Low-voltage branch switch power outage warning method and device based on multi-source data fusion

Abstract

The present application relates to the technical field of power systems, and in particular to a low-voltage branch switch power outage warning method and device based on multi-source data fusion. After obtaining multi-source data related to the target low-voltage branch switch, the convolution kernel size of the feature extraction model is determined according to the change trend and data type of each type of data. This not only improves the accuracy of feature extraction and reduces the dependence on manually set parameters, but also enhances the generalization ability of the model on multi-source data. After determining the convolution kernel size, the present application can input various types of data into the corresponding feature extraction model, obtain the corresponding feature vectors, fuse the various feature vectors to form a fusion vector, and input the fusion vector into the power outage warning model to obtain a warning result for the target low-voltage branch switch. When the warning result indicates that there is a power outage risk, a warning signal is issued in time to remind the operation and maintenance personnel to take measures to avoid power outage accidents.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and in particular to a method and device for early warning of power outages of low-voltage branch switches based on multi-source data fusion. Background Art

[0002] Low-voltage branch switches play a vital role in power systems. Their primary function is to control and protect low-voltage circuits, ensuring the safety and reliability of power supply. In industrial, commercial, and residential power applications, low-voltage branch switches are widely used in key locations such as distribution cabinets and branch lines, providing circuit on / off control, overload protection, and short-circuit protection. A failure in a low-voltage branch switch can cause a localized power outage, disrupting production activities, commercial operations, and residents' lives. It can even lead to safety accidents, resulting in significant economic losses and social impact.

[0003] Therefore, it is necessary to provide power outage warnings for low-voltage branch switches to alert operation and maintenance personnel to take appropriate measures, such as arranging maintenance in advance and adjusting operating modes, to avoid power outages. However, existing technologies for power outage warnings for low-voltage branch switches have many shortcomings. For example, traditional warning methods mainly rely on a single data source, such as current or temperature. These data often cannot fully reflect the operating status of the low-voltage branch switch, resulting in inaccurate warning results. In addition, traditional warning methods usually use fixed algorithms or models when processing data, lacking the ability to adapt to data changes, further limiting the accuracy and reliability of the warning. Summary of the Invention

[0004] The purpose of this application is to solve at least one of the above-mentioned technical defects, especially the technical defect of low accuracy and reliability in the power outage warning of low-voltage branch switches in the prior art.

[0005] The present application provides a low-voltage branch switch power outage early warning method based on multi-source data fusion, the method comprising:

[0006] Acquire multi-source data related to the target low-voltage branch switch collected from different dimensions;

[0007] Determining, based on the change and data type of each type of data in the multi-source data, a convolution kernel size in a feature extraction model when extracting features from each type of data, wherein each type of data corresponds to a feature extraction model;

[0008] Input each type of data into the corresponding feature extraction model to obtain the feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector;

[0009] The fusion vector is input into a preset power outage warning model to obtain a warning result of the target low-voltage branch switch output by the power outage warning model.

[0010] Optionally, the acquiring multi-source data collected from different dimensions and related to the target low-voltage branch switch includes:

[0011] Acquire a current signal of the target low-voltage branch switch collected by a current sensor, and calculate the current harmonic content and three-phase imbalance based on the current signal;

[0012] Acquiring the contact temperature of the target low-voltage branch switch collected by a temperature sensor, and the vibration acceleration and vibration frequency of the target low-voltage branch switch collected by a vibration sensor;

[0013] Acquiring the ambient humidity of the installation environment of the target low-voltage branch switch collected by a humidity sensor, and the environmental pollution level of the installation environment collected by a pollution level monitoring device;

[0014] The current harmonic content, the three-phase imbalance, the contact temperature, the vibration acceleration, the vibration frequency, the ambient humidity, and the ambient pollution level are used as multi-source data related to the target low-voltage branch switch.

[0015] Optionally, determining the convolution kernel size in the feature extraction model when extracting features for each type of data according to the change status and data type of each type of data in the multi-source data includes:

[0016] For each type of data in the multi-source data:

[0017] Determine the change index data of this type of data according to the change of this type of data within a preset period;

[0018] Determine type indicator data of the data according to the data type of the data;

[0019] Based on the change index data and type index data of this type of data, the convolution kernel size in the feature extraction model when extracting features for this type of data is determined.

[0020] Optionally, determining the change indicator data of the data according to the change of the data within a preset period includes:

[0021] Determine the local change rate of the data at each moment according to the data value of the data at each moment within a preset period;

[0022] According to the local change rate of this type of data at each moment and the preset period, the change index data of this type of data is determined.

[0023] Optionally, the calculation formula for determining the local change rate of the data of this type at each moment according to the data value of the data of this type at each moment within the preset period is:

[0024]

[0025] in, is the local rate of change at time t, x(t) is the data value at time t, yes The data value at the moment;

[0026] The calculation formula for determining the change index data of this type of data based on the local change rate of this type of data at each moment and the preset period is:

[0027]

[0028] in, For the change indicator data, is the local rate of change at time t, is the local change rate at time t-1, and T is the preset period.

