environmental protection.

[0035] Furthermore, after determining whether the power transmission and transformation equipment to be detected has an abnormality according to the type information, the method may also include: determining the cause of the abnormality; querying the preset abnormality handling strategy library according to the cause of the abnormality, determining the preset abnormality handling strategy corresponding to the abnormal cause as the target handling strategy; and providing the target handling strategy to the operation and maintenance personnel. Among them, the preset abnormality handling strategy library includes multiple preset abnormality handling strategies, and each preset abnormality handling strategy corresponds to at least one abnormal cause. The multiple preset abnormality handling strategies are pre-acquired and determined by clustering based on a large number of historical abnormality handling records.

[0036] Based on the above embodiment, the physical field data is reduced in order and reconstructed according to the frequency domain data obtained by conversion, so that the reconstructed target state data can remove the noise data in the original physical field data while retaining the valid information in the original physical field data. Furthermore, through the reconstructed target state data, it is possible to quickly and accurately determine whether the power transmission and transformation equipment to be detected has any abnormality, thereby improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment.

[0037] In some embodiments, according to the frequency domain data corresponding to each physical field in the target state data, each physical field data in the target state data is reduced in order to obtain the reduced target state data. When the method is specifically implemented, it may also include the following contents: S1: for the frequency domain data corresponding to each physical field data in the target state data, determining the cyclostationary spectral density matrix corresponding to each physical field; S2: Decomposing the cyclostationary spectral density matrix to obtain a modal matrix corresponding to the cyclostationary spectral density matrix; S3: Obtaining the reduced-order target state data according to a preset number of dominant modes in the modal matrix.

[0038] Among them, the above-mentioned cyclostationary spectral density matrix can be a mathematical tool for describing random processes, especially when analyzing time series data. It combines the concepts of cyclostationarity and spectral density to more effectively process and analyze multivariate random signals with periodic characteristics. In addition, for time series data of multiple physical field variables (such as electric field, temperature field and pressure field), the cyclostationary spectral density matrix can provide a global perspective to reveal the correlation and dynamic relationship between variables.

[0039] In some embodiments, the decomposing the cyclostationary spectral density matrix to obtain a modal matrix corresponding to the cyclostationary spectral density matrix may include: Decomposing the cyclostationary spectral density matrix by eigenvalue decomposition to obtain multiple modal data with different energy proportions; According to the modal data of the multiple different energy proportions, a modal matrix corresponding to the cyclostationary spectral density matrix is ​​constructed.

[0040] The dominant mode mentioned above may be obtained by sorting the modal data according to the energy proportion of the modal data, and selecting the dominant mode according to the sorted modal data.

[0041] In some embodiments, the method may further include the following contents when determining the type information of the reconstructed target state data based on the reconstructed target state data according to the preset classification prediction model: S1: Obtaining a weight coefficient corresponding to each physical field data; S2: performing weighted processing on different physical field data in the reconstructed target state data according to the weight coefficient to obtain processed target state data; S3: Determine type information of the reconstructed target state data according to the processed target state data using the preset classification prediction model.

[0042] The weight coefficients mentioned above may be the contribution of the importance index of each physical field data, which may be obtained specifically according to the following method: S1: using the principal component analysis method, performing standardization processing on the physical field data to obtain standardized data; S2: calculating the covariance matrix and eigenvalue decomposition of the data obtained by the standardization process, and selecting the principal component to obtain the contribution of the importance index of each physical field data; S3: Determine the weight coefficient of the importance index of each physical field data according to the contribution of the importance index of each physical field data.

[0043] In addition, the weights of each physical field can be set manually according to the role and influence of each physical field in the system. For example, since the electric field strength may have a greater impact on power transmission and transformation equipment, it can be given a higher weight coefficient.

[0044] In some embodiments, the frequency domain conversion process is performed on each physical field data in the target state data to obtain the frequency domain data corresponding to the target state data. When the method is specifically implemented, the following contents may also be included: S1: Determine whether the target state data contains abnormal values; S2: When it is determined that the target state data contains abnormal values, use a preset data assimilation model to assimilate the abnormal values ​​in the target state data; wherein the preset data assimilation model is a model constructed according to a data assimilation algorithm; S3: Perform frequency domain conversion processing on each physical field data in the target state data obtained after the assimilation processing to obtain frequency domain data corresponding to the target state data.

