A transformer micro-water data prediction method and system

Abstract

The application discloses a transformer micro-water data prediction method and system, and the method comprises the following steps: obtaining micro-water sample data, and pre-processing the micro-water sample data; extracting time characteristics in the pre-processed micro-water sample data, and obtaining data variation rules based on the time characteristics; and obtaining a prediction result of the micro-water data by using an extreme gradient boosting model according to the pre-processed micro-water sample data and the data variation rules. The application extracts time characteristics and analyzes data variation rules, combines the extreme gradient boosting model for prediction, assigns weights to the micro-water sample data to improve the prediction accuracy, and enhances the generalization ability of the model. The application generates a micro-water data variation rule graph and a prediction graph, directly displays data dynamics and prediction trends, provides a scientific basis for transformer maintenance and management, effectively reduces the fault risk caused by excessive micro-water content, and improves the overall stability and safety of the power system.

Description

Technical Field

[0001] This invention relates to the field of transformer data prediction technology, and in particular to a method and system for predicting transformer micro-water levels. Background Technology

[0002] Moisture content plays a crucial role in power systems, directly impacting the performance of insulation materials and the stable operation of transformers. Specifically, moisture content is a direct indicator of the degree of moisture absorption in insulation materials. Ideally, when moisture content is maintained at a low level, the insulation resistance of the insulation material remains at a high value, effectively preventing accidental current leakage and ensuring the safe and reliable operation of the transformer under rated parameters. However, once the moisture content exceeds a preset safety threshold, the insulation performance of the insulation material will decline sharply. This not only weakens its ability to block current but also significantly increases the probability of transformer failure, potentially even triggering serious power accidents.

[0003] Therefore, how to accurately and efficiently predict the moisture content in transformers has become a critical issue that urgently needs to be addressed in the field of power technology. This not only requires researchers to have a deep understanding of the complex relationship between moisture content and insulation performance, but also necessitates the development of advanced and reliable monitoring and prediction technologies to achieve real-time and precise control of the moisture content in transformers, thereby ensuring the overall safety and stability of the power system. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, this invention is proposed. Therefore, this invention provides a method for predicting micro-water content in transformers to solve the problem of low accuracy in existing methods for predicting micro-water content in transformers.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for predicting transformer micro-water levels, comprising:

[0008] Obtain micro-water sample data and preprocess the micro-water sample data;

[0009] Extract the temporal features from the preprocessed micro-water sample data, and obtain the data change pattern based on the temporal features;

[0010] Based on the preprocessed micro-water sample data and the data variation patterns, the prediction results of the micro-water data are obtained using an extreme gradient boosting model.

[0011] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, the prediction results of micro-water data obtained by using the extreme gradient boosting model include:

[0012] The preprocessed micro-water sample data and the data variation pattern are input into the extreme gradient boosting model, and weights are assigned to the preprocessed micro-water sample data.

[0013] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, it further includes:

[0014] The preprocessed micro-water sample data and the data change patterns are divided into training and testing sets in the extreme gradient boosting model according to a first preset ratio.

[0015] A portion of micro-water sample data and data variation patterns are selected from the training set according to a second preset ratio, and used as a sampling set.

[0016] The extreme gradient boosting model is trained using the training set, and the trained extreme gradient boosting model is used for evaluation on the test set.

[0017] Several decision trees are constructed based on the training set, the sampling set, and the test set, and several predicted values ​​are output through the several decision trees;

[0018] The extreme gradient boosting model integrates the predicted values ​​to output the prediction results of the micro-water data.

[0019] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, the prediction results include:

[0020] The system monitors the transformer's real-time temperature, operating parameters, and moisture levels.

[0021] Analyze the correlation between the real-time temperature information and the real-time micro-water data;

[0022] Analyze the relationship between the real-time operating parameter information and the real-time micro-water data;

[0023] Based on the correlation between the real-time temperature information and the real-time micro-water data, and the relationship between the real-time operating parameter information and the real-time micro-water data, the prediction results are optimized.

[0024] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, the method involves: extracting time features from the preprocessed micro-water sample data, and obtaining the data change patterns based on the time features, including...

[0025] Record the timestamps in the preprocessed micro-water sample data;

[0026] The time interval between the preprocessed micro-water sample data is obtained based on the timestamp.

[0027] The data change pattern is obtained based on the timestamp and the time interval.

[0028] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, it further includes:

[0029] A micro-water data change pattern diagram is generated based on the preprocessed micro-water sample data and the data change pattern.

