Relay protection equipment state evaluation system and method based on data analysis

Abstract

This invention discloses a data analysis-based relay protection equipment condition assessment system and method, relating to the field of industrial data analysis technology. The system acquires historical inspection records of relay protection equipment in substations, establishes a fault database, collects current operating status data of target equipment, calculates the error between the status data and standard values ​​to obtain the operating status assessment value of the target equipment, combines the anomaly recognition results of the target equipment's panel pressure plate image to assess the current abnormal state of the target equipment, collects the target equipment's target semantic features, compares the target semantic features with a set of defective equipment, calculates the similarity between the target equipment and the defective equipment, and performs a weighted calculation with the current abnormal state assessment value to obtain the overall assessment value of the target equipment. Based on the sequence of overall assessment values, the system predicts the future overall assessment value of the target equipment and triggers corresponding early warnings and maintenance decision suggestions based on the overall assessment value.

Description

Technical Field

[0001] This invention relates to the field of industrial data processing technology, specifically to a system and method for assessing the condition of relay protection equipment based on data analysis. Background Technology

[0002] With the deepening of smart grid construction, relay protection equipment, as a core device for ensuring the safe and stable operation of the power system, has seen its operation and maintenance management level become an increasingly important focus of the industry. Currently, substations generally use a periodic manual inspection model to conduct functional tests and status checks on relay protection equipment. This model relies on on-site operation by maintenance personnel and the recording of test data using paper forms. This process is not only cumbersome and inefficient, but also prone to problems such as information errors, version inconsistencies, and difficulties in traceability during data collection, aggregation, and archiving.

[0003] The condition assessment of relay protection equipment relies heavily on a comprehensive review of its family-related defects, setting configuration consistency, and switchboard activation / deactivation status. In existing technologies, family-related defect verification primarily relies on manual comparison of equipment models, batches, and historical defect reports, which is prone to omissions of critical defects due to information delays or human error. Furthermore, the verification of protection settings and switchboard switches requires item-by-item visual inspection, which is labor-intensive, subjective, and cannot ensure strict consistency between operating parameters and design requirements. Especially in large-scale substation group operation and maintenance scenarios, manual verification is insufficient for cross-site, multi-device collaborative analysis and risk warning, significantly weakening the overall reliability of the relay protection system. Therefore, there is an urgent need to develop a data analysis-based method for relay protection equipment condition assessment to automate defect verification, improve the accuracy of setting and switchboard verification, and dynamically assess condition risks, thereby comprehensively improving the safe operation level of the relay protection system. Summary of the Invention

[0004] The purpose of this invention is to provide a data analysis-based system and method for assessing the condition of relay protection equipment, in order to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a data analysis-based method for assessing the condition of relay protection equipment, the method comprising: Step S1: Collect several current operating data of the target device and image data of the current cabinet pressure plate. Calculate the operating deviation value avg of the target device based on the error between the operating status data and the corresponding standard value. The anomaly recognition model identifies the abnormal area of ​​the image data to determine the image deviation value. Calculate the current abnormal state evaluation value R of the target device based on the operating deviation value and the image deviation value. Step S2: Extract the equipment information from the historical inspection records of the relay protection equipment in the substation and generate a set of defective equipment. Compare the equipment information of the target equipment with the set of defective equipment. Calculate the current associated defect intensity G of the target equipment based on the similarity of the most similar relay protection equipment. Step S3: Weighted summation of the abnormal state evaluation value R and the associated defect intensity G to obtain the overall evaluation value H of the target equipment; Step S4: Compare the current overall assessment value H with the alarm threshold W to determine whether the target device should be given a fault status warning; predict the overall assessment value of the target device in the future time period based on the historical overall assessment value H sequence. Based on the overall assessment prediction value The alarm threshold W is compared to determine whether the target device will receive a fault status warning in the future.

[0006] Furthermore, the calculation of the operating deviation value avg in step S1 includes: calculating the deviation rate of all operating parameters and taking the average to generate the operating deviation value avg. The formula for calculating the deviation rate of the i-th operating parameter is: , in, Vr represents the deviation rate of the i-th operating parameter. i Vs represents the current value of the i-th operating parameter of the target device. i This represents the standard value of the i-th running parameter; The image deviation value includes the area ratio c of the abnormal region in the image data and the visual difference value f, which indicates the deviation of pixel attributes before and after the abnormal region has an abnormality. The formula for the abnormal state evaluation value R is: R = avg × (1 + γ × f × c) Where γ represents the magnification factor.