[0045] Based on the above embodiments, the quality of multiple physical field data and the accuracy of the power transmission and transformation equipment monitoring can be improved, and the stable operation of the system and the reliability of the prediction can be guaranteed.

[0046] In some embodiments, the method of reconstructing the reduced-order target state data to obtain reconstructed target state data may further include the following when the method is specifically implemented: S1: According to a preset energy threshold, the modal data corresponding to each physical field in the reduced-order target state data is screened to obtain the screened target state data; wherein the preset energy threshold is determined according to the energy proportion of the modal data corresponding to each physical field data in the reconstructed target state data; S2: Reconstruct the filtered target state data to obtain reconstructed target state data.

[0047] Among them, the above energy threshold can be used to filter out the dominant mode. The higher the energy threshold setting value, the more modal information is retained. For example, when the energy threshold is 90%, the main modal information can be retained and the noise can be appropriately reduced; when the energy threshold is 95%, it can be used for high-precision reconstruction; when the energy threshold is 99%, all information is retained as much as possible, which is suitable for precise analysis, but the dimensionality reduction effect is poor.

[0048] Specifically, the modal data corresponding to each physical field in the reduced-order target state data is screened, and taking an energy threshold of 90% as an example, the modal data with an energy accumulation ratio greater than or equal to 90% is screened out.

[0049] Based on the above embodiments, not only the amount of data is reduced, but also the key features of the system are retained. When combined with the random forest model, the reduced-order data can greatly improve the monitoring accuracy, realize predictive maintenance and accurate positioning of equipment failures, thereby improving the stability and operation efficiency of the system.

[0050] In some embodiments, according to the frequency domain data corresponding to each physical field data in the target state data, each physical field data in the target state data is reduced in order to obtain the reduced target state data. When the method is specifically implemented, it may also include the following contents: By using a preset reduced-order model, each physical field data in the target state data is subjected to reduced-order processing according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced-order target state data.

[0051] The above-mentioned preset reduced-order model can be a mathematical or computational model that retains key features or main dynamic behaviors by reducing the dimensions of the system or data. Its core goal is to reduce the computational complexity without losing important information, thereby making tasks such as analysis, prediction, and control more efficient. For example, the preset reduced-order model can include models constructed by algorithms such as principal component analysis, dynamic mode decomposition, cyclostationary spectral orthogonal decomposition, balanced truncation reduction, and kernel principal component analysis.

[0052] Specifically, taking kernel principal component analysis as an example, a suitable kernel function is selected to map the frequency domain data corresponding to each physical field data in the target state data to a high-dimensional feature space, the kernel matrix is ​​centrally processed, the principal components are determined, and the frequency domain data corresponding to each physical field data in the target state data is mapped to a low-dimensional space using the selected eigenvectors, and each physical field data in the target state data is reduced in order to obtain the reduced-order target state data.

[0053] In some embodiments, the method of reconstructing the reduced-order target state data to obtain reconstructed target state data may further include the following when the method is specifically implemented: The reduced-order target state data is reconstructed using a preset reconstruction model to obtain the reconstructed target state data, wherein the preset reconstruction model is a model constructed according to a preset machine learning algorithm.

[0054] As can be seen from the above, an abnormality detection method for power transmission and transformation equipment provided in an embodiment of the present specification obtains target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes multiple physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; each physical field data in the target state data is subjected to frequency domain conversion processing to obtain corresponding frequency domain data; according to the frequency domain data corresponding to each physical field data in the target state data, each physical field data in the target state data is subjected to corresponding order reduction processing to obtain reduced-order target state data; the reduced-order target state data is subjected to reconstruction processing to obtain reconstructed target state data; a preset classification prediction model is used to process the reconstructed target state data to determine the type information of the reconstructed target state data; according to the type information, it is determined whether the power transmission and transformation equipment to be detected has an abnormality. In this way, the physical field data is reduced in order and reconstructed according to the frequency domain data obtained by conversion, so that the reconstructed target state data can remove the noise data in the original physical field data while retaining the valid information in the original physical field data. Furthermore, through the reconstructed target state data, it is possible to quickly and accurately determine whether the power transmission and transformation equipment to be tested has any abnormality, thereby improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment.

[0055] See also Figure 2 As shown, the embodiment of this specification also provides a specific electronic device, wherein the electronic device includes a network communication port 201, a processor 202 and a memory 203, and the above structures are connected through internal cables so that each structure can perform specific data interaction.