[0030] After the step of obtaining the prediction results of micro-water data based on the extreme gradient boosting model, the method further includes:

[0031] A micro-water data prediction map is generated based on the prediction results.

[0032] As a preferred embodiment of the transformer micro-water data prediction method of the present invention, the preprocessing of the micro-water sample data includes:

[0033] The micro-water sample data were analyzed;

[0034] If duplicate values ​​exist in the micro-water sample data, then the duplicate values ​​are deleted;

[0035] If there are missing values ​​in the micro-water sample data, the missing values ​​are deleted, or the mean of the values ​​before and after the missing value is taken to fill the missing value.

[0036] If there are outliers in the micro-water sample data, the outliers are deleted, or the average of the values ​​before and after the outlier is used to replace the outlier.

[0037] Secondly, the present invention provides a system for predicting transformer micro-water data, comprising,

[0038] The data acquisition module is used to acquire micro-water sample data;

[0039] The preprocessing module is used to preprocess the micro-water sample data;

[0040] The feature extraction module is used to extract the time features from the preprocessed micro water sample data and obtain the data change pattern based on the time features;

[0041] The prediction module is used to obtain the prediction results of the micro-water data based on the preprocessed micro-water sample data and the data change pattern using an extreme gradient boosting model.

[0042] Thirdly, the present invention provides a computing device, comprising:

[0043] Memory and processor;

[0044] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the transformer micro-water data prediction method.

[0045] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the transformer micro-water data prediction method.

[0046] Compared with existing technologies, the beneficial effects of this invention are as follows: By extracting time features and analyzing data change patterns, this invention combines extreme gradient boosting models for prediction. It not only assigns weights to micro-water sample data to improve prediction accuracy, but also enhances the model's generalization ability by constructing multi-decision tree integrated prediction results through the division of training, testing, and sampling sets. Furthermore, this invention considers the correlation between transformer real-time temperature information, operating parameter information, and micro-water data, further optimizing the prediction results. It generates micro-water data change pattern graphs and prediction graphs, intuitively displaying data dynamics and prediction trends, providing a scientific basis for transformer maintenance and management, effectively reducing the risk of faults caused by excessive micro-water content, and improving the overall stability and safety of the power system. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0048] Figure 1 This is a schematic diagram of the overall process of the transformer micro-water data prediction method according to an embodiment of the present invention. Detailed Implementation

[0049] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0050] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0052] This invention is described in detail with reference to the schematic diagrams. When describing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0053] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the system or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0054] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0055] Example 1

[0056] Reference Figure 1As an embodiment of the present invention, a method for predicting transformer micro-water data is provided, comprising:

[0057] S100: Acquire micro-water sample data and preprocess the micro-water sample data;

[0058] It should be noted that trace amounts of water in transformer insulating oil and solid insulating materials exist in dissolved, emulsified, or free forms. Trace water sample data refers to relatively accurate and complete historical data on trace water recorded and preserved through various methods. Understandably, some historical data may contain inaccuracies or incompleteness, and these low-quality sample data can affect the accuracy and reliability of the prediction model. For example, human error or equipment malfunction during data recording may lead to errors in the recording of trace water data, thus affecting subsequent predictions. Therefore, data preprocessing is necessary.

[0059] Preferably, the micro-water sample data is tested; if there are duplicate values ​​in the micro-water sample data, the duplicate values ​​are deleted; if there are missing values ​​in the micro-water sample data, the missing values ​​are deleted, or the mean of the values ​​before and after the missing values ​​is taken to fill the missing values; if there are outliers in the micro-water sample data, the outliers are deleted, or the mean of the values ​​before and after the outliers is taken to replace the outliers.

[0060] In one possible embodiment, the occurrence of duplicate values ​​typically includes those caused by measurement errors. For example, some traditional sensors may only be accurate to a certain degree, such as ±5 ppm (parts per million). When the actual moisture content fluctuates within the measurement accuracy range, duplicate measurements may occur. Furthermore, if the sensor has poor stability, its zero-point drift or measurement sensitivity may change after long-term operation, leading to repeated inaccurate measurements. The occurrence of duplicate values ​​also includes those caused by the internal stability of the transformer. For example, a well-sealed transformer operating under a stable load may maintain a relatively constant moisture content for a period of time. The occurrence of duplicate values ​​also includes those caused by data recording and processing. For example, write errors in the storage database or abnormal data caching mechanisms may cause previously received data to be rewritten to the storage unit. During data transmission, signal loss or repeated transmission may also result in the receiving end receiving duplicate moisture data. It is understood that the occurrence of duplicate values ​​includes, but is not limited to, the reasons described above.