[0007] Furthermore, the formula for calculating the associated defect strength G in step S2 is as follows: , in, This indicates the similarity between the target device and the j-th relay protection device. This indicates the service life of the j-th relay protection device. The number of faults of the j-th relay protection device within a unit time period is represented by N, and N represents the number of the most similar relay protection devices.

[0008] Furthermore, the formula for calculating the overall evaluation value H in step S3 is as follows: H = α × R + β × G Here, α and β both represent weights, and α+β=1.

[0009] Furthermore, in step S4, the overall evaluation prediction value The calculation formula is: , in, This represents the overall assessment prediction value corresponding to the next k time steps. This represents the overall evaluation value corresponding to the current time step within the historical overall evaluation value H sequence. hL represents the number of future time steps, and hL represents the overall evaluation value corresponding to the Lth time step in the past within the historical overall evaluation value H sequence.

[0010] A data analysis-based relay protection equipment condition assessment system includes: a data acquisition and preprocessing module, a real-time condition assessment module, a related defect analysis module, and an early warning and decision-making module; The data acquisition and preprocessing module is used to acquire historical inspection records of relay protection equipment in substations, extract semantic features related to equipment information to establish a set of defective equipment, and establish an anomaly recognition model to identify abnormal areas based on abnormal images of the switchgear pressure plates of relay protection equipment. The real-time status assessment module is used to collect the current operating data of the target equipment, calculate the error between the operating data and the standard value to obtain the operating deviation value avg of the target equipment, and combine the anomaly recognition model to identify the anomaly of the image of the screen cabinet pressure plate of the target equipment, and perform a combined assessment of the current abnormal status of the target equipment to obtain the abnormal status assessment value R. The associated defect analysis module is used to collect the semantic features of the target device's equipment information and use them as target semantic features. It compares the target semantic features with the set of defective devices, calculates the associated defect strength G between the target device and the defective devices, and performs a weighted operation with the current abnormal state evaluation value R to obtain the overall evaluation value H of the target device. The early warning and decision-making module is used to assess the operating status of relay protection equipment based on the overall assessment value, predict the future overall assessment value of the target equipment based on the sequence of overall assessment values, and trigger corresponding early warning and operation and maintenance decision-making suggestions based on the overall assessment value.

[0011] Furthermore, the data acquisition and preprocessing module includes a historical data management unit and an image recognition unit. The historical data management unit is used to standardize the collected semantic features. The image recognition unit is used to identify abnormal areas in the cabinet pressure plate image by training an anomaly recognition model based on machine vision algorithms.

[0012] Furthermore, the real-time status assessment module includes: a parameter calibration unit, a visual difference management unit, and an abnormal status assessment unit; the parameter calibration unit is used to periodically update the standard values ​​of each operating parameter and calculate the deviation rate of each operating parameter; the visual difference management unit is used to compare the area ratio of abnormal area images to the current cabinet pressure plate image data and the difference value of visual differences between abnormal area images and normal state images; the abnormal status assessment unit is used to calculate the abnormal status assessment value R of the target equipment.

[0013] Furthermore, the associated defect analysis module includes a semantic comparison unit, an associated defect intensity management unit, and an overall evaluation value management unit. The semantic comparison unit is used to compare the semantic features of the target device with the semantic features of the relay protection devices in the defective device set, and calculate the similarity of the semantic features of each piece of device information. The associated defect intensity management unit is used to obtain the equipment operating years and the number of faults per unit time period of the relay protection devices in the defective device set, and calculate the current associated defect intensity G of the target device. The overall evaluation value management unit is used to obtain the abnormal state evaluation value R and the associated defect intensity G of the target device, and perform a weighted calculation to obtain the overall evaluation value H of the target device. Furthermore, the early warning and decision-making module includes: an alarm threshold management unit, an evaluation value prediction unit, and an information prompting unit. The alarm threshold management unit manages the alarm thresholds W of the relay protection equipment; the evaluation value prediction unit obtains the overall evaluation value H of the target equipment before a certain time step, sets the number of prediction time steps k, and calculates the predicted value of the overall evaluation value of the target equipment. The information prompting unit is used to provide alarm information according to the alarm policy.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. By integrating real-time operational data, visual images, historical static attributes (device information of the target device), and fault records, a multi-dimensional evaluation system is constructed. Compared to traditional methods that rely on only a single data source, this system can more comprehensively depict the device's status.

