A digital twin model construction method and system for transformer fault diagnosis

Abstract

The present invention discloses a method and system for constructing a digital twin model for transformer fault diagnosis, comprising: dividing the operating status of each transformer device into information dimensions, defining matching factors for data points under each information dimension, and constructing a registration library by setting standard values ​​for each matching factor under each information dimension; obtaining a sampling sequence under the operating status of each transformer device, determining the deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtaining a comprehensive score for the corresponding information dimension based on the deviation determination result, thereby evaluating the availability of the current sampling sequence; and obtaining a target data set based on the availability evaluation results of all sampling sequences, thereby constructing a digital twin model for transformer fault diagnosis. Based on data registration technology, a data set for model construction is screened from a large number of original data sets, effectively capturing valid data and improving the fault diagnosis accuracy and efficiency of the digital twin model.

Description

Technical Field

[0001] The present invention relates to the field of digital twin technology, and in particular to a method and system for constructing a digital twin model for transformer fault diagnosis. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Currently, transformer equipment condition assessment primarily targets a broad range of equipment, employing mechanisms and causal relationship models based on theoretical analysis, computational simulation, and experimental testing, along with standardized evaluation criteria. Evaluation parameters and thresholds are primarily determined based on statistical analysis of extensive test data and expert experience. However, due to the complexity of transformer failure mechanisms, the diversity of operating environments, and differences in equipment manufacturing processes and operating conditions, establishing rigorous, comprehensive, and accurate assessment and prediction models is difficult. Furthermore, standardized fixed threshold determination methods struggle to ensure applicability across diverse equipment.

[0004] With the development of digital twin technology, research on transformer equipment status assessment methods based on digital twin technology has been carried out one after another. In order to enable the digital twin model to realize the prediction function of equipment status, a large amount of operating data is currently used to construct the digital twin model; however, there is no unified planning and management of these transformer operating status sample sets, resulting in some useless data participating in model construction and affecting the prediction accuracy. This will also cause the contribution of useful data to be weakened in the entire model construction process and prediction process, resulting in a lack of effective application of the accumulated useful data. Summary of the Invention

[0005] In order to solve the above problems, the present invention proposes a method and system for constructing a digital twin model for transformer fault diagnosis. Based on data registration technology, the data set used to build the model is screened out from a large number of original data sets, accurately capturing valid data and improving the fault diagnosis accuracy and efficiency of the digital twin model.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a method for constructing a digital twin model for transformer fault diagnosis, comprising:

[0008] The operating status of each device in the transformer is divided into information dimensions, and matching factors are defined for data points in each information dimension. By setting the standard value of each matching factor in each information dimension, a registration library is constructed.

[0009] Obtain the sampling sequence of each device in the transformer under its operating state, determine the deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result to evaluate the availability of the current sampling sequence;

[0010] The target data set is obtained according to the availability evaluation results of all sampling sequences, and a digital twin model for transformer fault diagnosis is constructed based on it.

[0011] As an optional implementation method, an original data set including normal operating status data, abnormal operating status data and operating status simulation data of each transformer device is obtained, and a sampling sequence of the operating status of each transformer device is obtained after preprocessing the original data set.

[0012] As an optional implementation, the information dimensions include: basic information, historical case information, operating condition information, meteorological information, alarm information, initial information and online monitoring information.

[0013] As an optional implementation method, the information dimension under each sampling sequence is weighted, and the sum of the weights of all information dimensions is 1. The matching factor under each information dimension is also weighted, and the sum of the weights of all matching factors in a single information dimension is 1. At the same time, the weights and standard values ​​are adjusted according to the fault diagnosis accuracy feedback of the constructed digital twin model.

[0014] As an optional implementation, the process of determining the degree of deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library includes: determining the degree of deviation based on the proportion of the absolute value of the difference between the sampling value and the standard value to the standard value, and converting the evaluation score of the matching factor based on the degree of deviation.

[0015] As an optional implementation, the comprehensive score of the information dimension is: the comprehensive score of the information dimension is obtained according to the evaluation scores and corresponding weights of all matching factors in the information dimension.

[0016] As an optional implementation, the process of usability evaluation is as follows: based on the comprehensive scores of all information dimensions in the current sampling sequence and the corresponding weights, an evaluation score of the current sampling sequence is obtained, and usability evaluation is performed based on comparison with the qualified score.