[0029] Optionally, the calculation formula for determining the convolution kernel size in the feature extraction model when extracting features for this type of data based on the change index data and type index data of this type of data is:

[0030]

[0031] in, is the convolution kernel size in the feature extraction model, round() represents the rounding function, and are the maximum and minimum values ​​of the convolution kernel, respectively. Is the adjustment coefficient used to control type indicator data The influence of the change on the convolution kernel size.

[0032] Optionally, before determining the convolution kernel size in the feature extraction model when extracting features for each type of data based on the change status and data type of each type of data in the multi-source data, the method further includes:

[0033] Cleaning the multi-source data to obtain cleaned multi-source data;

[0034] The cleaned multi-source data is normalized to obtain normalized multi-source data.

[0035] The present application also provides a low-voltage branch switch power outage warning device based on multi-source data fusion, comprising:

[0036] A data acquisition module, used to acquire multi-source data related to the target low-voltage branch switch collected from different dimensions;

[0037] a convolution kernel determination module, configured to determine, based on the change and data type of each type of data in the multi-source data, a convolution kernel size in a feature extraction model when extracting features from each type of data, wherein each type of data corresponds to a feature extraction model;

[0038] The feature extraction and fusion module is used to input various types of data into the corresponding feature extraction models, obtain the feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector;

[0039] The power outage warning module is used to input the fusion vector into a preset power outage warning model to obtain the warning result of the target low-voltage branch switch output by the power outage warning model.

[0040] The present application also provides a computer-readable storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any of the above embodiments.

[0041] The present application also provides a computer device, comprising: one or more processors, and a memory;

[0042] The memory stores computer-readable instructions, which, when executed by the one or more processors, execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any one of the above embodiments.

[0043] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0044] The low-voltage branch switch power outage warning method and device based on multi-source data fusion provided by the present application can first obtain multi-source data collected from different dimensions and closely related to the target low-voltage branch switch. These data may include but are not limited to various types of monitoring data such as current, voltage, temperature, and humidity; then, the present application can determine the convolution kernel size in the specific feature extraction model used when extracting features for each type of data based on the changing trend and respective data types of each type of data in these multi-source data. This not only improves the accuracy of feature extraction and ensures that the model can better understand and process each type of data, but also reduces the dependence on manually set parameters; and, by determining the convolution kernel size in the above manner, the feature extraction model can better adapt to the characteristics of different data, thereby improving the generalization ability on multi-source data. Furthermore, each type of data in the present application strictly corresponds to a feature extraction model specially designed for it, so that when each type of data is input into their respective corresponding feature extraction models, the feature vectors output by each feature extraction model can be obtained, and these feature vectors each represent the characteristics of their corresponding data; then, the present application can fuse these feature vectors to form a comprehensive fusion vector, which concentrates the feature information of all data; finally, the present application can input this fusion vector into a pre-set power outage warning model. After the model is calculated and analyzed, the warning result for the target low-voltage branch switch output by the power outage warning model is finally obtained. When the warning result indicates that the target low-voltage branch switch is at risk of power outage, a warning signal can be issued in time to remind the operation and maintenance personnel to take corresponding measures, such as arranging maintenance in advance, adjusting the operation mode, etc., to avoid the occurrence of power outage accidents, thereby providing reliable warning support for the stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0046] Figure 1 A flowchart of a low-voltage branch switch power outage warning method based on multi-source data fusion provided in an embodiment of the present application;

[0047] Figure 2 A schematic diagram of a process for determining the convolution kernel size in a feature extraction model when extracting features for each type of data provided in an embodiment of the present application;

[0048] Figure 3A schematic diagram of the structure of a low-voltage branch switch power outage warning device based on multi-source data fusion provided in an embodiment of the present application;

[0049] Figure 4 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0051] In one embodiment, Figure 1 As shown, Figure 1 A flowchart of a low-voltage branch switch power outage warning method based on multi-source data fusion provided in an embodiment of the present application is provided; the present application provides a low-voltage branch switch power outage warning method based on multi-source data fusion, the method may include:

[0052] S110: Acquire multi-source data collected from different dimensions and related to the target low-voltage branch switch.

[0053] In this step, when issuing a power outage warning for the target low-voltage branch switch, the present application can first obtain multi-source data related to the target low-voltage branch switch collected from different dimensions, so that the multi-source data can be used to predict whether the target low-voltage branch switch has a power outage risk.