[0015] 2. A correlation defect intensity index was constructed. By deeply comparing the semantic features corresponding to the target equipment's information with a set of defective equipment with historical faults, the inherent defect risk of inheriting specific models, batches, or manufacturers can be quantified. This enables the assessment work to go beyond single equipment and gain the ability to analyze "family defects." By introducing a similarity calculation mechanism based on equipment correlation features into family defect risk assessment, it is possible to prevent and proactively repair target equipment prone to failure in batches.

[0016] 3. By setting an overall evaluation value H obtained by weighted summation of associated defect intensity G and abnormal state evaluation value R, the relay protection equipment is made transparent, interpretable and highly practical in state evaluation, effectively improving the intelligent level of power system operation and maintenance. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the data analysis-based relay protection equipment condition assessment system of the present invention; Figure 2 This is a flowchart illustrating the data analysis-based condition assessment method for relay protection equipment according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example: Figures 1-2 As shown, the present invention provides a technical solution: a relay protection equipment condition assessment system and method based on data analysis.

[0020] The methods include: Step S1: Collect several current operating data of the target device and image data of the current cabinet pressure plate. Calculate the operating deviation value avg of the target device based on the error between the operating status data and the corresponding standard value. The anomaly recognition model identifies the abnormal area of ​​the image data to determine the image deviation value. Calculate the current abnormal state evaluation value R of the target device based on the operating deviation value and the image deviation value. In step S1, calculating the operating deviation value avg includes: calculating the deviation rate of all operating parameters and taking the average to generate the operating deviation value avg. The formula for calculating the deviation rate of the i-th operating parameter is as follows: , in, Vr represents the deviation rate of the i-th operating parameter. i Vs represents the current value of the i-th operating parameter of the target device. i This represents the standard value of the i-th running parameter; Image deviation values ​​include the area ratio (c) of the abnormal region in the image data and the visual difference value (f) indicating the deviation of pixel attributes before and after the presence or absence of an abnormal region. The formula for the abnormal state evaluation value R is: R = avg × (1 + γ × f × c) Where γ represents the magnification factor.

[0021] In step S1, an anomaly assessment is performed on the current state of the target device based on multi-dimensional runtime data. Simultaneously, anomaly regions in the target device's current image data are identified using an anomaly recognition model, and pixel-level analysis is conducted to determine the image deviation values ​​of these anomaly regions.

[0022] Gram-level analysis includes collecting the area ratio c of abnormal regions in the overall image data; gram-level analysis also includes encoding colors (digital encoding) in grayscale space, RGB three primary color space (red (R), green (G), blue (B)) or HSV color space (hue (H), saturation (S), lightness (V)) on a pixel-by-pixel basis.

[0023] For example, different grayscale ranges correspond to a numerical code, and different red proportion ranges correspond to a numerical code. The numerical code is greater than 0 and increases linearly with the grayscale or color intensity. The difference between the codes corresponding to all abnormal pixels within the abnormal region and the codes corresponding to all theoretical pixels in the normal region is calculated, and the average of all differences constitutes the visual difference value f.

[0024] Furthermore, the construction of the anomaly recognition model includes: collecting several historical images of cabinet pressure plates, constructing and training a visual anomaly recognition model for online monitoring of cabinet pressure plate status, and training the labeled image anomalies using a convolutional neural network or U-net model so that the model can effectively identify the abnormal areas of the cabinet pressure plate images and label the corresponding abnormal areas with significant, quantified image deviation values ​​(area proportion c, visual difference value f).

[0025] Furthermore, the abnormal state assessment value R integrates the analysis of abnormal conditions in the relay protection equipment's operating parameters and the cabinet pressure plate images. Visual anomalies in the cabinet pressure plates typically include unclear markings, uncleaned stains, and corrosion. The formula for calculating the abnormal state assessment value reflects a risk superposition effect: when the equipment cabinet appears completely normal, the value of f is 0, and R is entirely determined by the average deviation of the operating parameters. However, when a visual anomaly is detected in the cabinet pressure plate image (current image data), the value of f will be greater than 0. It acts as an amplifying factor on the parameter deviation, meaning that even if the parameter deviation itself is small, a significant visual anomaly, such as pressure plate burning, will cause the R value to rise sharply. This more accurately reflects the severity of multiple concurrent risks, facilitating accurate subsequent maintenance decision-making recommendations.

