Power grid multi-level evaluation method and system based on power grid device traceability record items

Abstract

Provided are a power grid multi-level evaluation method and system based on power grid device traceability record items. The system comprises: a power grid device record item collection and processing module, a historical-record-item analysis and pattern identification module, a handling pattern rating and optimization module, and a functional-fault pre-determination and decision support module. By means of a functional-fault handling record screening mechanism, a handling pattern rating mechanism and an ineffective-measure deletion mechanism, the quality of functional-fault handling measures and the reliability of a functional-fault pattern library are improved, and ineffective or inefficient handling measures are effectively reduced; and by means of establishing and applying a functional-fault pre-determination model, the accuracy and timeliness of functional-fault pre-determination for a power grid device are improved.

Description

A method and system for multi-level evaluation of power grids based on power grid equipment traceability records. Technical Field

[0001] This invention belongs to the field of power system automation, and in particular relates to a method and system for multi-level evaluation of power grids based on power grid equipment traceability record items. Background Technology

[0002] The field of power system automation involves using advanced information technology, control theory, and computer technology to monitor, control, optimize, and protect power systems. This includes real-time monitoring of grid conditions, automatic adjustment of generation and load, and prediction and prevention of functional failures to ensure the stable, efficient, and safe operation of the power system.

[0003] Multilevel evaluation of power grids is an important component of power system automation. It provides a comprehensive understanding of the power grid's operating status and performance through multi-level assessments.

[0004] Fault handling measures can vary in quality, with some measures being ineffective or even useless. This not only wastes resources but may also delay fault repair time, increasing the risk to grid operations. Due to the lack of effective screening and rating mechanisms, some ineffective or inefficient fault handling measures may be widely applied, which not only increases operating costs but may also negatively impact the stable operation of the grid. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-level evaluation method for power grids based on power grid equipment traceability records, aiming to solve the technical problems existing in the prior art as identified in the background art.

[0006] This invention is implemented as follows: a multi-level evaluation method for power grids based on power grid equipment traceability records, the method comprising:

[0007] Collect real-time and previous records of power grid equipment operation from the power grid equipment, including operating parameters, status monitoring records, environmental parameters and previous maintenance records, and preprocess the acquired records.

[0008] Based on previous records, analyze and identify patterns and anomalies in the previous records, analyze potential functional failure patterns and the probability of functional failures in power grid equipment, and establish a functional failure pattern library. Combined with previous records, content related to the same type of functional failure is used as storage tags and stored in the functional failure pattern library.

[0009] Rate the treatment patterns for functional impairments in the previous records, set level thresholds, and delete treatment patterns that are below the level thresholds.

[0010] Establish a functional failure prediction model, combine real-time records of power grid equipment to predict potential functional failure patterns and timing points, and simultaneously search the functional failure pattern library based on the type of functional failure to obtain the handling patterns for that type of functional failure from previous records, and plan the functional failure handling decisions according to the handling time nodes.

[0011] As a further aspect of the present invention, the analysis and identification of patterns and anomalies in previously recorded items, the analysis of potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and the establishment of a functional failure pattern library specifically include:

[0012] Establish a functional malfunction identification model, analyze the previously collected records of power grid equipment operation, identify the normal operation mode of the equipment, detect the record points that deviate from the normal mode, and mark the record points.

[0013] A secondary analysis is performed on the marked record items, and the development trend of potential functional impairment is judged in combination with the subsequent record items of the record item. The probability and pattern of functional impairment caused by the record item are also judged.

[0014] Establish a functional failure pattern library, store the functional failure-related content in the previous records of power grid equipment operation in the functional failure pattern library, and use the functional failure type as the storage label to store all previous records related to the functional failure type under the storage label of that type.

[0015] As a further aspect of the present invention, the step of rating the processing patterns for functional impairments in previously recorded items, setting a level threshold, and deleting processing patterns below the level threshold specifically includes:

[0016] Filter out all records related to the management of functional impairment from the previous records, including the time of occurrence of functional impairment and the type of functional impairment, and extract the management measures, management time and management results of the type of functional impairment from the functional impairment pattern library according to the functional impairment type;

[0017] Establish rating rules, including indicators such as processing time, processing cost, functional impairment recovery time, and the time of impact of highly impaired functional impairment, and conduct level assessments of processing measures and results;

[0018] According to the rating rules, a level threshold is set, and functional impairment treatment measures that are rated below the level threshold are defined as invalid measures and removed from the functional impairment pattern library.

[0019] As a further aspect of the present invention, the establishment of a functional failure prediction model, which combines real-time records of power grid equipment to predict potential functional failure patterns and timing points, specifically includes:

[0020] Establish a functional impairment prediction model and train the functional impairment prediction model using previously recorded items stored in the functional impairment pattern library;

[0021] Substitute the real-time records of power grid equipment operation into the functional failure prediction model to predict the possible functional failure modes and timing of power grid equipment.

[0022] Based on the predicted type of functional impairment, the processing mode for that type of functional impairment is retrieved from the previous records in the functional impairment pattern library. Based on the predicted time of functional impairment, the time window from the current time to the possible occurrence of the functional impairment is determined. Combined with the processing mode, the processing time node and functional impairment processing steps are set for the predicted functional impairment.