[0054] Among them, the target low-voltage branch switch in this application refers to a low-voltage branch switch that requires a power outage warning. The types of low-voltage branch switches include but are not limited to molded case circuit breakers, miniature circuit breakers, leakage circuit breakers, etc. The multi-source data of this application may include but are not limited to various types of monitoring data such as current, voltage, temperature, humidity, vibration, and degree of contamination. These data can comprehensively reflect the operating status of the target low-voltage branch switch. By analyzing and processing these multi-source data, it is possible to more accurately predict whether the target low-voltage branch switch has a power outage risk, thereby improving the accuracy and reliability of the warning.

[0055] S120: Determine the convolution kernel size in the feature extraction model when extracting features for each type of data according to the change and data type of each type of data in the multi-source data, wherein each type of data corresponds to a feature extraction model.

[0056] In this step, after obtaining multi-source data related to the target low-voltage branch switch collected from different dimensions through S110, the present application can determine the convolution kernel size in the feature extraction model when performing feature extraction on each type of data based on the change and data type of each type of data in the multi-source data. In this way, feature extraction can be performed on different types of data according to convolution kernels of different sizes.

[0057] Specifically, after the present application obtains multi-source data related to the target low-voltage branch switch collected from different dimensions, since the multi-source data includes multiple types of data, the present application can determine the corresponding feature extraction model and the convolution kernel size of the convolution layer in the feature extraction model for each type of data to ensure that each feature extraction model can perform the most effective feature extraction for its corresponding data type.

[0058] For example, for current data, due to its periodic and mutational characteristics, this application can design a smaller convolution kernel for it to capture the local characteristics of the current data; and for temperature data, due to its relatively smooth changes, this application can design a larger convolution kernel for it to capture the global characteristics of the temperature data.

[0059] In this way, the present application can fully utilize the feature information of different types of data to improve the accuracy and efficiency of feature extraction. At the same time, since each type of data corresponds to a dedicated feature extraction model, this also avoids the problem of inaccurate feature extraction caused by using a unified model to process different types of data in traditional methods.

[0060] S130: Input each type of data into the corresponding feature extraction model respectively, obtain the feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector.

[0061] In this step, after determining the convolution kernel size in the feature extraction model when extracting features for each type of data through S120, the present application can use each feature extraction model to extract the corresponding type of data respectively, and after obtaining each feature vector, fuse each feature vector to obtain a fusion vector.

[0062] Specifically, after inputting each type of data into the corresponding feature extraction model, each feature extraction model will extract features from the input data based on the size of the convolution kernel designed. It can be understood that the feature extraction process is essentially a convolution operation on the data, extracting the local features of the data by sliding the convolution kernel on the data. Since each feature extraction model in this application is specially designed for its corresponding data type, they can accurately extract the feature information of the corresponding data and form a feature vector.

[0063] After obtaining the individual eigenvectors, the present application can fuse them to form a comprehensive fusion vector. The fusion process can be performed using a variety of methods, such as splicing, weighted summation, and the like. The splicing method is to directly connect the individual eigenvectors together to form a longer vector; the weighted summation method is to assign a weight to each eigenvector based on its importance, and then sum the weighted eigenvectors to obtain a fusion vector. Regardless of the method used, the purpose of the fusion vector is to integrate the information of the individual eigenvectors for subsequent power outage warning analysis.

[0064] S140: Inputting the fusion vector into a preset power outage warning model to obtain a warning result of the target low-voltage branch switch output by the power outage warning model.

[0065] In this step, after the various feature vectors are fused through S130 to obtain the fused vector, the present application can also input the fused vector into a preset power outage warning model, so that the power outage risk of the target low-voltage branch switch can be predicted through the power outage warning model, and the corresponding warning result can be output.

[0066] Among them, the power outage warning model preset in this application can be a support vector machine model. Support vector machine (SVM) is a classification method based on statistical learning theory. It maximizes the intervals between different categories by finding the optimal hyperplane to achieve prediction. Of course, the power outage warning model of this application can also be a neural network model, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). These models can automatically learn the complex features of the data and perform classification or regression prediction. In addition, ensemble learning methods such as random forests, gradient boosting trees, etc. can also be used to combine the prediction results of multiple base classifiers to improve the accuracy of the warning. The specific model selection can be set according to the actual situation and is not limited here.

[0067] Furthermore, when the present application adopts a support vector machine model as a power outage warning model, the model can select a radial basis function (RBF) kernel and set corresponding hyperparameters. Subsequently, the present application can use the training data to train the model, find the optimal hyperplane through an optimization algorithm (such as the sequential minimum optimization algorithm SMO), and adjust the model hyperparameters to improve the model performance. When the training is completed, the present application can input the fusion vector into the trained support vector machine model. The model makes predictions based on the input fusion vector and outputs a warning result. The warning result can be divided into two categories: normal operation of the low-voltage disconnect switch and the risk of power outage of the low-voltage disconnect switch. When the model predicts that there is a risk of power outage of the low-voltage branch switch, it can issue a warning signal in time to remind the operation and maintenance personnel to take corresponding measures, such as arranging maintenance in advance, adjusting the operation mode, etc., to avoid the occurrence of power outage accidents.