[0061] In one possible embodiment, the occurrence of missing values ​​typically includes missing values ​​caused by sensor malfunction, such as interference with communication lines and signal transmission interruption in harsh electromagnetic environments, preventing the transmission of micro-water data from the sensor to the data storage device; missing values ​​also include missing values ​​caused by environmental factors, such as degraded sensor battery performance or unstable performance of electronic components at extremely low temperatures, which may lead to incomplete or missing data acquisition; missing values ​​also include missing values ​​caused by human factors, such as accidentally deleting files containing micro-water data when cleaning the data storage device, or failing to migrate some data successfully due to operational errors during data migration. It is understood that the occurrence of missing values ​​includes, but is not limited to, the reasons described above.

[0062] In one possible embodiment, outliers typically include those caused by measurement errors, such as inaccurate standard materials used in calibration or non-compliant calibration procedures, which can lead to deviations in sensor measurements. After incorrect calibration, the sensor may measure the actual normal trace moisture content as excessively high or low. Outliers can also be caused by internal transformer faults, such as problems with the transformer's sealing structure, like aging or cracked gaskets or fissures in the welded joints, allowing a large amount of external moisture to enter the transformer and causing a sharp increase in trace moisture content, resulting in outliers. This can cause trace moisture data to far exceed the normal range, seriously threatening the transformer's insulation performance. Outliers can also occur during data processing or transmission, such as signal attenuation in the transmission line, noise interference, or errors in the data transmission protocol, all of which can lead to errors in the received data and generate outliers. It is understood that the occurrence of outliers includes, but is not limited to, the reasons described above.

[0063] In this embodiment, the preprocessing steps for micro-water sample data include: calculating the proportion of missing values; if the total data volume is large and the proportion of missing values ​​is extremely small, the records containing missing values ​​can be directly deleted; if the data is to be retained, the average of the preceding and following values ​​can be used to fill the missing values; marking values ​​that violate prior business knowledge as outliers, such as total hydrocarbon content which cannot be negative, and calculating the proportion of outliers; if the total data volume is large and the proportion of outliers is extremely small, the records containing outliers can be directly deleted; if the data is to be retained, the average of the preceding and following values ​​can be used to fill the outliers; for duplicate values, simply delete the duplicate values.

[0064] S102: Extract the temporal features from the pre-processed micro-water sample data, and obtain the data change pattern based on the temporal features;

[0065] It should be noted that time characteristics can be divided into continuous time characteristics and discrete time characteristics. In actual working conditions, time characteristics are mainly discrete time characteristics.

[0066] In one possible implementation, in order to understand the changing trend of micro-water data over the next week, the collection time is divided into the following characteristics using a time decomposition method: which week, which day, which hour of each day, which week of each month, which day of each month, and which day of each week. These data can capture the patterns of data change within a short period.

[0067] In this embodiment, the time feature includes at least a timestamp and a time interval. A timestamp is a number representing a specific moment, typically measured in seconds or milliseconds, calculated from a reference point, and used to record the exact time of the specific moisture content in the transformer at a given moment. A time interval refers to the period between two different instants, which can be viewed as a line segment with a beginning and an end. In data analysis and data science, understanding the time interval is crucial for analyzing trends, patterns, and behaviors over a period of time. Therefore, the data variation pattern can be obtained by recording the timestamps in the preprocessed moisture sample data; obtaining the time intervals between the preprocessed moisture sample data based on the timestamps; and then using the timestamps and time intervals.

[0068] Preferably, the timestamps in the pre-processed micro water sample data are recorded; the time intervals between the pre-processed micro water sample data are obtained based on the timestamps; and the data change patterns are obtained based on the timestamps and time intervals.

[0069] Preferably, a micro-water data change pattern diagram is generated based on the pre-processed micro-water sample data and the data change pattern;

[0070] Preferably, a micro-water data prediction map is generated based on the prediction results.

[0071] Specifically, the timestamps of each preprocessed micro-water sample data are recorded. Micro-water sample data with the same time interval are then filtered out using these timestamps. For example, the timestamps of the first data point are January 1, 2024, 00:00; the second data point is January 2, 2024, 00:00; the third data point is January 3, 2024, 00:00, and so on, with each data point having a one-day time interval. Alternatively, the timestamps of the first data point are January 1, 2024, 00:00; the second data point is January 1, 2024, 08:00; the third data point is January 1, 2024, 16:00, and so on, with each data point having an eight-hour time interval. Understandably, by using the recorded timestamps and micro-water sample data with the same time intervals, time-based data variation patterns can be obtained. Furthermore, a time-based micro-water data variation pattern graph can be generated based on the preprocessed micro-water sample data and the time-based data variation patterns.