[0056] The network communication port 201 can be specifically used to obtain target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes multiple physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data.

[0057] The processor 202 can be specifically used to perform frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; perform corresponding order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data to obtain reduced-order target state data; reconstruct the reduced-order target state data to obtain reconstructed target state data; use a preset classification prediction model to process the reconstructed target state data to determine the type information of the reconstructed target state data; and determine whether there is an abnormality in the power transmission and transformation equipment to be detected based on the type information.

[0058] The memory 203 may be specifically used to store corresponding instruction programs.

[0059] Based on the above method, the relevant structural performance of the electronic equipment can be effectively utilized, the data processing speed of the electronic equipment can be improved, and the abnormality detection method of the power transmission and transformation equipment can be efficiently implemented.

[0060] In this embodiment, the network communication port 201 can be a virtual port that is bound to different communication protocols so that different data can be sent or received. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. In addition, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM, CDMA, etc.; it can also be a Wifi chip; it can also be a Bluetooth chip.

[0061] In this embodiment, the processor 202 may be implemented in any appropriate manner. For example, the processor may take the form of a microprocessor or a processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) executable by the (micro)processor, a logic gate, a switch, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, etc. This specification does not limit this.

[0062] In this embodiment, the memory 203 may include multiple levels. In a digital system, anything that can store binary data can be a memory; in an integrated circuit, a circuit with a storage function that has no physical form is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0063] The embodiment of the present specification also provides a computer-readable storage medium based on the above-mentioned abnormality detection method for power transmission and transformation equipment, wherein the computer-readable storage medium stores computer program instructions, which, when executed, implement the following: obtaining target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes multiple physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; performing corresponding order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data to obtain reduced-order target state data; reconstructing the reduced-order target state data to obtain reconstructed target state data; determining type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; and determining whether the power transmission and transformation equipment to be detected has an abnormality according to the type information.

[0064] In this embodiment, the storage medium includes but is not limited to a random access memory (RAM), a read-only memory (ROM), a cache, a hard disk (HDD) or a memory card. The memory may be used to store computer program instructions. The network communication unit may be an interface for network connection communication set in accordance with the standard specified by the communication protocol.

[0065] In this embodiment, the functions and effects specifically implemented by the program instructions stored in the computer-readable storage medium can be explained in comparison with other implementations and will not be described in detail here.

[0066] See also Figure 3 At the software level, the embodiment of this specification also provides an abnormality detection device for power transmission and transformation equipment, which may specifically include the following structural modules: The data acquisition module 301 is used to acquire target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; A frequency domain conversion module 302 is used to perform frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; The order reduction processing module 303 is used to perform corresponding order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced-order target state data; A reconstruction processing module 304 is used to reconstruct the reduced-order target state data to obtain reconstructed target state data; A classification prediction module 305 is used to determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; The abnormality determination module 306 is used to determine whether the power transmission and transformation equipment to be detected has an abnormality according to the type information.

[0067] In some embodiments, the above-mentioned order reduction processing module 303, when implemented specifically, determines the cyclostationary spectral density matrix corresponding to each physical field data in the target state data for the frequency domain data corresponding to the physical field data; decomposes the cyclostationary spectral density matrix to obtain the modal matrix corresponding to the cyclostationary spectral density matrix; and obtains the reduced-order target state data according to a preset number of dominant modes in the modal matrix.

[0068] In some embodiments, the above-mentioned classification prediction module 305, when implemented, obtains a weight coefficient corresponding to each of the physical field data; performs weighted processing on different physical field data in the reconstructed target state data according to the weight coefficient to obtain processed target state data; and uses the preset classification prediction model to determine the type information of the reconstructed target state data according to the processed target state data.

[0069] In some embodiments, the frequency domain conversion module 302, when implemented, determines whether the target state data contains abnormal values; when it is determined that the target state data contains abnormal values, the abnormal values ​​in the target state data are assimilated using a preset data assimilation model; wherein the preset data assimilation model is a model constructed according to a data assimilation algorithm; and each physical field data in the target state data obtained after the assimilation process is subjected to frequency domain conversion processing to obtain frequency domain data corresponding to the target state data.