[0023] Another objective of this invention is to provide a multi-level power grid evaluation system based on power grid equipment traceability records, the system comprising:

[0024] The power grid equipment record item acquisition and processing module is used to collect real-time and previous records of power grid equipment operation from the power grid equipment, including operating parameters, status monitoring records, environmental parameters and previous maintenance records, and to preprocess the acquired records.

[0025] The previous record item analysis and pattern recognition module is used to analyze and identify patterns and anomalies in previous records based on previous records, analyze potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and establish a functional failure pattern library. Combined with previous records, content related to the same type of functional failure is used as storage tags and stored in the functional failure pattern library.

[0026] The processing mode rating and optimization module is used to rate the processing modes of functional impairment in the previous records, set level thresholds, and delete processing modes that are below the level thresholds.

[0027] The Fault Prediction and Decision Support Module is used to establish a fault prediction model. By combining real-time records of power grid equipment, it predicts the potential fault patterns and timing of equipment failures. At the same time, it searches the fault pattern library based on the type of fault to obtain the handling mode for that type of fault from previous records, and plans the fault handling decision according to the handling time node.

[0028] As a further embodiment of the present invention, the previous record item analysis and pattern recognition module includes:

[0029] The record item pattern recognition unit is used to establish a functional failure recognition model, analyze the previously collected records of the operation of the power grid equipment, identify the normal operation mode of the equipment, detect record item points that deviate from the normal mode, and mark the record item points.

[0030] The functional impairment probability detection and assessment unit performs secondary analysis on the marked record items, combines the subsequent record items of the record item to judge the development trend of potential functional impairment, and judges the probability and pattern of functional impairment caused by the record item.

[0031] The Fault Pattern Library Management Unit establishes a fault pattern library, stores fault-related content from previous records of power grid equipment operation in the fault pattern library, and stores all previous records related to the fault type under the corresponding storage tag.

[0032] As a further embodiment of the present invention, the processing mode rating and optimization module includes:

[0033] The dysfunction management record item filtering unit is used to filter out all records related to dysfunction management from previous record items, including the time of occurrence of dysfunction and the type of dysfunction, and extract the management measures, management time and management results of the type of dysfunction from the dysfunction pattern library according to the dysfunction type;

[0034] The treatment measure rating unit is used to formulate rating rules, including indicators such as treatment time, treatment cost, functional impairment recovery time, and the time of impact of highly impaired functional impairment, and to evaluate the treatment measures and treatment results.

[0035] According to the rating rules, a level threshold is set, and functional impairment treatment measures that are rated below the level threshold are defined as invalid measures and removed from the functional impairment pattern library.

[0036] As a further embodiment of the present invention, the functional impairment prediction and decision support module includes:

[0037] The functional impairment prediction model building training unit is used to build a functional impairment prediction model and train the functional impairment prediction model using previous records stored in the functional impairment pattern library.

[0038] The functional failure mode time analysis and prediction unit is used to input the real-time records of the operation of power grid equipment into the functional failure prediction model to predict the possible functional failure modes and time points of the power grid equipment.

[0039] The functional impairment processing time step planning unit is used to retrieve the processing mode for the type of functional impairment from the previous record in the functional impairment mode library according to the predicted functional impairment type, and determine the time window from the current time to the possible occurrence of the functional impairment based on the predicted functional impairment time. In combination with the processing mode, it sets the processing time node and functional impairment processing steps for the predicted functional impairment.

[0040] The beneficial effects of this invention are:

[0041] Through meticulous screening of dysfunction management records, comprehensive rating rule development, and a rigorous mechanism for removing ineffective measures, the quality of dysfunction management actions and the reliability of the dysfunction pattern library have been significantly improved. This not only helps optimize future dysfunction management strategies but also effectively reduces ineffective or inefficient management actions.

[0042] By establishing and applying a functional failure prediction model, the accuracy and timeliness of predicting power grid equipment functional failures have been improved. Furthermore, by planning and implementing countermeasures in advance, potential risks to power grid operation have been effectively reduced. This proactive approach to functional failure management significantly enhances the reliability and stability of the power grid, providing robust decision support for power grid operators. Attached Figure Description

[0043] Figure 1 is a flowchart of a multi-level power grid evaluation method based on power grid equipment traceability record items provided by an embodiment of the present invention;

[0044] Figure 2 is a flowchart of the process provided by the embodiment of the present invention to analyze and identify patterns and anomalies in previously recorded items, analyze potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and establish a functional failure pattern library.

[0045] Figure 3 is a flowchart of rating the processing mode of functional impairment in the previous recorded items and setting the level threshold according to an embodiment of the present invention;

[0046] Figure 4 is a flowchart of the process of establishing a functional failure prediction model and predicting the potential functional failure mode and time point of the equipment by combining the real-time records of the power grid equipment, according to an embodiment of the present invention.