[0068] Furthermore, when the power outage warning model is used to predict the warning results of the target low-voltage branch switch, the power outage warning model can output one or more warning levels, such as "low risk", "medium risk", "high risk", etc., so that operation and maintenance personnel can take corresponding measures according to the warning level. The division of warning levels can be set according to actual conditions, for example, the thresholds corresponding to different warning levels are determined based on historical data and expert experience. When the warning result output by the power outage warning model reaches or exceeds a certain threshold, the warning level of the target low-voltage branch switch can be upgraded to the corresponding level. In this way, the operation and maintenance personnel can more intuitively understand the power outage risk of the target low-voltage branch switch, and make more accurate decisions. In addition, the present application can also record and store information such as warning results and warning levels for subsequent analysis and tracing.

[0069] In the above embodiment, first, multi-source data closely related to the target low-voltage branch switch collected from different dimensions can be obtained. These data may include but are not limited to various types of monitoring data such as current, voltage, temperature, and humidity; then, the present application can determine the convolution kernel size in the specific feature extraction model used when extracting features for each type of data based on the changing trend and respective data types of each type of data in these multi-source data. This not only improves the accuracy of feature extraction and ensures that the model can better understand and process each type of data, but also reduces dependence on manually set parameters; and, by determining the convolution kernel size in the above manner, the feature extraction model can better adapt to the characteristics of different data, thereby improving the generalization ability on multi-source data. Furthermore, each type of data in the present application strictly corresponds to a feature extraction model specially designed for it, so that when each type of data is input into their respective corresponding feature extraction models, the feature vectors output by each feature extraction model can be obtained, and these feature vectors each represent the characteristics of their corresponding data; then, the present application can fuse these feature vectors to form a comprehensive fusion vector, which concentrates the feature information of all data; finally, the present application can input this fusion vector into a pre-set power outage warning model. After the model is calculated and analyzed, the warning result for the target low-voltage branch switch output by the power outage warning model is finally obtained. When the warning result indicates that the target low-voltage branch switch is at risk of power outage, a warning signal can be issued in time to remind the operation and maintenance personnel to take corresponding measures, such as arranging maintenance in advance, adjusting the operation mode, etc., to avoid the occurrence of power outage accidents, thereby providing reliable warning support for the stable operation of the power system.

[0070] In one embodiment, the multi-source data collected from different dimensions and related to the target low-voltage branch switch in S110 may include:

[0071] S111: Acquire a current signal of the target low-voltage branch switch collected by a current sensor, and calculate current harmonic content and three-phase imbalance according to the current signal.

[0072] S112: Acquire the contact temperature of the target low-voltage branch switch collected by the temperature sensor, and the vibration acceleration and vibration frequency of the target low-voltage branch switch collected by the vibration sensor.

[0073] S113: Acquire the ambient humidity of the installation environment of the target low-voltage branch switch collected by the humidity sensor, and the environmental pollution level of the installation environment collected by the pollution level monitoring device.

[0074] S114: Using the current harmonic content, the three-phase imbalance, the contact temperature, the vibration acceleration, the vibration frequency, the ambient humidity, and the ambient pollution level as multi-source data related to the target low-voltage branch switch.

[0075] In this embodiment, when acquiring multi-source data related to the target low-voltage branch switch, the present application uses a variety of sensors and monitoring devices to ensure the comprehensiveness and accuracy of the data.

[0076] Specifically, the present application can install a high-precision current sensor in the circuit of the target low-voltage branch switch, which can collect the current signal of the target low-voltage branch switch in real time and accurately during operation. Through these collected current signals, the system further calculates and analyzes key electrical measurement data such as current harmonic content and three-phase imbalance. The calculation process of current harmonic content can use the fast Fourier transform (FFT) algorithm, which can accurately extract the amplitude and phase information of each harmonic component from the complex current signal, and then scientifically characterize the size of the current harmonic content by calculating the total harmonic distortion (THD). The three-phase imbalance can be determined by calculating the degree of difference between the three-phase currents, thereby accurately reflecting the balance state of the three-phase current.

[0077] In addition, to comprehensively monitor the operating status of the low-voltage shunt switch, this application also installs a temperature sensor on the contact of the target low-voltage shunt switch. This temperature sensor can monitor and record the temperature changes of the target low-voltage shunt switch contacts in real time, promptly identifying potential overheating risks. At the same time, this application also installs a vibration sensor on the mechanical components of the target low-voltage shunt switch. This sensor is responsible for collecting mechanical vibration signals and extracting vibration characteristic parameters such as vibration acceleration and vibration frequency through advanced signal processing algorithms, so as to evaluate the health status of the mechanical components.