[0070] In some embodiments, the above-mentioned reconstruction processing module 304, when specifically implemented, performs screening processing on the modal data corresponding to each physical field in the reduced-order target state data according to a preset energy threshold to obtain the screened target state data; wherein the preset energy threshold is determined according to the energy proportion of the modal data corresponding to each physical field data in the reconstructed target state data; and performs reconstruction processing on the screened target state data to obtain the reconstructed target state data.

[0071] In some embodiments, the above-mentioned order reduction processing module 303, when implemented, can also use a preset order reduction model to perform order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data to obtain the reduced-order target state data.

[0072] In some embodiments, the above-mentioned reconstruction processing module 304, when implemented, can also use a preset reconstruction model to reconstruct the reduced-order target state data to obtain the reconstructed target state data, wherein the preset reconstruction model is a model constructed according to a preset machine learning algorithm.

[0073] It should be noted that the units, devices or modules described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. For the convenience of description, the above devices are described separately by functions divided into various modules. Of course, when implementing this specification, the functions of each module can be implemented in the same or more software and / or hardware, or the modules that implement the same function can be implemented by a combination of multiple sub-modules or sub-units. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0074] As can be seen from the above, based on an abnormality detection device for power transmission and transformation equipment provided in an embodiment of the present specification, the relationship between the pore throat diameter and the crude oil content in the first layer obtained by laser confocal microscopy, and the relationship between the transverse relaxation time and the nuclear magnetic resonance signal intensity in the first layer obtained by nuclear magnetic resonance testing, can determine the target relationship between the transverse relaxation time and the pore throat diameter, and then the distribution information of the pore throat diameter in the second layer can be determined based on the target relationship, the transverse relaxation time in the second layer and the nuclear magnetic resonance signal intensity, so as to improve the accuracy of the pore size characterization in the oil reservoir without destroying the core, so as to improve the development effect of the oil and gas reservoir.

[0075] In a specific scenario example, a method for detecting abnormalities of power transmission and transformation equipment provided in this specification can be applied to quickly realize high-precision and lightweight order reduction processing of multiple physical field data related to the power transmission and transformation equipment, thereby being able to efficiently and accurately detect and determine whether there are abnormalities in the power transmission and transformation equipment, effectively reducing the amount of calculation and improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment. The specific implementation process may include the following contents.

[0076] See also Figure 4 As shown in the figure, the abnormal processing of multi-physics field data. Specifically, the multi-physics field data observed in real time is monitored for abnormal data; after the abnormal data is detected, it is fed back to the monitoring system module for inspection, and after it is confirmed as abnormal data, the data assimilation substitution by deep learning is performed; in the dynamic operation process, the "abnormal data value" is integrated with the new observation data, and the normal value of the abnormal data is assimilated through the existing data; then the processed data is provided to the digital twin reduction algorithm, and the data assimilation algorithm can be re-optimized according to the feedback of the algorithm until the data obtained by the reduction algorithm no longer detects abnormal data signals.

[0077] See also Figure 5 As shown, S1: Circularly Symmetric Power Orthogonal Decomposition (CSPOD) combined with random forest algorithm with abnormal data detection.

[0078] Specifically, by performing the orthogonal decomposition of the cyclostationary spectrum and adjusting the preset energy threshold, modal data with different energy proportions are obtained, information is fully retained, and the accuracy of modal reconstruction is improved.

[0079] The cyclostationary spectrum orthogonal decomposition is a purely data-driven reduced-order prediction model, which has no requirements for the type of model, takes little computing time, and does not require model training in advance. It also has high accuracy and rapid response. By adjusting the energy threshold, the order reduction process is performed, and according to different energy proportions, only a small number of low-order modes can be used to invert the approximation of the full-order data.

[0080] Furthermore, the random forest algorithm with abnormal data detection can use data assimilation capabilities to judge and assimilate abnormal data, ensure the reliability and accuracy of data as much as possible, and reduce the defects of sparse observation (sampling) data. Figure 6 As shown, the input and output physical parameters (i.e., the reconstructed target state data) are selected, and the training, test, and validation sets are selected. The model is accurately evaluated. If it fails, parameter and sample optimization are performed. If it passes, the reduced-order model is stored and output.

[0081] S2: Multi-physics field variable coupling weighted algorithm.

[0082] In the case of coupling relationships between multiple physical field variables, the weight coefficients of each physical field data are considered, and the coupling terms are reasonably weighted during the calculation process. This multi-physical field variable coupling weighting algorithm helps to improve the accuracy and stability of the calculation.

[0083] S3: Build a preset classification prediction model.