[0072] S104: Based on the preprocessed micro-water sample data and the data change pattern, the prediction results of micro-water data are obtained by using the extreme gradient boosting model;

[0073] It should be noted that the extreme gradient boosting model is an algorithm based on gradient boosting decision trees. It internally implements the gradient boosting tree model and has made many optimizations to the algorithm in the model, achieving high accuracy while maintaining extremely fast speed.

[0074] In this embodiment of the application, establishing an extreme gradient boosting model includes initializing the parameters of the extreme gradient boosting model according to business needs and data characteristics; inputting preprocessed micro-water sample data and data change patterns into the extreme gradient boosting model, and assigning weights to the preprocessed micro-water sample data.

[0075] In one possible embodiment, the extreme gradient boosting model can accept various data as inputs; for example, when predicting the trend of micro-water data in transformers, continuous variables such as transformer oil temperature (because oil temperature affects the solubility of micro-water), running time, and load rate can be used as input features. These numerical features can provide the model with rich information to establish the relationship between input and output. In this application, the input is preprocessed micro-water sample data and the data change pattern.

[0076] In this embodiment, a weight is assigned to each micro-water sample data, which is very useful in some cases, such as when there is a data imbalance problem. For example, if when collecting micro-water data of a transformer, the data collected under normal transformer conditions is much more than the data collected under abnormal transformer conditions, in order to make the model pay more attention to the data under abnormal conditions, the samples under abnormal conditions can be assigned higher weights. In this way, the model will pay more attention to these samples during the training process, thereby better learning the patterns of abnormal conditions.

[0077] In this embodiment of the application, the preprocessed micro-water sample data and data change patterns are input into the extreme gradient boosting model, and a series of algorithms are used to obtain the prediction results of the micro-water data. The prediction results include micro-water prediction data for at least a period of time in the future. It can be understood that after obtaining the prediction results, a micro-water data prediction map can be generated based on the prediction results.

[0078] Preferably, the preprocessed micro-water sample data and data change patterns are input into the extreme gradient boosting model, and weights are assigned to the preprocessed micro-water sample data.

[0079] Preferably, the preprocessed micro-water sample data and data variation patterns are divided into a training set and a test set in the extreme gradient boosting model according to a first preset ratio; a portion of micro-water sample data and data variation patterns are selected from the training set according to a second preset ratio as a sampling set; the extreme gradient boosting model is trained using the training set, and the trained extreme gradient boosting model is used for evaluation on the test set; several decision trees are constructed based on the training set, the sampling set, and the test set, and several predicted values ​​are output through the several decision trees; the several predicted values ​​are integrated through the extreme gradient boosting model to output the prediction result of the micro-water data;

[0080] In this embodiment of the application, the training set is a dataset used to train the model. It contains input features and corresponding labels. The model learns from this data to discover patterns and relationships in the data. The test set is another part of the data that is independent of the training set. It is used to evaluate the generalization ability of the model. The test set should contain data that does not appear in the training set.

[0081] In this embodiment of the application, the first preset ratio is the ratio of dividing the sample data into a training set and a test set. For example, the first preset ratio can be set to 8 to 2.

[0082] In this embodiment of the application, the sampling set refers to a portion of the dataset randomly selected from the training set when constructing each tree, including sample sampling, i.e., selecting a portion of micro-water sample data from the training set, and feature sampling, i.e., extracting corresponding sample features from the data change patterns.

[0083] In this embodiment of the application, the second preset ratio is to randomly extract a certain proportion (e.g., 70%) of sample data and sample features from the training set as a sampling set before each iteration builds the decision tree.

[0084] Specifically, using the training set data, multiple decision trees are constructed iteratively. Each tree attempts to correct the prediction error of the previous tree; this process is called gradient boosting. After the extreme gradient boosting model is trained, it can be used to predict the test set data to obtain the prediction results. The prediction results are then compared with the sample data of the test set. Various performance metrics, such as mean squared error and mean absolute error in regression tasks, and accuracy and recall in classification tasks, are calculated to evaluate the model's performance on unseen data, in order to determine whether the model is overfitting or has good generalization ability.