[0078] Not only that, the present application also takes into account the influencing factors of the installation environment of the low-voltage branch switch. Therefore, the present application installs a humidity sensor and a pollution level monitoring device in the installation environment of the target low-voltage branch switch. The humidity sensor can adopt a capacitive humidity sensor, which can collect environmental humidity data in real time and accurately. The pollution level monitoring device can scientifically evaluate the degree of environmental pollution by measuring key parameters such as the conductivity or pollution deposition density on the surface of the insulator, thereby providing an important reference basis for the maintenance and maintenance of the target low-voltage branch switch. Through these comprehensive monitoring methods, the present application can comprehensively improve the operational safety and reliability of the low-voltage branch switch.

[0079] In one embodiment, Figure 2 As shown, Figure 2 A schematic diagram of a process for determining the convolution kernel size in a feature extraction model when extracting features for each type of data provided in an embodiment of the present application; determining the convolution kernel size in the feature extraction model when extracting features for each type of data based on the change and data type of each type of data in the multi-source data in S120 may include:

[0080] S121: For each type of data in the multi-source data: determine change index data of the type of data according to changes in the type of data within a preset period.

[0081] S122: Determine type indicator data of the data according to the data type of the data.

[0082] S123: Based on the change index data and type index data of the data of this type, determine the convolution kernel size in the feature extraction model when performing feature extraction on the data of this type.

[0083] In this embodiment, when determining the convolution kernel size in the feature extraction model when extracting features for each type of data, the present application can first determine a change index data for each type of data in the multi-source data based on its change within a preset period. This preset period can be set according to actual needs, such as one day, one week, or one month. The change index data can reflect the fluctuation and change trend of this type of data over a period of time, and is one of the important bases for determining the size of the convolution kernel.

[0084] Next, this application can also determine a type indicator data based on the data type of this type of data. The data type can refer to the physical meaning or source of the data, for example, current and voltage are electrical measurement data, temperature and humidity are environmental parameter data, etc. Different types of data have different characteristics and information content. Therefore, type indicator data can help this application better understand the essential attributes of the data and provide a useful reference for determining the size of the convolution kernel.

[0085] For example, the type index data of the multi-source data in this application is the importance index data of each type of data in the multi-source data in the low-voltage branch switch power outage warning; this application can use the AHP hierarchical analysis method or the entropy weight method to determine the type index data of the multi-source data; wherein, the type index data of the multi-source data is shown in the following table:

[0086] Table 1 shows the type indicator data of multi-source data

[0087]

[0088] As can be seen from Table 1, when the multi-source data of the present application includes current harmonic content, three-phase imbalance, contact temperature, vibration acceleration, vibration frequency, ambient humidity and environmental pollution level, the type index data of current harmonic content is 0.15, the type index data of three-phase imbalance is 0.15, the type index data of contact temperature is 0.25, the type index data of vibration acceleration is 0.15, the type index data of vibration frequency is 0.1, the type index data of ambient humidity is 0.1, and the type index data of environmental pollution level is 0.1.

[0089] Finally, this application can comprehensively consider the change indicator data and type indicator data of this type of data, and determine the convolution kernel size in the feature extraction model when extracting features from this type of data through certain algorithms or rules. Moreover, this process is a dynamic adjustment and optimization process, which can ensure that each feature extraction model can perform the most effective feature extraction for its corresponding data type, thereby improving the accuracy and reliability of the warning.

[0090] In one embodiment, determining the change indicator data of the data type in S121 based on the change of the data type within a preset period may include:

[0091] S1211: Determine the local change rate of the data of this type at each moment according to the data value of the data of this type at each moment within a preset period.

[0092] S1212: Determine change index data of this type of data according to the local change rate of this type of data at each moment and the preset period.

[0093] In this embodiment, when determining the change indicator data for each type of data within a preset period, this application first focuses on the local change characteristics of the data at each specific moment. The local change rate can reflect the degree of data fluctuation over a short period of time and is a key indicator for capturing data change trends.

[0094] Specifically, this application can determine the local rate of change of this type of data at each moment within a preset period by calculating the ratio of the difference between the data values ​​at adjacent moments to the data value at the previous moment. This process can generate a local rate of change sequence that records in detail the subtle changes in the data at each point in time.

[0095] Subsequently, the application can further process this local change rate sequence and determine the overall change index data for this type of data within a preset period by statistically analyzing the mean, standard deviation, or other statistical characteristics of the local change rate. This change index data can comprehensively reflect the fluctuation level and change trend of the data over a period of time, providing an important basis for determining the convolution kernel size in the feature extraction model.

[0096] In this way, the present application can more accurately grasp the dynamic characteristics of each type of data, thereby ensuring that the feature extraction model can perform the most effective feature extraction for its corresponding data type.

[0097] In one embodiment, in S1211, the calculation formula for determining the local change rate of the data of this type at each moment in a preset period is as follows:

[0098]

[0099] in, is the local rate of change at time t, x(t) is the data value at time t, yes The data value at the moment.