[0047] Figure 5 is a structural block diagram of a multi-level power grid evaluation system based on power grid equipment traceability record items provided in an embodiment of the present invention;

[0048] Figure 6 is a structural block diagram of the previous record item analysis and pattern recognition module provided in an embodiment of the present invention;

[0049] Figure 7 is a structural block diagram of the processing mode rating and optimization module provided in an embodiment of the present invention;

[0050] Figure 8 is a structural block diagram of the functional impairment prediction and decision support module provided in an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0052] It is understood that the terms "first," "second," etc., used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this application, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0053] Figure 1 is a flowchart of a multi-level power grid evaluation method based on power grid equipment traceability record items provided by an embodiment of the present invention. As shown in Figure 1, the multi-level power grid evaluation method based on power grid equipment traceability record items is characterized in that the method includes steps S100, S200, S300, and S400, as follows:

[0054] S100 collects real-time and previous records of power grid equipment operation from the power grid equipment. These records include operating parameters, specifically electrical parameters such as voltage, current, power, and frequency, as well as equipment switching status and operating time; status monitoring records, specifically physical conditions such as equipment temperature, vibration, and noise, and software system operation logs; environmental parameters, specifically temperature, humidity, wind speed, and pollution levels of the environment in which the equipment is located; and maintenance of previous records. The acquired records are preprocessed, including data cleaning, format standardization, missing value handling, and outlier detection and processing, to ensure data quality and consistency.

[0055] The preprocessing section specifically includes:

[0056] 1. Data cleaning: Remove useless or erroneous data to ensure data accuracy and integrity.

[0057] step:

[0058] Remove duplicates: Check and delete duplicate records.

[0059] Error detection: Identify and correct obvious errors, such as negative voltage or current.

[0060] Non-logic value processing: Check whether the data conforms to physical laws, such as the relationship between voltage and current.

[0061] 2. Ensure consistent formatting to guarantee data from different sources for subsequent analysis.

[0062] step:

[0063] Standardize units: For example, standardize all voltage values ​​to volts (V) and current values ​​to amperes (A), etc.

[0064] Standardize date and time formats: unify all timestamps to the same format (such as ISO 8601).

[0065] Data type conversion: Ensure that numeric, character, and other data types are consistent.

[0066] 3. Handling missing values ​​to avoid bias in analysis results due to missing values.

[0067] step:

[0068] Identify missing values: Mark records with all missing values.

[0069] Missing value imputation: Methods such as mean, median, and mode can be used, or techniques such as interpolation and forward imputation can be employed.

[0070] Delete missing records: If the percentage of missing values ​​is too high (e.g., more than 30%), consider deleting the record.

[0071] 4. Outlier detection and handling: Identify and process unreasonable data points to improve the quality of the dataset.

[0072] step:

[0073] Statistical methods: Use Z-Score or IQR methods to identify outliers.

[0074] Visualization methods: Use visualization tools such as box plots and scatter plots to assist in identification.

[0075] Solution:

[0076] Correction: If the outlier is clearly erroneous, it can be corrected to a value within a reasonable range.

[0077] Deletion: If the outlier is unreasonable and cannot be corrected, consider deleting the record directly.

[0078] Labeling: Retain outliers and label them so that their impact can be considered in subsequent analysis.

[0079] The preprocessing standards should meet the following requirements:

[0080] Completeness: Data records should be complete without missing data, or the percentage of missing data should be below a set threshold (e.g., 10%).

[0081] Accuracy: The cleaned data should be free of errors and conform to physical laws.

[0082] Consistency: All data formats should be consistent to ensure consistency in data processing and analysis.

[0083] Reasonableness: Data processed for anomalies should be within a reasonable range and conform to the actual operating conditions of the equipment. S200: Based on preprocessed records, analyze and identify patterns and anomalies in previous records. Patterns include periodic patterns, trend patterns, and seasonal patterns. Analyze potential functional failure patterns and their probabilities for occurrence in power grid equipment. Functional failure patterns include electrical faults, mechanical faults, and software faults. Subsequently, establish a functional failure pattern library. Combining previous records, classify faults based on their nature, cause, and impact. Content related to similar functional failures is stored as tags in the functional failure pattern library.

[0084] More specifically, as shown in Figure 2, step S200 includes the following three steps:

[0085] S210 will first establish a functional impairment recognition model, which is based on a machine learning algorithm, specifically the support vector machine algorithm, to establish the functional impairment recognition model.

[0086] Where s(x, n) is the anomaly score, ranging from 0 to 1. Records with anomaly scores close to 1 are considered anomalous, potentially indicating device malfunction or potential problems; records with scores close to 0 are considered normal. x represents a specific record, which may include device operating parameters, status information, environmental conditions, etc. h(x) reflects the ease with which record x can be isolated. The shorter the path length, the easier it is for record x to be isolated, indicating a possible anomaly or potential malfunction. E(h(x)) represents the average isolation depth of record x. The shorter the average path length, the more likely the record is to be anomalous. n represents the total number of records, and c(n) is a baseline value used to compare and calculate the anomaly score for each record.

[0087] The model aims to identify normal operating modes from previous records of power grid equipment operation, detect and mark records that deviate from these normal modes. These records may be early signs of potential functional failures and are therefore specifically marked.

[0088] S220, performs secondary analysis on the marked record items, specifically:

[0089] 1. Time series data extraction

[0090] Objective: To extract the marked record points and records at several time points before and after them to form time series data.

[0091] step:

[0092] Determine the window period (e.g., 5 minutes before or after, 10 minutes before or after) to retrieve the corresponding record items.

[0093] The recorded items include parameters such as voltage, current, power, frequency, temperature, vibration, and noise.

[0094] 2. Trend Analysis

[0095] Objective: To identify trends in the changes of recorded items over time.

[0096] step:

[0097] Calculate the rate of change: Calculate the rate of change of each parameter within the window period for the extracted time series data.