[0026] Step S2: Extract the equipment information from the historical inspection records of the relay protection equipment in the substation and generate a set of defective equipment. Compare the equipment information of the target equipment with the set of defective equipment. Calculate the current associated defect intensity G of the target equipment based on the similarity of the most similar relay protection equipment. In step S2, the equipment information includes the equipment model, equipment code, manufacturer, and equipment version information. Since the equipment code contains manufacturing information, a unique serial number, and a check digit, and because the manufacturing information in the equipment codes of equipment in the same batch is exactly the same, the equipment code only contains manufacturing information.

[0027] The formula for calculating the associated defect strength G in step S2 is: , in, This indicates the similarity between the target device and the j-th relay protection device. This indicates the service life of the j-th relay protection device. The number of faults of the j-th relay protection device within a unit time period is represented by N, and N represents the number of the most similar relay protection devices.

[0028] In step S2, the defective equipment set is a structured collection of historical semantics and fault data during the maintenance of relay protection equipment. The aim is to establish a correlation between the characteristic information of relay protection equipment in the same batch and the fault, so as to carry out unified maintenance of target equipment in the same batch in the future.

[0029] For each historical record related to relay protection equipment, the system automatically extracts and structures a series of key static semantic features (equipment information). During the collection of static semantics, the static features can be encoded, such as pre-encoding equipment manufacturers, equipment models, and equipment versions into word vectors, or using word vector models such as Word2Vec and BERT word embedding models to calculate the cosine similarity between the word vector of the target equipment and the corresponding word vectors of other relay protection equipment, in order to determine the similarity between the equipment.

[0030] In step S2, the associated defect strength G aims to quantify the degree of association between the target device and a group of historically frequently failing "problem devices." In the formula for calculating the associated defect strength G, sim... j The similarity is the one calculated above; d j It is the average number of failures of the j-th most similar device in the defective equipment set within a unit time period. This value is directly obtained from historical maintenance records; y j It is the service life of the j-th equipment, where max(y) is the value of the j-th equipment. jThe design of formula 1) aims to avoid an excessively small denominator due to equipment with less than one year of operation, which could lead to an unreasonable amplification effect. The underlying logic of this formula is that if the target equipment is highly similar in semantic features to several historically faulty devices, i.e., sim... j The value is close to 1, and those devices have a history of failures and the devices themselves fail very frequently, as shown by d. j High value, especially frequent failures in its early stages of operation (y j The value is small, resulting in 1 / y j The larger the G-value, the greater the contribution of these historically faulty devices to the target device's G-value. By summing the contributions of all historically faulty devices, the G-value becomes a quantitative indicator of the potential "family history" risk of the target device. For example, if a certain model, batch, and version of equipment frequently fails, it indicates that devices with similar semantic characteristics may all carry a "family defect" risk. Therefore, devices with the same or highly similar semantic characteristics have an extremely high risk of failing.

[0031] Step S3: Weighted summation of the abnormal state evaluation value R and the associated defect intensity G to obtain the overall evaluation value H of the target equipment; The formula for calculating the overall evaluation value H in step S3 is as follows: H = α × R + β × G Here, α and β both represent weights, and α+β=1.

[0032] In step S3, the overall evaluation value H combines the equipment's "family defects" (related defect strength G) and current operational defects (abnormal state evaluation value R), integrating real-time operational data, visual images, historical static attributes (equipment information), and fault records to construct a multi-dimensional evaluation system. Compared to traditional methods that rely on only a single data source, this approach can more comprehensively depict the equipment's status and provide fault status early warnings, enabling more accurate maintenance decisions for each piece of equipment based on these warnings.

[0033] Step S4: Compare the current overall assessment value H with the alarm threshold W to determine whether the target device should be given a fault status warning; predict the overall assessment value of the target device in the future time period based on the historical overall assessment value H sequence. Based on the overall assessment prediction value The alarm threshold W is compared to determine whether the target device will receive a fault status warning in the future.

[0034] Among them, the overall evaluation prediction value in step S4 The calculation formula is: , in, This represents the overall assessment prediction value corresponding to the next k time steps. This represents the overall evaluation value corresponding to the current time step within the historical overall evaluation value H sequence. Let k represent the number of future time steps, and hL represent the overall evaluation value corresponding to the Lth time step in the historical overall evaluation value sequence H. Also, k must be less than L.