[0084] Specifically, by inputting the reconstructed target state data of the sample, an efficient and accurate preset classification prediction model is trained. The model can process a large amount of data in a short time and output results that are close to or even more accurate than traditional methods.

[0085] S4: Performance evaluation and optimization.

[0086] According to actual application requirements, on the basis of ensuring the accuracy of modal reduction and multi-physics field data coupling weighted algorithm, further optimize the calculation speed and memory usage. Parallel computing technology, specific hardware acceleration or algorithm optimization can be used to improve the calculation efficiency.

[0087] S5: Verification and application.

[0088] By comparing and verifying with traditional methods, the accuracy and efficiency of its calculation results are evaluated. At the same time, the model training part is compared with the data in the real scene to verify its generalization ability and stability. Finally, this standardized process is documented with detailed instructions and packaging management to ensure its reusability and continuous optimization. This will help communication and collaboration between users and improve the accuracy and efficiency of the entire process.

[0089] Based on the above embodiments, the computational cost and memory cost can be greatly reduced by modal order reduction processing under the conditions of stability, accuracy and efficiency, and the accuracy of the calculation results can be effectively guaranteed. In addition, it is possible to efficiently and accurately detect and determine whether there is an abnormality in the power transmission and transformation equipment, effectively reducing the amount of calculation and improving the efficiency and accuracy of abnormal operation and maintenance diagnosis of the power transmission and transformation equipment. In addition, the application of the order reduction method in the field of power transmission and transformation equipment can bring the following potential benefits: based on the reduced order model algorithm and the multi-physics field rapid response model, the computational efficiency is significantly improved compared with the full-order model; a reduced order model is constructed to realize a high-precision and lightweight reduced order simulation model with second-level calculation, and realize fast and accurate prediction of the physical field; a reduced order model with universality and flexibility is realized, and different types of multi-physics field models can be used, and it can be adjusted and optimized according to the characteristics of specific physical problems; the model can obtain results in a short time in engineering scenarios such as real-time control and decision support, so that the system can quickly respond to and adapt to different operating requirements.

[0090] The advantages of the algorithm are illustrated by taking two special application scenarios as examples. The GIS bus model (Geographic Information System, GIS) is the steady-state multi-physical field data under a fixed current, while the converter valve model is the unsteady-state multi-physical field data obtained by current pulses.

[0091] GIS busbar model 1) See Figure 7 As shown, the comparison between the original physical field and the reconstructed physical field after reduction is shown in Table 1.

[0092] Taking the GIS busbar model as an example, it is found that a high accuracy can be achieved within seconds.

[0093] 2) See Figure 8 As shown, the point cloud comparison of the true value and predicted result of random forest with abnormal data detection, Among them, the prediction error and prediction time under different input parameters are shown in Table 2:

[0094] Converter valve model (non-steady current pulse data) 1) See Fig. 9 As shown in Figure 2, the cyclostationary spectral orthogonal decomposition (CSPOD) is used to analyze the fourth-order typical mode of magnetic field intensity; see Fig.10 As shown in Table 3, the comparison between the original physical field of magnetic field intensity and the reconstructed physical field after reduction is shown in Table 3:

[0095] Taking the converter valve as an example, even for non-steady-state data, the accuracy is still guaranteed to be around 4%.

[0096] 2) See Fig.11 As shown, the original physical field of the magnetic field intensity of the random forest with abnormal data detection (left) is compared with the reconstructed physical field after reduction (right). Through quantitative comparison, the average error that meets the project indicator requirements is less than 5% (compared with the original simulation result data), and the estimated average error is 2.68%.

[0097] Although the present specification provides method operation steps as described in the embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative means. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only order of execution. When the device or client product in practice is executed, it can be executed in sequence or in parallel according to the method shown in the embodiments or the drawings (for example, a parallel processor or a multi-threaded processing environment, or even a distributed data processing environment). The terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, product or device including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements inherent to such a process, method, product or device. In the absence of more restrictions, it is not excluded that there are other identical or equivalent elements in the process, method, product or device including the elements. The words first, second, etc. are used to represent the name, and do not represent any particular order.

[0098] Those skilled in the art also know that, in addition to implementing the controller in a purely computer-readable program code, the controller can be made to implement the same function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered as a hardware component, and the devices for implementing various functions included therein can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules for implementing the method and structures within the hardware component.