[0017] In a second aspect, the present invention provides a digital twin model construction system for transformer fault diagnosis, comprising:

[0018] A registration library construction model is configured to divide the operating status of each device of the transformer into information dimensions, define matching factors for data points under each information dimension, and construct a registration library by setting standard values ​​for each matching factor under each information dimension;

[0019] The evaluation model is configured to obtain the sampling sequence of each device in the transformer under the operating status, determine the deviation between each sample value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result, thereby evaluating the usability of the current sampling sequence;

[0020] The model building model is configured to obtain the target data set based on the availability evaluation results of all sampling sequences, so as to build a digital twin model for transformer fault diagnosis.

[0021] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the method described in the first aspect.

[0022] In a fourth aspect, the present invention provides a computer-readable storage medium for storing computer instructions, wherein when the computer instructions are executed by a processor, the method described in the first aspect is performed.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] The present invention proposes a method and system for constructing a digital twin model for transformer fault diagnosis. Based on data registration technology, accurate data sets for model construction are screened out from a large number of original data sets, a consistent data processing method is implemented, and effective data is accurately captured, laying a good foundation for the construction and analysis of digital twins, thereby improving the fault diagnosis accuracy and efficiency of digital twin models.

[0025] The present invention targets data of different types, sources and formats, and processes data format conversion, data cleaning, data filling, data evaluation and other processing for data caused by system reasons or other non-system reasons, standardizes non-standardized data, and structures a large amount of unstructured data that cannot be directly calculated by computers, and then performs alignment to ensure alignment accuracy and efficiency.

[0026] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0028] Figure 1Flowchart of a method for constructing a digital twin model for transformer fault diagnosis provided in Example 1 of the present invention;

[0029] Figure 2 Schematic diagram of the registration library provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0030] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0031] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0032] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that 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 comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0034] Example 1

[0035] like Figure 1 As shown, this embodiment provides a method for constructing a digital twin model for transformer fault diagnosis, including:

[0036] The operating status of each device in the transformer is divided into information dimensions, and matching factors are defined for data points in each information dimension. By setting the standard value of each matching factor in each information dimension, a registration library is constructed.

[0037] Obtain the sampling sequence of each device in the transformer under its operating state, determine the deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result to evaluate the availability of the current sampling sequence;

[0038] The target data set is obtained according to the availability evaluation results of all sampling sequences, and a digital twin model for transformer fault diagnosis is constructed based on it.

[0039] In this embodiment, the original data set used to build the digital twin model includes normal operating status data, abnormal operating status data of each transformer device, and operating status simulation data generated by the simulation tool, thereby forming a multi-block, multi-level, and multi-type multi-dimensional data set;

[0040] Among them, normal operating status data can be obtained from the real-time records of the equipment's daily operating status detection system; abnormal operating status data is more difficult to obtain because, under normal circumstances, key components are less likely to fail. Since there is very little data when the equipment fails, you can consider obtaining such data from other channels, such as other departments within the enterprise, or partners upstream and downstream of the supply chain; at the same time, use simulation tools to generate operating status simulation data, and combine this data with physical sensor data to prevent the lack of fault data from becoming an obstacle to evaluation.

[0041] In this embodiment, data preprocessing is performed on the original data set, specifically including:

[0042] (1) Methods for handling missing data values;

[0043] Missing values ​​are common in data due to various factors such as storage and retrieval. Missing data can be divided into single-variable missing and multi-variable missing based on the amount of missing data.

[0044] When the missing data is univariate and does not affect the results of data analysis, it is considered to be completely random missing. The processing method in this case is list deletion or pairwise deletion.

[0045] When the missing data is a single variable and it is impossible to determine whether the missing data is caused by machine or human operation, it is usually considered that the missing result meets the random missing characteristics. When processing missing values, smooth prediction or single value interpolation can be used to fill the data.

[0046] When the missing data are multivariate and it is difficult to find the existing pattern in the data list, the missing data can be supplemented by maximum likelihood estimation or maximum expected value algorithm.

[0047] (2) Methods for handling data outliers;

[0048] Outliers are often called "outliers". The following methods are commonly used to deal with outliers:

[0049] a) Simple statistical analysis. After obtaining the data, a simple descriptive statistical analysis can be performed on the data. For example, the maximum and minimum values ​​can be used to determine whether the value of this variable exceeds the reasonable range. For example, if the customer's age is -20 or 200, it is obviously unreasonable and is an outlier.

[0050] b) Principle. If the data follows a normal distribution, In principle, an outlier is a value in a set of measured values ​​that deviates from the mean by more than 3 standard deviations. The probability of a value other than This is a very rare event with a low probability. If the data does not follow a normal distribution, it can also be described by how many times the standard deviation is away from the mean.