[0035] In step S4, a fault status warning is determined based on the comparison with the overall evaluation value H at the current moment. The core purpose of this step is to determine whether to issue a fault status warning based on a multi-dimensional evaluation system of the target equipment from multiple perspectives. Specifically, for target equipment with abnormal operation or family-related defects, the evaluation method of this application will output a fault status warning, thereby increasing the maintenance frequency of the aforementioned target equipment.

[0036] Furthermore, using the recent historical overall assessment value H series, the overall assessment value of the target equipment in the future period is predicted. Overall assessment and prediction values This represents the maximum overall assessment value H that the target device may reach within a preset future time period. Based on this overall assessment value, a predicted value is generated. The alarm status is compared with the set alarm threshold W, and a fault status warning for the target device is given in the future period, so as to provide suggestions for multiple maintenance decisions in the future preset period based on the fault status warning. Specific implementation examples: Construct a set of defective devices; Obtain information on all relay protection devices with historical maintenance records in the substation, extract their semantic features, and construct a set of defective devices. , of which Taiwanese equipment The semantic feature vector is: , Examples of dimension definitions are as follows: (Equipment model can use alphanumeric codes) Manufacturers may use the last 6 digits of the Unified Social Credit Code; Hardware version number, Software version number Equipment service life Number of failures in the last 12 months.

[0038] Calculate the current abnormal state assessment value R of the target equipment; Collect the following five operating parameters: operating voltage, return coefficient, operating time, insulation resistance, and contact resistance. The standard values ​​were 220V, 0.95, 30ms, 100MΩ and 1mΩ, respectively. The current values ​​of the target sampling device were 222V, 0.92, 28.5ms, 98MΩ and 1.25mΩ, respectively. The deviation rates were 0.0091, 0.0316, 0.05, 0.02 and 0.25, respectively. The abnormal state evaluation value R was 0.0721. The abnormal image data is identified and abnormal areas are determined based on the anomaly detection model. The proportion of abnormal image area to image area is 3%. In this example, the weighted value of color difference and grayscale difference is used as the visual difference value f, where color difference fe=0.18, grayscale difference fg=0.22, f=0.6×0.18+0.4×0.22=0.196, f×c=0.00588, and the magnification factor γ=10. R=0.0721×(1+0.0588)=0.07634.

[0039] Step S3: Calculate the associated defect strength G and the overall evaluation value H; Based on semantic feature matching, a device with a semantic similarity of 0.9 was matched. This device has been in operation for 11 years and has experienced 2 failures. Therefore, N=1. =0.9, =11, =2, the calculated associated defect strength G=0.9×0.18=0.1636; With α and β both set to 0.5, the calculated overall evaluation value H is: H=0.5×0.07634+0.5×0.1636=0.119; Step S4: Status assessment, early warning, and trend prediction; Set the alarm threshold W=0.15, If we obtain the overall evaluation values ​​[0.08, 0.09, 0.1, 0.105, 0.112, 0.117, 0.119] corresponding to the past 7 time units, then: =0.08, G=0.119, k=3, L=7.

[0040] Calculate the predicted value after 3 units of time: =0.119 + 3 × 0.0057 = 0.1361; If the overall evaluation value of the target device does not exceed the alarm threshold after 3 unit time periods, the target device will not be listed as a key monitoring device during the detection process of the third unit time period.

