A health status assessment method and system for multi-source heterogeneous power data

Abstract

The present invention discloses a health status assessment method and system for multi-source heterogeneous power data. The health status assessment method and system for multi-source heterogeneous power data include: S110, constructing a unified data model for multi-source heterogeneous power data; S120, extracting multidimensional data features based on the unified data model; S130, constructing a health status assessment model based on the extracted data features; S140, establishing a real-time health status monitoring and alarm mechanism based on the health status assessment model; the present application provides a real-time monitoring and alarm mechanism for the health status of multi-source heterogeneous power data, accurately detecting abnormal situations of power data through a multi-feature assessment model, and providing convenience for operation and maintenance personnel.

Description

Technical Field

[0001] The present invention relates to a time series feature model training technology, and in particular to a health status assessment method and system for multi-source heterogeneous power data. Background Art

[0002] The current power system is developing rapidly towards intelligence and digitalization, and the safety and reliability of equipment operation have become key factors affecting the stability of the power grid. However, power equipment usually operates in a complex and changeable environment. Its operating status is affected by multiple factors such as load fluctuations, external environment, and equipment aging. It is prone to failures and even large-scale power outages. In order to ensure the stable operation of the equipment, it is necessary to monitor the health status of the equipment in real time and warn of potential risks in a timely manner. However, the operation data of power equipment has the characteristics of multi-source heterogeneity, nonlinearity, and high dimensionality. Traditional monitoring methods are often difficult to efficiently integrate multi-source data and cannot comprehensively and accurately evaluate the health status of the equipment. At the same time, the existing evaluation methods lack dynamics and intelligence, and only rely on fixed thresholds or single features for monitoring, which is prone to false alarms or missed alarms. Therefore, there is an urgent need for a health status assessment method that can integrate multi-source heterogeneous data, dynamically evaluate the operating status of equipment, and support intelligent warnings to improve the accuracy and timeliness of power equipment operation and maintenance and ensure the safe and reliable operation of the power grid. Summary of the invention

[0003] The present invention provides a method for evaluating the health status of multi-source heterogeneous power data, comprising:

[0004] S110, construct a unified data model for multi-source heterogeneous power data;

[0005] S120, extracting multidimensional data features according to a unified data model;

[0006] S130, constructing a health status assessment model based on the extracted data features;

[0007] S140. Establish a real-time health status monitoring and alarm mechanism based on the health status assessment model.

[0008] A method for assessing the health status of multi-source heterogeneous power data as described above, wherein the method for constructing a unified data model for multi-source heterogeneous power data comprises the following sub-steps:

[0009] Time-align heterogeneous data;

[0010] Perform data cleaning and standardization on heterogeneous data after time alignment.

[0011] A method for assessing the health status of multi-source heterogeneous power data as described above, wherein the method for extracting multidimensional data features according to a unified data model comprises the following sub-steps:

[0012] Preliminary extraction of multidimensional data features;

[0013] Calculate statistical characteristics of preliminary feature data;

[0014] According to the statistical characteristics, the importance of feature data is evaluated.

[0015] As described above, a health status assessment method for multi-source heterogeneous power data is described, in which different features are weighted by feature importance, the impact of key features on the health status is highlighted, and the role of secondary features is weakened. Time series extraction values ​​are used to reflect the operating status of the equipment in real time. The benchmark value is used as a comparison basis to measure the degree of deviation between the current state of the equipment and the normal state, thereby constructing a health status assessment model.

[0016] A health status assessment method for multi-source heterogeneous power data as described above, wherein a real-time health status monitoring and alarm mechanism is established based on a health status assessment model, and time series characteristic data of equipment operation is continuously collected. In combination with the characteristic contribution value and the baseline value, the health score of the current equipment is dynamically calculated using the health score formula, and a health status curve is generated by real-time monitoring to display the health status change trend of the equipment in the time series.

[0017] The present invention also provides a health status assessment system for multi-source heterogeneous power data, including: a unified data module, a multi-dimensional feature data extraction module, a health status assessment module, and a monitoring and alarm module.