[0100] As you can understand, the formula for calculating the local rate of change provides a way to quantify the degree of data fluctuation over a short period of time. By calculating the ratio of the difference between data values ​​at adjacent moments to the value at the previous moment, this application can capture subtle changes in the data at each point in time, which is crucial for understanding the dynamic characteristics of data.

[0101] It's worth noting that the calculation of the local rate of change isn't limited to the formula above; this application can flexibly adjust it based on actual needs. For example, when calculating the difference between data values ​​at adjacent moments, this application can use different time intervals or introduce weighting factors to emphasize the importance of data changes at certain moments. These adjustments can further enhance the local rate of change's ability to capture the dynamic characteristics of the data.

[0102] In S1212, the calculation formula for determining the change index data of this type of data is determined based on the local change rate of this type of data at each moment and the preset period:

[0103]

[0104] in, For the change indicator data, is the local rate of change at time t, is the local change rate at time t-1, and T is the preset period.

[0105] It's understandable that the calculation formula for the aforementioned change index, R, comprehensively considers the fluctuations in the local change rate over the time series. By calculating the ratio of the difference between the local change rates at adjacent moments to the local change rate at the previous moment, and accumulating and averaging these differences, we arrive at an indicator that reflects the overall trend of data change over a preset period. This indicator not only considers the local variation characteristics of the data at each point in time, but also the continuity and stability of these local variations over the time series. Therefore, it can more comprehensively and accurately describe the dynamic characteristics of the data.

[0106] After determining the change indicator data and type indicator data for each type of data, the present application may use certain algorithms or rules to combine these two indicator data to determine the convolution kernel size in the feature extraction model when extracting features for this type of data.

[0107] In one embodiment, in S123, based on the change index data and type index data of the data, the calculation formula for determining the convolution kernel size in the feature extraction model when extracting features for the data of this type is:

[0108]

[0109] in, is the convolution kernel size in the feature extraction model, round() represents the rounding function, and are the maximum and minimum values ​​of the convolution kernel, respectively. Is the adjustment coefficient used to control type indicator data The influence of the change on the convolution kernel size.

[0110] In this embodiment, when determining the convolution kernel size in the feature extraction model for each type of data, this application uses an algorithm that comprehensively considers data variation indicators and data type indicators. The core concept of this algorithm is to dynamically adjust the convolution kernel size based on data variation and data type to adapt to the characteristics of different data, thereby improving the feature extraction effect.

[0111] In the above formula, K represents the final convolution kernel size. and These represent the upper and lower limits of the adjustable range of the convolution kernel size. These two values ​​can be set according to actual needs to ensure that the convolution kernel size varies within a reasonable range.

[0112] R represents the data's change indicator, reflecting the data's fluctuation level and trend within a preset period. By calculating the local rate of change and conducting statistical analysis, this application can derive an R metric that quantifies the data's dynamic characteristics. A larger R metric indicates more dramatic data fluctuations and a more pronounced trend, necessitating a larger convolution kernel to capture these characteristics.

[0113] C represents the data type indicator data, which reflects the physical meaning or source of the data and embodies the importance of the data in the low-voltage branch switch power outage warning. In this application, different types of data have different characteristics and information content, so the value of C will also vary. By introducing the type indicator data C, this application can better understand the essential attributes of the data and provide a useful reference for determining the size of the convolution kernel.

[0114] In one embodiment, before determining the convolution kernel size in the feature extraction model when extracting features for each type of data based on the change status and data type of each type of data in the multi-source data in S120, the following steps may also be included:

[0115] S115: Cleaning the multi-source data to obtain cleaned multi-source data.

[0116] S116: performing a normalization operation on the cleaned multi-source data to obtain normalized multi-source data.

[0117] In this embodiment, before determining the convolution kernel size in the feature extraction model for each type of data, preprocessing the multi-source data is a crucial step. The preprocessing process of this application mainly includes two steps: data cleaning and data normalization.

[0118] The purpose of data cleaning is to remove noise, outliers, and missing values ​​from the raw data to ensure data accuracy and reliability. In the low-voltage branch switch power outage warning system, multi-source data may come from different sensors and monitoring devices, so the quality and format of the data may vary. Through data cleaning, this application can identify and correct these differences, resulting in a cleaner and tidier data set. Data cleaning can include operations such as removing duplicate data, filling missing values, and smoothing noisy data.

[0119] Data normalization is to scale the data according to a certain ratio so that it falls into a small specific interval to facilitate subsequent data processing and feature extraction. In multi-source data, different types of data may have different dimensions and value ranges, which may cause the feature extraction model to deviate when processing data. Through data normalization, this application can convert all data to the same scale, thereby improving the consistency and accuracy of feature extraction. Commonly used data normalization methods include minimum-maximum normalization, Z-score standardization, etc. This application can be selected according to actual conditions and is not limited here.