[0099] Through the description of the above embodiments, it can be known that those skilled in the art can clearly understand that the present specification can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present specification can essentially be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a mobile terminal, a server, or a network device, etc.) to execute the methods described in each embodiment of the present specification or some parts of the embodiments.

[0100] Although the present specification is described through embodiments, those skilled in the art will appreciate that there are many modifications and changes to the present specification without departing from the spirit of the present specification, and it is intended that the appended claims include these modifications and changes without departing from the spirit of the present specification.

Claims

1. A method for detecting abnormality of power transmission and transformation equipment, characterized in that: include: Acquire target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; Performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; According to the frequency domain data corresponding to each physical field data in the target state data, performing corresponding order reduction processing on each physical field data in the target state data to obtain the reduced-order target state data; Reconstructing the reduced-order target state data to obtain reconstructed target state data; Determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; It is determined whether the power transmission and transformation equipment to be detected has any abnormality according to the type information.

2. The method according to claim 1, characterized in that: The step of performing order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field in the target state data to obtain the reduced-order target state data includes: For the frequency domain data corresponding to each physical field data in the target state data, determining a cyclostationary spectral density matrix corresponding to each physical field; Decomposing the cyclostationary spectral density matrix to obtain a modal matrix corresponding to the cyclostationary spectral density matrix; The reduced-order target state data is obtained according to a preset number of dominant modes in the modal matrix.

3. The method according to claim 2, characterized in that The method of using a preset classification prediction model to process the reconstructed target state data to determine type information of the reconstructed target state data includes: Obtaining a weight coefficient corresponding to each of the physical field data; According to the weight coefficient, weighted processing is performed on different physical field data in the reconstructed target state data to obtain processed target state data; The preset classification prediction model is used to determine type information of the reconstructed target state data according to the processed target state data.

4. The method according to claim 3, characterized in that The performing frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data includes: Determining whether the target state data contains abnormal values; When it is determined that the target state data contains abnormal values, the abnormal values ​​in the target state data are assimilated using a preset data assimilation model; wherein the preset data assimilation model is a model constructed according to a data assimilation algorithm; Each physical field data in the target state data obtained after assimilation is converted into a frequency domain to obtain the corresponding frequency domain data.

5. The method according to claim 4, characterized in that The reconstructing the reduced-order target state data to obtain reconstructed target state data includes: According to a preset energy threshold, the modal data corresponding to each physical field in the reduced-order target state data is screened to obtain the screened target state data; wherein the preset energy threshold is determined according to the energy proportion of the modal data corresponding to each physical field data in the reconstructed target state data; The filtered target state data is reconstructed to obtain reconstructed target state data.

6. The method according to claim 1, characterized in that The step of performing order reduction processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data to obtain the reduced-order target state data comprises: By using a preset reduced-order model, each physical field data in the target state data is subjected to reduced-order processing according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced-order target state data.

7. The method according to claim 6, characterized in that The reconstructing the reduced-order target state data to obtain reconstructed target state data includes: The reduced-order target state data is reconstructed using a preset reconstruction model to obtain the reconstructed target state data, wherein the preset reconstruction model is a model constructed according to a preset machine learning algorithm.

8. An abnormality detection device for power transmission and transformation equipment, characterized in that: include: A data acquisition module, used to acquire target state data of the power transmission and transformation equipment to be detected; wherein the target state data includes a plurality of physical field data, and the physical field data includes at least one of the following: electric field, magnetic field and temperature field data; A frequency domain conversion module is used to perform frequency domain conversion processing on each physical field data in the target state data to obtain corresponding frequency domain data; A reduction order processing module, used for performing corresponding reduction order processing on each physical field data in the target state data according to the frequency domain data corresponding to each physical field data in the target state data, so as to obtain the reduced order target state data; A reconstruction processing module, used for reconstructing the reduced-order target state data to obtain reconstructed target state data; A classification prediction module, used to determine type information of the reconstructed target state data by processing the reconstructed target state data using a preset classification prediction model; The abnormality judgment module is used to determine whether there is an abnormality in the power transmission and transformation equipment to be detected according to the type information.

9. An electronic device, characterized in that: It comprises a processor and a memory for storing instructions executable by the processor, and when the processor executes the instructions, the steps of the abnormality detection method for power transmission and transformation equipment described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the abnormality detection method for power transmission and transformation equipment described in any one of claims 1 to 7 are implemented.

Images