[0018] Unified data module: build a unified data model for multi-source heterogeneous power data;

[0019] Multi-dimensional feature data extraction module: extract multi-dimensional data features based on a unified data model;

[0020] Health status assessment module: constructs a health status assessment model based on the extracted data features;

[0021] Monitoring and alarm module: Establish real-time health status monitoring and alarm mechanism based on the health status assessment model.

[0022] In the health status assessment system for multi-source heterogeneous power data as described above, a method for constructing a unified data model for multi-source heterogeneous power data comprises the following sub-steps:

[0023] Time-align heterogeneous data;

[0024] Perform data cleaning and standardization on heterogeneous data after time alignment.

[0025] In the health status assessment system for multi-source heterogeneous power data as described above, the method for extracting multidimensional data features according to the unified data model comprises the following sub-steps:

[0026] Preliminary extraction of multidimensional data features;

[0027] Calculate statistical characteristics of preliminary feature data;

[0028] According to the statistical characteristics, the importance of feature data is evaluated.

[0029] As described above, a health status assessment system for multi-source heterogeneous power data is described, in which different features are weighted by feature importance, the impact of key features on the health status is highlighted, and the role of secondary features is weakened. Time series extraction values ​​are used to reflect the operating status of the equipment in real time. The benchmark value is used as a comparison basis to measure the degree of deviation between the current status of the equipment and the normal status, thereby constructing a health status assessment model.

[0030] A health status assessment system for multi-source heterogeneous power data as described above is established, in which a real-time health status monitoring and alarm mechanism is established based on a health status assessment model, and the time series characteristic data of equipment operation is continuously collected. In combination with the characteristic contribution value and the baseline value, the health score of the current equipment is dynamically calculated using the health score formula, and a health status curve is generated by real-time monitoring to display the health status change trend of the equipment in the time series.

[0031] The beneficial effects achieved by the present invention are as follows: a real-time monitoring and alarm mechanism is provided for the health status of multi-source heterogeneous power data, and an abnormal situation of power data is accurately detected through a multi-feature evaluation model, providing convenience for operation and maintenance personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0033] Figure 1 This is a flow chart of a method for evaluating the health status of multi-source heterogeneous power data provided in Example 1 of the present application;

[0034] Figure 2 This is a schematic diagram of a health status assessment system for multi-source heterogeneous power data provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0035] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0036] Embodiment 1

[0037] like Figure 1 As shown, the first embodiment of the present application provides a method for evaluating the health status of multi-source heterogeneous power data:

[0038] Step S110: constructing a unified data model for multi-source heterogeneous power data;

[0039] The main sources of power data include sensors, monitoring equipment and historical records. Sensor data is real-time monitoring information of power equipment operation, usually collected at a high frequency, and has strong time series characteristics; the status information recorded by monitoring equipment is relatively comprehensive, usually collected by cameras, smart meters, circuit breaker status sensors, etc. in the power system, with a low collection frequency, but the record of key states is relatively accurate and can provide time tags; historical record data comes from equipment maintenance logs and long-term archived operation data. The data timeliness is weak, but it is crucial for long-term trend analysis and model training. The method of building a unified data model for multi-source heterogeneous power data includes the following sub-steps:

[0040] Step S111, time-aligning heterogeneous data;

[0041] There may be inconsistent timestamps, data delays or missing, and different time references between heterogeneous data. The time alignment method is used to unify asynchronous data from different sources into a standard time series for comprehensive analysis and to ensure that different data can be accurately associated at the same time point. Specifically, the formula is used: T aligned ={t1,t2,...,t n},t i =max(T sensor1 ,T sensor2 ,...) to time align heterogeneous data, where T aligned represents the aligned unified time series set, which consists of multiple time points {t1, t2, ..., t n}, the time points are arranged in ascending order, t i represents the i-th specific time point in the unified time series,

[0042] (T sensor1 ,T sensor2,...) represents the original timestamp sets of different data sources. Each set contains the timestamps of all sampling points of a data source. max means that among the timestamps of different data sources, the maximum timestamp at the current moment is taken as the aligned time point to ensure that the unified time point can cover all data sources.