[0120] After completing the data cleaning and normalization operations, this application obtains preprocessed multi-source data. This data is now ready for feature extraction and can be further input into the feature extraction model for subsequent convolution kernel size determination and feature extraction. Through this preprocessing process, this application can ensure the quality and consistency of multi-source data, thereby improving the accuracy and reliability of the low-voltage shunt switch power outage warning system.

[0121] The following describes a low-voltage branch switch power outage warning device based on multi-source data fusion provided in an embodiment of the present application. The low-voltage branch switch power outage warning device based on multi-source data fusion described below and the low-voltage branch switch power outage warning method based on multi-source data fusion described above can be referenced to each other.

[0122] In one embodiment, Figure 3 As shown, Figure 3 This is a schematic diagram of the structure of a low-voltage branch switch power outage warning device based on multi-source data fusion provided in an embodiment of the present application. The present application also provides a low-voltage branch switch power outage warning device based on multi-source data fusion, which can include a data acquisition module 210, a convolution kernel determination module 220, a feature extraction and fusion module 230, and a power outage warning module 240, specifically including the following:

[0123] The data acquisition module 210 is used to acquire multi-source data collected from different dimensions and related to the target low-voltage branch switch.

[0124] The convolution kernel determination module 220 is used to determine the convolution kernel size in the feature extraction model when extracting features for each type of data according to the change and data type of each type of data in the multi-source data, wherein each type of data corresponds to a feature extraction model.

[0125] The feature extraction and fusion module 230 is used to input various types of data into corresponding feature extraction models, obtain feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector.

[0126] The power outage warning module 240 is configured to input the fusion vector into a preset power outage warning model to obtain a warning result of the target low-voltage branch switch output by the power outage warning model.

[0127] In the above embodiment, first, multi-source data closely related to the target low-voltage branch switch collected from different dimensions can be obtained. These data may include but are not limited to various types of monitoring data such as current, voltage, temperature, and humidity; then, the present application can determine the convolution kernel size in the specific feature extraction model used when extracting features for each type of data based on the changing trend and respective data types of each type of data in these multi-source data. This not only improves the accuracy of feature extraction and ensures that the model can better understand and process each type of data, but also reduces dependence on manually set parameters; and, by determining the convolution kernel size in the above manner, the feature extraction model can better adapt to the characteristics of different data, thereby improving the generalization ability on multi-source data. Furthermore, each type of data in the present application strictly corresponds to a feature extraction model specially designed for it, so that when each type of data is input into their respective corresponding feature extraction models, the feature vectors output by each feature extraction model can be obtained, and these feature vectors each represent the characteristics of their corresponding data; then, the present application can fuse these feature vectors to form a comprehensive fusion vector, which concentrates the feature information of all data; finally, the present application can input this fusion vector into a pre-set power outage warning model. After the model is calculated and analyzed, the warning result for the target low-voltage branch switch output by the power outage warning model is finally obtained. When the warning result indicates that the target low-voltage branch switch is at risk of power outage, a warning signal can be issued in time to remind the operation and maintenance personnel to take corresponding measures, such as arranging maintenance in advance, adjusting the operation mode, etc., to avoid the occurrence of power outage accidents, thereby providing reliable warning support for the stable operation of the power system.

[0128] In one embodiment, the present application also provides a computer-readable storage medium, which stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any of the above embodiments.

[0129] In one embodiment, the present application further provides a computer device, including: one or more processors, and a memory.

[0130] The memory stores computer-readable instructions, which, when executed by the one or more processors, execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any one of the above embodiments.

[0131] Schematically, as Figure 4 As shown, Figure 4 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 can be provided as a server. Figure 4 Computer device 300 includes a processing component 302, which further includes one or more processors, and memory resources represented by memory 301 for storing instructions executable by processing component 302, such as application programs. The application programs stored in memory 301 may include one or more modules, each corresponding to a set of instructions. Furthermore, processing component 302 is configured to execute the instructions to implement the low-voltage branch switch power outage warning method based on multi-source data fusion according to any of the above-described embodiments.

[0132] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or the like.

[0133] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0134] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0135] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referenced to each other.