[0098] Rate of change = [(current value - previous value) / previous value] × 100%

[0099] Create trend charts: Visualize changes in time series data by creating line charts, bar charts, etc., to easily identify upward, downward, or fluctuating trends.

[0100] 3. Pattern Comparison

[0101] Objective: To compare the marked record items with known patterns in the dysfunction pattern library.

[0102] step:

[0103] Feature extraction: Extracting features from labeled time series, such as mean, standard deviation, maximum value, minimum value, peak value, frequency components, etc.

[0104] Similarity calculation: The labeled pattern is compared with known fault patterns in the pattern library using similarity measurement methods (such as cosine similarity, Euclidean distance, dynamic time warping (DTW)).

[0105] 4. Risk Assessment

[0106] Objective: To calculate the probability that a marked record item will lead to a functional impairment and to assess its risk level.

[0107] step:

[0108] Statistical analysis: Based on historical data, logistic regression or Bayesian methods are used to calculate the probability that each labeled record point leads to functional impairment.

[0109] Set risk level standards (e.g., low, medium, high), for example:

[0110] Low risk: Probability < 0.2

[0111] Medium risk: 0.2 ≤ probability < 0.5

[0112] High risk: Probability ≥ 0.5

[0113] Generate a risk assessment report: summarize the recorded items, their risk levels, corresponding failure modes and their probabilities, in order to facilitate subsequent decision-making.

[0114] S220 uses secondary analysis of the time-series data of marked record items and their subsequent records to identify trends over time, such as increases, decreases, or fluctuations. It compares the marked record items with known patterns in a functional impairment pattern library to identify similarities and determine the existence of known failure modes. Statistical methods are used to calculate the probability that a marked record item will cause functional impairment and assess its risk level. The development trend of potential functional impairment is determined by combining the subsequent records of that record item, and the probability and pattern of functional impairment caused by that record item are also determined. Secondary analysis not only focuses on individual record items but also considers changes in subsequent records to determine the development trend of potential functional impairment. This trend analysis allows for a more accurate prediction of the probability of functional impairment and possible functional impairment patterns. This trend-based analysis method significantly improves the accuracy of functional impairment prediction because it considers the dynamic process of functional impairment development.

[0115] S230 Finally, a functional failure pattern library will be established, storing all relevant functional failure information from previous records of power grid equipment operation. Each functional failure type is used as a storage tag, and all previous records related to that type of functional failure are stored under the corresponding tag. The advantage of this method is that it provides a centralized information repository for each functional failure type, facilitating rapid retrieval and analysis, thereby supporting more effective functional failure prediction and handling.

[0116] Overall, by establishing a functional malfunction identification model and a functional malfunction pattern library, not only has the accuracy and efficiency of identifying potential functional malfunctions from previously recorded items been improved, but a solid foundation has also been laid for subsequent functional malfunction prediction and handling. This pattern recognition and storage method based on previously recorded items enables grid operators to better understand and predict equipment functional malfunctions, thereby taking more effective preventive and response measures, significantly improving the operational safety and reliability of the power grid.

[0117] S300: Rate the handling patterns of functional impairment in the previously recorded items, set a level threshold, and delete the handling patterns that are below the level threshold.

[0118] More specifically, as shown in Figure 3, step S300 includes the following three steps:

[0119] S310 begins by filtering through previous records of the power grid equipment to extract all detailed records related to malfunction handling. These records include not only the exact time of the malfunction and the specific type of malfunction, but also the handling measures taken for each type of malfunction, the handling time, and the handling results. The key to this step is ensuring that the analyzed records are comprehensive and specific, providing an accurate basis for subsequent rating and optimization.

[0120] S320 establishes a detailed set of rating rules to classify treatment measures and outcomes. These rules cover multiple key indicators, including treatment time, treatment cost, recovery time, and duration of impact. These indicators allow for a comprehensive evaluation of the effectiveness and efficiency of each treatment measure. The rating process involves quantitative analysis of historical treatment measures to ensure that each measure receives a fair assessment.

[0121] S330, based on the rating rules, sets a level threshold, defines management measures for functional impairments rated below the level threshold as invalid measures, and removes them from the functional impairment pattern library. The significant advantage of this step is that it ensures the records in the functional impairment pattern library remain of high quality and efficiency, thus providing a more reliable and accurate reference for future functional impairment prediction and management decisions.

[0122] For specific rating rules and level thresholds, the following examples are provided:

[0123] Processing time (T):

[0124] Scoring rules: The shorter the processing time, the higher the score.

[0125] Calculation method: For processing modes with a processing time of 1 hour or less, the score is 100 points; for each additional hour, the score decreases by 10 points; for processing modes with a processing time of more than 24 hours, the score is no higher than 20 points.

[0126] Processing cost (C):

[0127] Scoring rules: The lower the processing cost, the higher the score.

[0128] Calculation method: The actual processing cost is compared with the preset standard cost. The processing mode with a cost more than 50% lower than the standard cost gets 100 points. For every 10% increase in cost, the score decreases by 10 points. The processing mode with a cost more than 100% higher than the standard cost gets no more than 20 points.

[0129] Functional impairment recovery effect (R):

[0130] Scoring rules: The better the recovery of functional impairment, the higher the score.