[0041] A data analysis-based relay protection equipment condition assessment system includes: a data acquisition and preprocessing module, a real-time condition assessment module, a related defect analysis module, and an early warning and decision-making module; The data acquisition and preprocessing module extracts semantic features related to equipment information to establish a set of defective equipment, and establishes an anomaly recognition module to identify abnormal areas based on the abnormal images of the pressure plates of the relay protection equipment's cabinets. The data acquisition and preprocessing module includes a historical data management unit and an image recognition unit. The historical data management unit is used to standardize the collected semantic features. The image recognition unit is used to identify abnormal areas in the image of the cabinet pressure plate by training an anomaly recognition model based on machine vision algorithms. The real-time status assessment module collects the current operating data of the target equipment, calculates the error between the operating data and the standard value to obtain the operating deviation value avg of the target equipment, and combines the anomaly recognition model to identify the anomalies in the image of the target equipment's cabinet pressure plate, performing a combined assessment of the current abnormal state of the target equipment to obtain the abnormal state assessment value R. The real-time status assessment module includes: a parameter calibration unit, a visual difference management unit, and an abnormal state assessment unit. The parameter calibration unit periodically updates the standard values ​​of each operating parameter and calculates the deviation rate of each operating parameter; the visual difference management unit compares the area ratio of the abnormal area image to the current cabinet pressure plate image data and the visual difference value between the abnormal area image and the normal state image; the abnormal state assessment unit calculates the abnormal state assessment value R of the target equipment. The associated defect analysis module is used to collect the semantic features of the target device's equipment information and use them as target semantic features. It compares the target semantic features with the set of defective devices, calculates the associated defect strength G between the target device and the defective devices, and performs a weighted calculation with the current abnormal state evaluation value R to obtain the overall evaluation value H of the target device. The associated defect analysis module includes a semantic comparison unit, an associated defect strength management unit, and an overall evaluation value management unit. The semantic comparison unit compares the semantic features of the target device with the semantic features of the relay protection devices in the set of defective devices and calculates the similarity of the semantic features of each piece of equipment information. The associated defect strength management unit obtains the equipment operating years and the number of faults per unit time period of the relay protection devices in the set of defective devices and calculates the current associated defect strength G of the target device. The overall evaluation value management unit obtains the abnormal state evaluation value R and the associated defect strength G of the target device and performs a weighted calculation to obtain the overall evaluation value H of the target device. The early warning and decision-making module is used to assess the operating status of relay protection equipment based on the overall evaluation value, predict the future overall evaluation value of the target equipment based on the sequence of overall evaluation values, and trigger corresponding early warning and operation and maintenance decision suggestions based on the overall evaluation value. The early warning and decision-making module includes: an alarm threshold management unit, an evaluation value prediction unit, and an information prompting unit. The alarm threshold management unit manages the alarm threshold W of the relay protection equipment. The evaluation value prediction unit obtains the overall evaluation value H of the target equipment before a certain time step, sets the number of prediction time steps k, and calculates the predicted overall evaluation value of the target equipment. The information prompting unit is used to provide alarm information according to the alarm policy.

[0042] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A data analysis-based method for assessing the condition of relay protection equipment, characterized in that: The method includes the following steps: Step S1: Collect several current operating data of the target device and image data of the current cabinet pressure plate. Calculate the operating deviation value avg of the target device based on the error between the operating status data and the corresponding standard value. The anomaly recognition model identifies the abnormal area of ​​the image data to determine the image deviation value. Calculate the current abnormal state evaluation value R of the target device based on the operating deviation value and the image deviation value. Step S2: Extract the equipment information from the historical inspection records of the relay protection equipment in the substation and generate a set of defective equipment. Compare the equipment information of the target equipment with the set of defective equipment. Calculate the current associated defect intensity G of the target equipment based on the similarity of the most similar relay protection equipment. Step S3: Weighted summation of the abnormal state evaluation value R and the associated defect intensity G to obtain the overall evaluation value H of the target equipment; Step S4: Compare the current overall evaluation value H with the alarm threshold W to determine whether the target device should be given a fault status warning; Based on historical overall assessment values ​​(H-series), predict the overall assessment value of the target equipment for future periods. Based on the overall assessment prediction value The alarm threshold W is compared to determine whether the target device will receive a fault status warning in the future.

2. The data analysis-based condition assessment method for relay protection equipment according to claim 1, characterized in that, The step S1 of calculating the operating deviation value avg includes: calculating the deviation rate of all operating parameters and taking the average to generate the operating deviation value avg. The formula for calculating the deviation rate of the i-th operating parameter is: ， in, Vr represents the deviation rate of the i-th operating parameter. i Vs represents the current value of the i-th operating parameter of the target device. i This represents the standard value of the i-th running parameter; The image deviation value includes the area ratio c of the abnormal region in the image data and the visual difference value f, which indicates the deviation of pixel attributes before and after the abnormal region has an abnormality. The formula for the abnormal state evaluation value R is: R = avg × (1 + γ × f × c) Where γ represents the magnification factor.

3. The data analysis-based condition assessment method for relay protection equipment according to claim 1, characterized in that: The formula for calculating the associated defect strength G in step S2 is as follows: ， in, This indicates the similarity between the target device and the j-th relay protection device. This indicates the service life of the j-th relay protection device. The number of faults of the j-th relay protection device within a unit time period is represented by N, and N represents the number of the most similar relay protection devices.