[0136] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A low-voltage branch switch power outage warning method based on multi-source data fusion, characterized in that: The method comprises: Acquire multi-source data related to the target low-voltage branch switch collected from different dimensions; Determining, based on the change and data type of each type of data in the multi-source data, a convolution kernel size in a feature extraction model when extracting features from each type of data, wherein each type of data corresponds to a feature extraction model; Input each type of data into the corresponding feature extraction model to obtain the feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector; Inputting the fusion vector into a preset power outage warning model to obtain a warning result of the target low-voltage branch switch output by the power outage warning model; The acquiring of multi-source data related to the target low-voltage branch switch collected from different dimensions includes: Acquire a current signal of the target low-voltage branch switch collected by a current sensor, and calculate the current harmonic content and three-phase imbalance based on the current signal; Acquiring the contact temperature of the target low-voltage branch switch collected by a temperature sensor, and the vibration acceleration and vibration frequency of the target low-voltage branch switch collected by a vibration sensor; Acquiring the ambient humidity of the installation environment of the target low-voltage branch switch collected by a humidity sensor, and the environmental pollution level of the installation environment collected by a pollution level monitoring device; using the current harmonic content, the three-phase unbalance, the contact temperature, the vibration acceleration, the vibration frequency, the ambient humidity, and the ambient pollution level as multi-source data related to the target low-voltage branch switch; The step of determining the convolution kernel size in the feature extraction model when extracting features for each type of data according to the change status and data type of each type of data in the multi-source data includes: For each type of data in the multi-source data: Determine the change index data of this type of data according to the change of this type of data within a preset period; Determine type indicator data of the data according to the data type of the data; Based on the change index data and type index data of this type of data, the convolution kernel size in the feature extraction model when extracting features for this type of data is determined.

2. The low-voltage branch switch power outage early warning method based on multi-source data fusion according to claim 1 is characterized in that: Determining the change index data of this type of data according to the change of this type of data within a preset period includes: Determine the local change rate of the data at each moment according to the data value of the data at each moment within a preset period; According to the local change rate of this type of data at each moment and the preset period, the change index data of this type of data is determined.

3. The low-voltage branch switch power outage early warning method based on multi-source data fusion according to claim 2 is characterized in that: The calculation formula for determining the local change rate of this type of data at each moment based on the data value of this type of data at each moment within the preset period is: ； in, is the local rate of change at time t, x(t) is the data value at time t, yes The data value at the moment; The calculation formula for determining the change index data of this type of data based on the local change rate of this type of data at each moment and the preset period is: ； in, For the change indicator data, is the local rate of change at time t, is the local change rate at time t-1, and T is the preset period.

4. The low-voltage branch switch power outage early warning method based on multi-source data fusion according to claim 1 is characterized in that: The calculation formula for determining the convolution kernel size in the feature extraction model when extracting features for this type of data based on the change index data and type index data of this type of data is: ； in, is the convolution kernel size in the feature extraction model, round() represents the rounding function, and are the maximum and minimum values ​​of the convolution kernel, respectively. Is the adjustment coefficient used to control type indicator data The influence of the change on the convolution kernel size.

5. The low-voltage branch switch power outage warning method based on multi-source data fusion according to any one of claims 1 to 4, characterized in that: Before determining the convolution kernel size in the feature extraction model when extracting features for each type of data based on the change status and data type of each type of data in the multi-source data, the method further includes: Cleaning the multi-source data to obtain cleaned multi-source data; The cleaned multi-source data is normalized to obtain normalized multi-source data.

6. A low-voltage branch switch power outage warning device based on multi-source data fusion, characterized in that: include: A data acquisition module, used to acquire multi-source data related to the target low-voltage branch switch collected from different dimensions; a convolution kernel determination module, configured to determine, based on the change and data type of each type of data in the multi-source data, a convolution kernel size in a feature extraction model when extracting features from each type of data, wherein each type of data corresponds to a feature extraction model; The feature extraction and fusion module is used to input various types of data into the corresponding feature extraction models, obtain the feature vectors output by each feature extraction model, and fuse each feature vector to obtain a fusion vector; a power outage warning module, configured to input the fusion vector into a preset power outage warning model to obtain a warning result of the target low-voltage branch switch output by the power outage warning model; The data acquisition module includes: Acquire a current signal of the target low-voltage branch switch collected by a current sensor, and calculate the current harmonic content and three-phase imbalance based on the current signal; Acquiring the contact temperature of the target low-voltage branch switch collected by a temperature sensor, and the vibration acceleration and vibration frequency of the target low-voltage branch switch collected by a vibration sensor; Acquiring the ambient humidity of the installation environment of the target low-voltage branch switch collected by a humidity sensor, and the environmental pollution level of the installation environment collected by a pollution level monitoring device; using the current harmonic content, the three-phase unbalance, the contact temperature, the vibration acceleration, the vibration frequency, the ambient humidity, and the ambient pollution level as multi-source data related to the target low-voltage branch switch; The convolution kernel determination module includes: For each type of data in the multi-source data: Determine the change index data of this type of data according to the change of this type of data within a preset period; Determine type indicator data of the data according to the data type of the data; Based on the change index data and type index data of this type of data, the convolution kernel size in the feature extraction model when extracting features for this type of data is determined.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by one or more processors, enable the one or more processors to execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any one of claims 1 to 5.

8. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, execute the steps of the low-voltage branch switch power outage warning method based on multi-source data fusion as described in any one of claims 1 to 5.

Images