[0131] Calculation method: The processing mode that fully restores the original performance of the equipment receives 100 points; the processing mode that partially restores the performance but suffers from performance loss is scored according to the degree of loss, with a maximum loss of no more than 50 points; the processing mode that fails to restore the performance receives no more than 30 points.

[0132] Long-term effects (L):

[0133] Scoring rules: The less impact on the long-term stable operation of the power grid, the higher the score.

[0134] Calculation method: The treatment mode with no significant negative impact on the long-term operation of the power grid will be scored 100 points; the treatment mode with a slight impact will be scored according to the degree of impact, with a maximum score of 80 points; the treatment mode with a significant negative impact on the long-term operation of the power grid will be scored no higher than 40 points.

[0135] Overall rating calculation: Overall score = (T×0.3) + (C×0.3) + (R×0.2) + (L×0.2)

[0136] The rating of the processing mode is determined by comparing the overall score with the preset level threshold.

[0137] Level threshold setting:

[0138] High level (excellent): Overall score ≥ 85 points

[0139] Intermediate (Good): 70 points ≤ Overall score < 85 points

[0140] Low level (average): 55 points ≤ Overall score < 70 points

[0141] Unsatisfactory (Poor): Overall score < 55 points.

[0142] Step S300 significantly improves the quality of fault handling measures and the reliability of the fault pattern library through meticulous screening of fault handling records, comprehensive rating rule formulation, and a strict mechanism for removing ineffective measures. This not only helps optimize future fault handling strategies but also effectively reduces ineffective or inefficient handling measures, thereby improving the operating efficiency and reliability of power grid equipment.

[0143] S400 establishes a functional failure prediction model, combines real-time records of power grid equipment to predict potential functional failure modes and timing points, and simultaneously searches the functional failure mode library based on the type of functional failure to obtain the processing mode for that type of functional failure from previous records, and plans the functional failure processing decision according to the processing time node.

[0144] More specifically, as shown in Figure 4, step S400 includes the following three steps:

[0145] S410, Establish a functional impairment prediction model and train it using previously recorded items stored in a functional impairment pattern library. The functional impairment prediction model is built based on machine learning or record-mining techniques. The model training process involves using a large number of previously recorded items stored in the functional impairment pattern library, which contain detailed information on various functional impairment patterns and their related treatment measures. Through this training, the model can learn the inherent connections and patterns between different functional impairment patterns, thereby improving the accuracy and reliability of its predictions.

[0146] Specifically, for the functional impairment model, the random forest algorithm is selected. Random forest is an ensemble learning method that can effectively improve the accuracy and robustness of predictions by combining the prediction results of multiple decision trees.

[0147] The specific model building process is as follows:

[0148] I. Data Preparation:

[0149] Collect historical operating data of power grid equipment, including electrical parameters (voltage, current, power, frequency), condition monitoring records (temperature, vibration, noise), environmental parameters (temperature, humidity), and previous fault records.

[0150] Data cleaning is performed to remove missing and outlier values, ensuring data quality.

[0151] II. Feature Engineering:

[0152] Select relevant features such as voltage, current, equipment temperature, vibration, and equipment status.

[0153] The features are normalized to ensure they are within the same range.

[0154] III. Model Training:

[0155] Data preparation:

[0156] Data cleaning: handling missing and outlier values ​​to ensure data integrity and consistency.

[0157] Data partitioning: The cleaned data is divided into training set, validation set, and test set, typically 70%, 15%, and 15%, respectively.

[0158] Feature engineering:

[0159] Feature normalization: Scaling features to the same range to improve the stability and convergence speed of model training.

[0160] Feature selection: Select the most influential features through correlation analysis, feature importance ranking, and other methods.

[0161] Model training:

[0162] Choose a suitable machine learning algorithm and train the model using a training set. Iterate and optimize the model parameters to minimize the prediction error.

[0163] Cross-validation is used to evaluate the model's generalization ability and prevent overfitting.

[0164] Model evaluation:

[0165] The performance of a model is evaluated using a validation set, and commonly used metrics include accuracy, precision, recall, and F1 score.

[0166] If the model performance is poor, you need to go back to the feature engineering stage and adjust the feature selection or model parameters.

[0167] Model optimization:

[0168] Model performance can be further improved through hyperparameter tuning (such as grid search and random search).

[0169] Select the best model and perform a final evaluation on the test set.

[0170] Model Deployment and Monitoring:

[0171] The trained model is deployed to a real-time monitoring system to ensure that it can receive real-time data and predict functional failures.

[0172] Regularly monitor the model's performance to ensure its accuracy and effectiveness, and retrain the model if necessary.

[0173] Prediction of functional impairment:

[0174] In a real-time monitoring environment, the model makes predictions by receiving real-time data (such as current voltage, temperature, vibration, etc.).

[0175] Assuming that the real-time input feature data (such as voltage, temperature, vibration) is very similar to the training data, the model may predict the risk of "electrical fault" and provide the predicted time point.

[0176] Next steps:

[0177] Based on the prediction results, the system retrieves the corresponding historical processing modes from the functional failure mode library and plans corresponding maintenance and processing measures to minimize the impact of the failure on the power grid operation.