[0051] c) Model-based detection. First, establish a data model. Anomalies are objects that do not fit the model perfectly. If the model is a collection of clusters, anomalies are objects that do not significantly belong to any cluster. When using a regression model, anomalies are objects that are relatively far from the predicted value.

[0052] (3) After cleaning the data, it is necessary to fill in the data to remove outliers. Usually, smoothing prediction and interpolation methods are used to handle missing values ​​for simple data, and random forest algorithms are used to fill in complex data.

[0053] Finally, the sampling sequence of each device in the transformer operating state after preprocessing is obtained.

[0054] In this embodiment, building a registration library is an important step in data registration. Its purpose is to define the standard values ​​of each information dimension in the operating status of each transformer device as a benchmark, perform overall rationality comparison analysis on the matching data, and obtain a reasonable target data set.

[0055] The construction of the registration library should take into account the standard parameters of the transformer status data in all aspects. According to the type, meaning, and source of the transformer status assessment data, the information dimension is divided into multiple types. Each type includes multiple matching factors. The multi-type and multi-factor data are constructed into a registration library. The specific definition is as follows:

[0056] The information dimensions of the operating status of each transformer device include: basic information, historical case information, operating condition information, meteorological information, alarm information, initial information, online monitoring information, etc. Each information dimension includes multiple data points, each data point is defined as a separate matching factor, and each matching factor is independently sampled for measurement value. The measurement value of each matching factor has a corresponding reference standard value, which can come from historical data and expert experience, such as Figure 2 shown.

[0057] In this embodiment, the information dimension under each sampling sequence is weighted, and the sum of the weights of all information dimensions is 1. The matching factors under each information dimension are also weighted, and the sum of the weights of all matching factors in a single information dimension is 1. The above weights are aligned before data alignment and can be defined manually. At the same time, the weights are adjusted based on the diagnostic performance of the constructed digital twin model to achieve gradual tuning during the construction of the digital twin model.

[0058] In this embodiment, data registration is performed on the sampling sequences of each device in the operating state of the transformer according to the registration library to obtain the target data set for building the digital twin model for transformer fault diagnosis; specifically, the following steps are included:

[0059] (1) Evaluation of each matching factor:

[0060] By comparing the sample value of the i-th matching factor with the standard value of the matching factor, the deviation P of the sample value of the matching factor is obtained. The deviation refers to the ratio of the absolute value of the difference between the sample value and the standard value to the standard value, specifically P = |AX| / A; where A is the standard value and X is the sample value;

[0061] Then, the evaluation score Q of the matching factor is calculated based on the deviation i , when the deviation P is 0, Q i =1, the larger P is, the closer Qi is to 0, and Q i The value space is [0-1], that is, when the sampling value is within the standard value range, Q i =1, not within the standard value range, the value is determined according to the degree of deviation from the standard value. The farther away from the normal range, the closer the value is to 0;

[0062] It is understandable that the specific conversion method can be set according to the actual measurement data of different matching factors to set a reasonable value logic, which can be defined manually and is not limited here.

[0063] (2) Evaluation of each information dimension:

[0064] According to the evaluation scores and weights of all matching factors in the jth information dimension, the evaluation score W of the information dimension is obtained. j :

[0065]

[0066] Among them, N j is the number of matching factors in the jth information dimension; K i is the weight of the i-th matching factor in the information dimension. The weights of all matching factors meet the conditions:

[0067] (3) Usability evaluation of target dataset:

[0068] According to the evaluation scores and weights of all information dimensions in the current sampling sequence, the evaluation score S of the current sampling sequence is obtained;

[0069]

[0070] Where m is the number of information dimensions in the current sampling sequence; L j is the weight of the information dimension. The weights of all information dimensions meet the following conditions:

[0071] (4) Define the qualified score of the target data set based on different equipment types and expert experience 合格 , when the evaluation score of the target dataset>=S 合格 , then the target dataset is judged to be the target matching dataset.

[0072] In this embodiment, a digital twin model for transformer fault diagnosis is constructed based on the target data set obtained by the final alignment. After the digital twin model is used to diagnose the transformer fault, the digital twin model is verified according to the diagnosis accuracy. By determining which type of fault diagnosis accuracy is low, the weight of the information dimension, the weight of the matching factor, and the standard value of the data type with low fault diagnosis accuracy are adjusted until the fault diagnosis accuracy meets the requirements.