[0085] In this embodiment, the decision tree is integrated throughout the training and prediction processes. During training, the decision tree grows from the root node. Based on the data in the training set (or sample set) and the objective function information, the optimal splitting features and splitting points are selected by calculating the gain, and the internal nodes of the decision tree are gradually constructed. Nodes are continuously split until the stopping condition is met, forming a complete decision tree structure, including leaf nodes. The values ​​of the leaf nodes are updated according to the training set (or sample set) data, enabling the decision tree to make reasonable predictions on the training data. Multiple decision trees are constructed sequentially through an iterative training process. Each decision tree learns the patterns in the training data to a certain extent and is constrained by hyperparameters during the construction process to avoid overfitting.

[0086] It should be noted that the specific process of outputting several predicted values ​​through several decision trees is as follows: when a new data sample is input (such as the sample data in the test set), each decision tree starts from the root node and guides the sample step by step to a leaf node according to the characteristics of the sample and the splitting rules of its internal nodes. Each leaf node of the decision tree will output a predicted value.

[0087] In this embodiment of the application, the prediction results of all decision trees are integrated (e.g., by weighted summation or voting) to obtain the prediction results of the extreme gradient boosting model for micro water data. It can be understood that the integration method improves the generalization ability and prediction accuracy of the model by combining multiple prediction results.

[0088] Preferably, the method involves detecting the transformer's real-time temperature information, real-time operating parameter information, and real-time micro-water data; analyzing the correlation between real-time temperature information and real-time micro-water data; analyzing the relationship between real-time operating parameter information and real-time micro-water data; and optimizing the prediction results based on the correlation between real-time temperature information and real-time micro-water data, as well as the relationship between real-time operating parameter information and real-time micro-water data.

[0089] It should be noted that temperature has a significant impact on the form and content of trace water in transformer oil. Generally, as temperature increases, the solubility of water in the oil increases, and the trace water content may increase; conversely, as temperature decreases, the solubility of water decreases, and the trace water content may decrease. Therefore, it is necessary to collect real-time temperature information during transformer operation and analyze the correlation between temperature and trace water content.

[0090] It should also be noted that the transformer's operating conditions can affect the generation and accumulation of trace moisture. For example, changes in transformer load, partial discharge, and insulation aging can all lead to changes in trace moisture content. Monitoring transformer operating parameters, such as load current, voltage, and partial discharge levels, and analyzing the relationship between these parameters and trace moisture content can help predict trends in trace moisture data. If a transformer operates under high load for an extended period, it may accelerate the aging of insulation materials, thereby generating more trace moisture.

[0091] In this embodiment of the application, based on the correlation between real-time temperature information and real-time micro-water data, and the relationship between real-time operating parameter information and real-time micro-water data, inappropriate or unreasonable prediction data in the prediction results can be identified. For example, if the transformer temperature rises due to a decrease in heat dissipation capacity within the next 6 hours, the increase in transformer temperature usually leads to higher micro-water data. However, if the prediction data for the next 6 hours in the prediction results shows a downward trend, then unreasonable prediction data in the prediction results can be identified, and these unreasonable prediction data can be optimized.

[0092] It should be noted that a method for predicting transformer micro-water data is provided, which involves acquiring several micro-water sample data; preprocessing the micro-water sample data; extracting time features from the preprocessed micro-water sample data; and obtaining the data change patterns based on the time features, which at least include timestamps and time intervals; and obtaining the prediction results of the micro-water data based on the preprocessed micro-water sample data and the data change patterns, using an extreme gradient boosting model. By preprocessing the micro-water sample data to obtain its time features, and then using an extreme gradient boosting model to obtain the prediction results of the micro-water data, the accuracy of the prediction is improved.

[0093] The above is an illustrative scheme of a transformer micro-water data prediction method according to this embodiment. It should be noted that the technical solution of the transformer micro-water data prediction system and the technical solution of the transformer micro-water data prediction method described above belong to the same concept. For details not described in detail in the technical solution of the transformer micro-water data prediction system in this embodiment, please refer to the description of the technical solution of the transformer micro-water data prediction method described above.

[0094] The transformer micro-water data prediction system in this embodiment includes:

[0095] The data acquisition module is used to acquire micro-water sample data;

[0096] The preprocessing module is used to preprocess micro-water sample data;

[0097] The feature extraction module is used to extract the time features from the preprocessed micro water sample data and obtain the data change patterns based on the time features;

[0098] The prediction module is used to obtain the prediction results of micro-water data based on the preprocessed micro-water sample data and the data change patterns, using an extreme gradient boosting model.