[0178] Selection of model input features:

[0179] When training the model, it is necessary to select appropriate input features based on the characteristics of power grid equipment operation and historical records of functional failures. Specifically, feature selection will be conducted from the following aspects:

[0180] Electrical parameters:

[0181] Voltage

[0182] Current

[0183] Power

[0184] Frequency

[0185] Condition monitoring parameters:

[0186] Equipment temperature

[0187] Vibration

[0188] Noise

[0189] Equipment status:

[0190] Switch State

[0191] Running Time

[0192] Environmental parameters:

[0193] The temperature, humidity, wind speed, and pollution level of the environment in which the equipment is located.

[0194] Historical fault records:

[0195] Previous fault types, occurrence times, and handling methods, etc.

[0196] Time characteristics:

[0197] Timestamps, periodic time features (such as hours, days of the week, etc.)

[0198] The feature selection process includes data statistical analysis, correlation analysis, and feature importance assessment to ensure that there is a significant correlation between the selected features and the functional impairment.

[0199] In step S420, real-time records of power grid equipment operation are input into a trained fault prediction model to predict potential fault patterns and timing. These real-time records include equipment operating parameters and status monitoring records, providing immediate information about the equipment's current state. The model uses these records to predict potential fault patterns and the timing of fault occurrence. The significant advantage of this step is its ability to monitor equipment operating status in real time and predict potential faults promptly, thus enabling fault prevention and timely handling.

[0200] S430: Based on the predicted functional fault type, the system retrieves the corresponding handling mode for that type of functional fault from previous records in the functional fault mode library. Based on the predicted functional fault time, it determines the time window from the current time to the possible occurrence of the functional fault. Combining this with the handling mode, it sets the handling time nodes and functional fault handling steps for the predicted functional fault. After predicting the functional fault type, the system retrieves historical handling modes matching the predicted functional fault type from the functional fault mode library. These handling modes include effective handling measures, handling times, and handling results for the same type of functional fault in the past. Based on the predicted functional fault time, the system determines a time window from the current time to the possible occurrence of the functional fault, and sets the handling time nodes and detailed functional fault handling steps based on this. The advantage of this process is that it can plan and prepare corresponding handling measures in advance based on the predicted functional fault type and time, thereby minimizing the impact of functional faults on power grid operation.

[0201] Step S400, by establishing and applying a functional failure prediction model, not only improves the accuracy and timeliness of predicting power grid equipment functional failures, but also effectively reduces potential risks to power grid operation by planning handling measures in advance. This proactive functional failure management approach significantly enhances the reliability and stability of the power grid, providing strong decision support for power grid operators.

[0202] Figure 5 is a structural block diagram of a multi-level power grid evaluation system based on power grid equipment traceability record items provided in an embodiment of the present invention. The system includes:

[0203] The power grid equipment record item acquisition and processing module 100 is used to collect real-time and previous record items of power grid equipment operation from the power grid equipment, including operating parameters, status monitoring record items, environmental parameters and previous maintenance record items, and to preprocess the acquired record items.

[0204] The previous record item analysis and pattern recognition module 200 is used to analyze and identify patterns and anomalies in the previous record items based on the previous record items, analyze the potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and establish a functional failure pattern library. Combined with the previous record items, the content related to the same type of functional failure is used as storage tags and stored in the functional failure pattern library.

[0205] The processing mode rating and optimization module 300 is used to rate the processing modes of functional impairment in the previously recorded items, set level thresholds, and delete processing modes that are below the level thresholds.

[0206] The Fault Prediction and Decision Support Module 400 is used to establish a fault prediction model. By combining the real-time records of the power grid equipment, it predicts the potential fault patterns and time points of the equipment. At the same time, it searches the fault pattern library based on the type of fault to obtain the processing mode for the fault in the previous records, and plans the fault processing decision according to the processing time node.

[0207] Specifically:

[0208] 1. Power Grid Equipment Record Item Acquisition and Processing Module 100

[0209] Input: Real-time and previous records obtained from power grid equipment.

[0210] Processing: Preprocess the acquired records, including data cleaning, format standardization, missing value handling, and outlier detection.

[0211] Output: Preprocessed records (including running parameters, status monitoring records, environmental parameters, and previous maintenance records).

[0212] Data transmission method:

[0213] Use API (Application Programming Interface) or message queue technology (such as Kafka, RabbitMQ) to send the preprocessed data to the previous record analysis and pattern recognition module 200.

[0214] 2. Previous Record Item Analysis and Pattern Recognition Module 200

[0215] Input: Preprocessed record items from the power grid equipment record item acquisition and processing module 100.

[0216] Processing: Analyze and identify patterns and anomalies in the records, establish a functional impairment pattern library, and conduct functional impairment probability analysis.

[0217] Output: A library of dysfunction patterns, along with associated storage tags (processing patterns for the same type of dysfunction).

[0218] Data transmission method:

[0219] The generated dysfunction pattern library and storage tags are stored in the database through a database interface (such as SQL or NoSQL), and the relevant information is sent to the processing pattern rating and optimization module 300.

[0220] 3. Processing Mode Rating and Optimization Module 300

[0221] Input: Functional disorder pattern library and processing patterns from the previous record analysis and pattern recognition module 200.

[0222] Processing: Rate the processing modes, set level thresholds, and delete processing modes that are below the threshold.

[0223] Output: Optimized processing mode and rating information.

[0224] Data transmission method:

[0225] The optimized processing mode is updated and stored in the database through the database interface, and the rating information is passed to the functional impairment prediction and decision support module 400.