[0073] Example 2

[0074] This embodiment provides a digital twin model construction system for transformer fault diagnosis, including:

[0075] A registration library construction model is configured to divide the operating status of each device of the transformer into information dimensions, define matching factors for data points under each information dimension, and construct a registration library by setting standard values ​​for each matching factor under each information dimension;

[0076] The evaluation model is configured to obtain the sampling sequence of each device in the transformer under the operating status, determine the deviation between each sample value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result, thereby evaluating the usability of the current sampling sequence;

[0077] The model building model is configured to obtain the target data set based on the availability evaluation results of all sampling sequences, so as to build a digital twin model for transformer fault diagnosis.

[0078] It should be noted that the above modules correspond to the steps described in Example 1, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above Example 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0079] In further embodiments, there is also provided:

[0080] An electronic device includes a memory and a processor, and computer instructions stored in the memory and executed by the processor, wherein when the computer instructions are executed by the processor, the method described in Example 1 is performed. For the sake of brevity, no further details are given here.

[0081] It should be understood that in this embodiment, the processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), off-the-shelf field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0082] The memory may include a read-only memory and a random access memory, and provides instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0083] A computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the method described in Example 1 is performed.

[0084] The method in Example 1 can be directly implemented as a hardware processor, or can be implemented using a combination of hardware and software modules in the processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, it will not be described in detail here.

[0085] Those skilled in the art will appreciate that the units, i.e., algorithm steps, of the various examples described in conjunction with this embodiment can be implemented using electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. A method for constructing a digital twin model for transformer fault diagnosis, characterized in that: include: The operating status of each device in the transformer is divided into information dimensions, and matching factors are defined for data points in each information dimension. By setting the standard value of each matching factor in each information dimension, a registration library is constructed. Obtain the sampling sequence of each device in the transformer under its operating state, determine the deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result to evaluate the availability of the current sampling sequence; The target dataset is obtained based on the availability evaluation results of all sampling sequences, and a digital twin model for transformer fault diagnosis is constructed based on it. Among them, the information dimension under each sampling sequence is weighted, and the sum of the weights of all information dimensions is 1. The matching factor under each information dimension is weighted, and the sum of the weights of all matching factors in a single information dimension is 1. At the same time, the weights and standard values ​​are adjusted according to the fault diagnosis accuracy feedback of the constructed digital twin model; The comprehensive score of the information dimension is: the comprehensive score of the information dimension is obtained according to the evaluation scores and corresponding weights of all matching factors in the information dimension; The process of usability evaluation is as follows: according to the comprehensive scores of all information dimensions in the current sampling sequence and the corresponding weights, the evaluation score of the current sampling sequence is obtained, and the usability evaluation is performed based on the comparison with the qualified score.

2. A method for constructing a digital twin model for transformer fault diagnosis according to claim 1, characterized in that: An original data set including normal operating status data, abnormal operating status data and operating status simulation data of each transformer device is obtained, and a sampling sequence under the operating status of each transformer device is obtained after preprocessing the original data set.

3. The method for constructing a digital twin model for transformer fault diagnosis according to claim 1, wherein: The information dimensions include: basic information, historical case information, operating condition information, meteorological information, alarm information, initial information and online monitoring information.

4. The method for constructing a digital twin model for transformer fault diagnosis according to claim 1, wherein: The process of determining the degree of deviation between each sampling value in the sampling sequence and the standard value of the corresponding matching factor in the registration library includes: determining the degree of deviation based on the proportion of the absolute value of the difference between the sampling value and the standard value to the standard value, and converting the evaluation score of the matching factor based on the degree of deviation.

5. A digital twin model construction system for transformer fault diagnosis, which adopts a digital twin model construction method for transformer fault diagnosis according to any one of claims 1 to 4, characterized in that: include: A registration library construction model is configured to divide the operating status of each device of the transformer into information dimensions, define matching factors for data points under each information dimension, and construct a registration library by setting standard values ​​for each matching factor under each information dimension; The evaluation model is configured to obtain the sampling sequence of each device in the transformer under the operating status, determine the deviation between each sample value in the sampling sequence and the standard value of the corresponding matching factor in the registration library, and obtain a comprehensive score for the corresponding information dimension based on the deviation judgment result, thereby evaluating the usability of the current sampling sequence; The model building model is configured to obtain the target data set based on the availability evaluation results of all sampling sequences, so as to build a digital twin model for transformer fault diagnosis.

6. An electronic device, characterized in that: The method comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method according to any one of claims 1 to 4 is completed.

7. A computer-readable storage medium, characterized in that Used to store computer instructions, which, when executed by a processor, complete the method according to any one of claims 1 to 4.

Images