4. The data analysis-based condition assessment method for relay protection equipment according to claim 1, characterized in that: The formula for calculating the overall evaluation value H in step S3 is as follows: H = α × R + β × G Here, α and β both represent weights, and α+β=1.

5. The data analysis-based condition assessment method for relay protection equipment according to claim 1, characterized in that: The overall evaluation prediction value in step S4 The calculation formula is: ， in, This represents the overall assessment prediction value corresponding to the next k time steps. This represents the overall evaluation value corresponding to the current time step within the historical overall evaluation value H sequence. hL represents the number of future time steps, and hL represents the overall evaluation value corresponding to the Lth time step in the past within the historical overall evaluation value H sequence.

6. A data analysis-based relay protection equipment condition assessment system, used to execute the data analysis-based relay protection equipment condition assessment method according to any one of claims 1-5, characterized in that: The system includes: The module includes data acquisition and preprocessing, real-time status assessment, correlation defect analysis, and early warning and decision-making. The data acquisition and preprocessing module is used to acquire historical inspection records of relay protection equipment in substations, extract semantic features related to equipment information to establish a set of defective equipment, and establish an anomaly recognition model to identify abnormal areas based on abnormal images of the switchgear pressure plates of relay protection equipment. The real-time status assessment module is used to collect the current operating data of the target equipment, calculate the error between the operating data and the standard value to obtain the operating deviation value avg of the target equipment, and combine the anomaly recognition model to identify the anomaly of the image of the screen cabinet pressure plate of the target equipment, and perform a combined assessment of the current abnormal status of the target equipment to obtain the abnormal status assessment value R. The associated defect analysis module is used to collect the semantic features of the target device's equipment information and use them as target semantic features. It compares the target semantic features with the set of defective devices, calculates the associated defect strength G between the target device and the defective devices, and performs a weighted operation with the current abnormal state evaluation value R to obtain the overall evaluation value H of the target device. The early warning and decision-making module is used to assess the operating status of relay protection equipment based on the overall assessment value, predict the future overall assessment value of the target equipment based on the sequence of overall assessment values, and trigger corresponding early warning and operation and maintenance decision-making suggestions based on the overall assessment value.

7. The relay protection equipment condition assessment system based on data analysis according to claim 6, characterized in that: The data acquisition and preprocessing module includes: a historical data management unit and an image recognition unit; The historical data management unit is used to standardize the collected semantic features; The image recognition unit is used to identify abnormal areas in the image of the cabinet pressure plate by training an anomaly recognition model based on machine vision algorithms.

8. The relay protection equipment condition assessment system based on data analysis according to claim 6, characterized in that: The real-time status assessment module includes: a parameter calibration unit, a visual difference management unit, and an abnormal status assessment unit; The parameter calibration unit is used to periodically update the standard values ​​of each operating parameter and calculate the deviation rate of each operating parameter; The visual difference management unit is used to compare the area ratio of abnormal area images to the current cabinet pressure plate image data, and the difference value of visual difference between abnormal area images and normal state images. The abnormal state assessment unit is used to calculate the abnormal state assessment value R of the target device.

9. The relay protection equipment condition assessment system based on data analysis according to claim 6, characterized in that: The associated defect analysis module includes: a semantic comparison unit, an associated defect strength management unit, and an overall evaluation value management unit; The semantic comparison unit is used to compare the semantic features of the target device with the semantic features of the relay protection devices in the defective device set, and calculate the similarity of the semantic features of each piece of device information. The associated defect intensity management unit is used to obtain the equipment operating years and the number of faults in a unit time period of the relay protection equipment in the defective equipment set, and to calculate the current associated defect intensity G of the target equipment. The overall evaluation value management unit is used to obtain the abnormal state evaluation value R and the associated defect intensity G of the target equipment, and then perform a weighted calculation to obtain the overall evaluation value H of the target equipment.

10. The relay protection equipment condition assessment system based on data analysis according to claim 6, characterized in that: The early warning and decision-making module includes: an alarm threshold management unit, an evaluation value prediction unit, and an information prompt unit; The alarm threshold management unit is used to manage the alarm thresholds W of relay protection equipment; The evaluation value prediction unit is used to obtain the overall evaluation value H of the target device before the time step, set the number of prediction time steps k, and calculate the predicted value of the overall evaluation value of the target device. ; The information prompting unit is used to provide alarm information according to the alarm policy.

Images