[0226] 4. Functional Disorder Prediction and Decision Support Module 400

[0227] Input: Real-time records from the power grid equipment record acquisition and processing module 100, and optimized processing modes from the processing mode rating and optimization module 300.

[0228] Processing: Establish a functional impairment prediction model, make predictions based on real-time recorded items, and simultaneously search for corresponding processing patterns in the functional impairment pattern library to plan processing decisions.

[0229] Output: Predicted functional impairment pattern, timing of occurrence, and corresponding decision.

[0230] Data transmission method:

[0231] The predicted results and processing decisions are sent to the relevant response system (such as the operation and maintenance management system) via API or message queue so that the operation and maintenance personnel can respond in a timely manner.

[0232] Overall data flow diagram

[0233] Power grid equipment record item acquisition and processing module 100 → sends the pre-processed record items to the previous record item analysis and pattern recognition module 200.

[0234] The previously recorded item analysis and pattern recognition module 200 generates a functional disorder pattern library, which is then sent to the processing pattern rating and optimization module 300.

[0235] Processing mode rating and optimization module 300 → The optimized processing mode is updated to the database and simultaneously sent to the functional failure prediction and decision support module 400.

[0236] Functional Failure Prediction and Decision Support Module 400 → Sends prediction results to the operation and maintenance management system.

[0237] Figure 6 is a structural block diagram of the previous record item analysis and pattern recognition module provided in an embodiment of the present invention. As shown in Figure 6, the previous record item analysis and pattern recognition module includes:

[0238] The record item pattern recognition unit 210 is used to establish a functional failure recognition model, analyze the previously collected records of the operation of the power grid equipment, identify the normal operation mode of the equipment, detect record item points that deviate from the normal mode, and mark the record item points.

[0239] The functional impairment probability detection and assessment unit 220 performs secondary analysis on the marked record items, combines the subsequent record items of the record item to judge the development trend of potential functional impairment, and judges the probability and pattern of functional impairment caused by the record item.

[0240] The functional fault pattern library management unit 230 establishes a functional fault pattern library, stores the functional fault-related content existing in the previous records of the operation of power grid equipment in the functional fault pattern library, and stores all previous records related to the functional fault type under the storage tag of that type, using the functional fault type as the storage tag.

[0241] Figure 7 is a structural block diagram of the processing mode rating and optimization module provided in an embodiment of the present invention. As shown in Figure 7, the processing mode rating and optimization module includes:

[0242] The dysfunction treatment record item filtering unit 310 is used to filter out all records related to dysfunction treatment from previous record items, including the time of occurrence of dysfunction and the type of dysfunction, and extract the treatment measures, treatment time and treatment results of the type of dysfunction from the dysfunction pattern library according to the type of dysfunction.

[0243] The treatment measure rating unit 320 is used to formulate rating rules, including treatment time, treatment cost, functional impairment recovery time, and the time of impact of highly impaired functional impairment, and to evaluate the treatment measures and treatment results.

[0244] According to the rating rules, a level threshold of 330 is set. Functional impairment measures that are rated below the level threshold are defined as invalid measures and removed from the functional impairment pattern library.

[0245] Figure 8 is a structural block diagram of the functional impairment prediction and decision support module provided in an embodiment of the present invention. As shown in Figure 8, the functional impairment prediction and decision support module includes:

[0246] The functional impairment prediction model building training unit 410 is used to build a functional impairment prediction model and train the functional impairment prediction model using previously recorded items stored in the functional impairment pattern library.

[0247] The functional failure mode time analysis and prediction unit 420 is used to substitute the real-time records of the operation of the power grid equipment into the functional failure prediction model to predict the possible functional failure modes and time points of the power grid equipment.

[0248] The functional impairment processing time step planning unit 430 is used to retrieve the processing mode of the functional impairment type in the previous record in the functional impairment mode library according to the predicted functional impairment type, and determine the time window from the current time to the possible occurrence of the functional impairment based on the predicted functional impairment time, and set the processing time node and functional impairment processing steps for the predicted functional impairment in combination with the processing mode.

[0249] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0250] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, record library, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual record rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0251] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0252] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0253] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-level evaluation method for power grids based on power grid equipment traceability record items, characterized in that, The method includes: Collect real-time and previous records of power grid equipment operation from the power grid equipment, including operating parameters, status monitoring records, environmental parameters and previous maintenance records, and preprocess the acquired records. Based on previous records, analyze and identify patterns and anomalies in the previous records, analyze potential functional failure patterns and the probability of functional failures in power grid equipment, and establish a functional failure pattern library. Combined with previous records, content related to the same type of functional failure is used as storage tags and stored in the functional failure pattern library. Rate the treatment patterns for functional impairments in the previous records, set level thresholds, and delete treatment patterns that are below the level thresholds. Establish a functional failure prediction model, combine real-time records of power grid equipment to predict potential functional failure patterns and timing points, and simultaneously search the functional failure pattern library based on the type of functional failure to obtain the handling patterns for that type of functional failure from previous records, and plan the functional failure handling decisions according to the handling time nodes.