[0099] This embodiment also provides a computing device suitable for predicting transformer micro-water data, including:

[0100] The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement the transformer micro-water data prediction method proposed in the above embodiments.

[0101] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the transformer micro-water data prediction method as proposed in the above embodiments.

[0102] The storage medium proposed in this embodiment and the method for predicting transformer micro-water data proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0103] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

Claims

1. A method for predicting micro-water levels in transformers, characterized in that, include: Obtain micro-water sample data and preprocess the micro-water sample data; Extract the temporal features from the preprocessed micro-water sample data, and obtain the data change pattern based on the temporal features; Based on the preprocessed micro-water sample data and the data change pattern, the prediction results of the micro-water data are obtained using the extreme gradient boosting model. Among them, the prediction results of micro-water data obtained using the extreme gradient boosting model include: The preprocessed micro-water sample data and the data variation pattern are input into the extreme gradient boosting model, and weights are assigned to the preprocessed micro-water sample data. The system monitors the transformer's real-time temperature, operating parameters, and moisture levels. Analyze the correlation between the real-time temperature information and the real-time micro-water data; Analyze the relationship between the real-time operating parameter information and the real-time micro-water data; Based on the correlation between the real-time temperature information and the real-time micro-water data, and the relationship between the real-time operating parameter information and the real-time micro-water data, the prediction results are optimized.

2. The transformer micro-water data prediction method as described in claim 1, characterized in that, It also includes, The preprocessed micro-water sample data and the data change patterns are divided into training and testing sets in the extreme gradient boosting model according to a first preset ratio. A portion of micro-water sample data and data variation patterns are selected from the training set according to a second preset ratio, and used as a sampling set. The extreme gradient boosting model is trained using the training set, and the trained extreme gradient boosting model is used for evaluation on the test set. Several decision trees are constructed based on the training set, the sampling set, and the test set, and several predicted values ​​are output through the several decision trees; The extreme gradient boosting model integrates the predicted values ​​to output the prediction results of the micro-water data.

3. The transformer micro-water data prediction method as described in claim 2, characterized in that, Extracting temporal features from preprocessed micro-water sample data, and obtaining data variation patterns based on these temporal features, including: Record the timestamps in the preprocessed micro-water sample data; The time interval between the preprocessed micro-water sample data is obtained based on the timestamp. The data change pattern is obtained based on the timestamp and the time interval.

4. The transformer micro-water data prediction method as described in claim 3, characterized in that, It also includes, A micro-water data change pattern diagram is generated based on the preprocessed micro-water sample data and the data change pattern. After the step of obtaining the prediction results of micro-water data based on the extreme gradient boosting model, the method further includes: A micro-water data prediction map is generated based on the prediction results.

5. The transformer micro-water data prediction method as described in claim 1, characterized in that, Preprocessing of the micro-water sample data includes: The micro-water sample data were analyzed; If duplicate values ​​exist in the micro-water sample data, then the duplicate values ​​are deleted; If there are missing values ​​in the micro-water sample data, the missing values ​​are deleted, or the mean of the values ​​before and after the missing value is taken to fill the missing value. If there are outliers in the micro-water sample data, the outliers are deleted, or the average of the values ​​before and after the outlier is used to replace the outlier.

6. A system for predicting micro-water levels in transformers, characterized in that, include, The data acquisition module is used to acquire micro-water sample data; The preprocessing module is used to preprocess the micro-water sample data; The feature extraction module is used to extract the time features from the preprocessed micro water sample data and obtain the data change pattern based on the time features; The prediction module is used to obtain the prediction results of the micro water data based on the preprocessed micro water sample data and the data change pattern using an extreme gradient boosting model. Among them, the prediction results of micro-water data obtained using the extreme gradient boosting model include: The preprocessed micro-water sample data and the data variation pattern are input into the extreme gradient boosting model, and weights are assigned to the preprocessed micro-water sample data. The system monitors the transformer's real-time temperature, operating parameters, and moisture levels. Analyze the correlation between the real-time temperature information and the real-time micro-water data; Analyze the relationship between the real-time operating parameter information and the real-time micro-water data; Based on the correlation between the real-time temperature information and the real-time micro-water data, and the relationship between the real-time operating parameter information and the real-time micro-water data, the prediction results are optimized.

7. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the transformer micro-water data prediction method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the transformer micro-water data prediction method according to any one of claims 1 to 5.

Images