2. The method according to claim 1, characterized in that, The analysis and identification of patterns and anomalies in previously recorded items, the analysis of potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and the establishment of a functional failure pattern library specifically include: Establish a functional malfunction identification model, analyze the previously collected records of power grid equipment operation, identify the normal operation mode of the equipment, detect the record points that deviate from the normal mode, and mark the record points. A secondary analysis is performed on the marked record items, and the development trend of potential functional impairment is judged in combination with the subsequent record items of the record item. The probability and pattern of functional impairment caused by the record item are also judged. Establish a functional failure pattern library, store the functional failure-related content in the previous records of power grid equipment operation in the functional failure pattern library, and use the functional failure type as the storage label to store all previous records related to the functional failure type under the storage label of that type.

3. The method according to claim 2, characterized in that, The process of rating the handling patterns of functional impairments in previously recorded items, setting level thresholds, and deleting handling patterns below the level thresholds specifically includes: Filter out all records related to the management of functional impairment from the previous records, including the time of occurrence of functional impairment and the type of functional impairment, and extract the management measures, management time and management results of the type of functional impairment from the functional impairment pattern library according to the functional impairment type; Establish rating rules, including indicators such as processing time, processing cost, functional impairment recovery time, and the time of impact of highly impaired functional impairment, and conduct level assessments of processing measures and results; According to the rating rules, a level threshold is set, and functional impairment treatment measures that are rated below the level threshold are defined as invalid measures and removed from the functional impairment pattern library.

4. The method according to claim 3, characterized in that, The establishment of the functional failure prediction model, which combines real-time records of power grid equipment to predict potential functional failure patterns and timing, specifically includes: Establish a functional impairment prediction model and train the functional impairment prediction model using previously recorded items stored in the functional impairment pattern library; Substitute the real-time records of power grid equipment operation into the functional failure prediction model to predict the possible functional failure modes and timing of power grid equipment. Based on the predicted type of functional impairment, the processing mode for that type of functional impairment is retrieved from the previous records in the functional impairment pattern library. Based on the predicted time of functional impairment, the time window from the current time to the possible occurrence of the functional impairment is determined. Combined with the processing mode, the processing time node and functional impairment processing steps are set for the predicted functional impairment.

5. A multi-level power grid evaluation system based on power grid equipment traceability record items, characterized in that, The system includes: The power grid equipment record item acquisition and processing module is used to collect real-time and previous records of power grid equipment operation from the power grid equipment, including operating parameters, status monitoring records, environmental parameters and previous maintenance records, and to preprocess the acquired records. The previous record item analysis and pattern recognition module is used to analyze and identify patterns and anomalies in previous records based on previous records, analyze potential functional failure patterns and the probability of functional failure occurrence of power grid equipment, and establish a functional failure pattern library. Combined with previous records, content related to the same type of functional failure is used as storage tags and stored in the functional failure pattern library. The processing mode rating and optimization module is used to rate the processing modes of functional impairment in the previous records, set level thresholds, and delete processing modes that are below the level thresholds. The Fault Prediction and Decision Support Module is used to establish a fault prediction model. By combining real-time records of power grid equipment, it predicts the potential fault patterns and timing of equipment failures. At the same time, it searches the fault pattern library based on the type of fault to obtain the handling mode for that type of fault from previous records, and plans the fault handling decision according to the handling time node.

6. The system according to claim 5, characterized in that, The previous record analysis and pattern recognition module includes: The record item pattern recognition unit is used to establish a functional failure recognition model, analyze the previously collected records of the operation of the power grid equipment, identify the normal operation mode of the equipment, detect record item points that deviate from the normal mode, and mark the record item points. The functional impairment probability detection and assessment unit performs secondary analysis on the marked record items, combines the subsequent record items of the record item to judge the development trend of potential functional impairment, and judges the probability and pattern of functional impairment caused by the record item. The Fault Pattern Library Management Unit establishes a fault pattern library, stores fault-related content from previous records of power grid equipment operation in the fault pattern library, and stores all previous records related to the fault type under the corresponding storage tag.

7. The system according to claim 6, characterized in that, The processing mode rating and optimization module includes: The dysfunction management record item filtering unit is used to filter out all records related to dysfunction management from previous record items, including the time of occurrence of dysfunction and the type of dysfunction, and extract the management measures, management time and management results of the type of dysfunction from the dysfunction pattern library according to the dysfunction type; The treatment measure rating unit is used to formulate rating rules, including indicators such as treatment time, treatment cost, functional impairment recovery time, and the time of impact of highly impaired functional impairment, and to evaluate the treatment measures and treatment results. According to the rating rules, a level threshold is set, and functional impairment treatment measures that are rated below the level threshold are defined as invalid measures and removed from the functional impairment pattern library.

8. The system according to claim 7, characterized in that, The functional impairment prediction and decision support module includes: The functional impairment prediction model building training unit is used to build a functional impairment prediction model and train the functional impairment prediction model using previous records stored in the functional impairment pattern library. The functional failure mode time analysis and prediction unit is used to input the real-time records of the operation of power grid equipment into the functional failure prediction model to predict the possible functional failure modes and time points of the power grid equipment. The functional impairment processing time step planning unit is used to retrieve the processing mode for the type of functional impairment from the previous record in the functional impairment mode library according to the predicted functional impairment type, and determine the time window from the current time to the possible occurrence of the functional impairment based on the predicted functional impairment time. In combination with the processing mode, it sets the processing time node and functional impairment processing steps for the predicted functional impairment.

Images