[0043] Step S112: performing data cleaning and standardization on the heterogeneous data after time alignment;

[0044] After time alignment, the heterogeneous data is cleaned and standardized. First, the data integrity is checked and missing values ​​and outliers are processed. For missing values, interpolation is used to fill them to ensure the continuity of the time series; for outliers, statistical methods or machine learning algorithms are used to identify them, and abnormal data are removed or corrected according to needs. Subsequently, data from different sources are unified into the same numerical range and unit, and the data are mapped to a unified scale through normalization or standardization methods to eliminate dimensional differences between data. Finally, the cleaned and standardized data set T is generated. cleaned Provide high-quality input for subsequent feature extraction and model analysis to ensure data reliability and consistency.

[0045] Step S120: extract multidimensional data features according to the unified data model;

[0046] On the basis of a unified data model, we integrate multi-source heterogeneous power data (such as sensor data, monitoring equipment data, and historical record data), and use advanced data processing and analysis methods to extract key features based on their time series, laying the foundation for subsequent health status assessment and prediction. The specific method includes the following sub-steps:

[0047] Step S121, preliminary extraction of multi-dimensional data features;

[0048] Wavelet transform is used to extract short-term and long-term features of time series data, and the hidden patterns in the data are obtained through noise reduction and decomposition. Wavelet transform is a time-frequency analysis method that can analyze time series data in both the time domain and the frequency domain. The high-frequency part reflects the detailed features or mutation points in the data, and the low-frequency part retains the global trend or long-term change characteristics of the data. The data is denoised by wavelet transform to remove interference information, and the features on different time scales during the operation of power equipment are extracted to provide clear basic data for subsequent feature calculation and analysis. Specifically, the formula is used: Perform feature extraction, where X filtered represents the extracted features, represents the wavelet transform operation, where T cleaned represents the cleaned data set, and ψ represents the selected wavelet basis function.

[0049] Step S122, calculating statistical features of preliminary feature data;

[0050] Statistical methods are used to analyze the extracted preliminary feature data, extract key features that describe data distribution and change patterns, and comprehensively characterize the operating status characteristics of the equipment (such as stability, volatility, and change trends) through statistical feature extraction, providing key reference indicators for health assessment. Specifically, the formula is used: Calculate the statistical characteristics of the characteristic data, where μ represents the mean, reflecting the central tendency of the data, σ represents the standard deviation, reflecting the volatility and dispersion of the data, and r represents the rate of change, reflecting the magnitude of the change in the data during the observation time. i ∈X filtered , i represents the subscript of the feature data, and its value ranges from 1 to n.

[0051] Step S123: evaluating the importance of the feature data according to the statistical features;

[0052] According to the statistical characteristics, the contribution of each feature to the model prediction results is evaluated, the importance of the features is quantified, the features with greater impact on health status assessment are screened out, the model input is optimized, and the accuracy and interpretability of the assessment are improved. Specifically, the formula is used:

[0053] F={μ,σ,r} for importance assessment, where φ j represents the feature importance set, F represents the full set of statistical features, S represents the statistical feature subset, ν(F) represents the predicted value of the feature set, ν(S) represents the predicted value of the model when only the feature subset S is used, when φ j >0 means that feature j has a positive contribution to the prediction result of the model. j <0 means that feature j has a negative impact on the prediction results of the model.

[0054] Step S130: construct a health status assessment model based on the extracted data features;

[0055] According to the above extracted data features, a health status assessment model is constructed. j Weight different features to highlight the impact of key features on health status and weaken the role of secondary features; use time series to extract value X filtered , real-time reflection of the operating status of the equipment, enhancing the real-time and responsiveness of the evaluation model; using the benchmark value X baseline As a basis for comparison, measure the degree of deviation between the current state of the equipment and the normal state to ensure that the evaluation has a clear reference system. Specifically, the formula is used:

[0056] Construct a health status assessment model, where H represents the health score, ranging from 0 to 100. The higher the value, the healthier the device operation status. j represents the importance of feature j, represents the eigenvalue of feature j, Indicates the reference value of feature j.

[0057] Step S140: Establish a real-time health status monitoring and alarm mechanism based on the health status assessment model;

[0058] According to the health status assessment model, a real-time health status monitoring and alarm mechanism is established. First, the time series characteristic data of equipment operation is continuously collected, and the health score H of the current equipment is dynamically calculated using the health score formula in combination with the characteristic contribution value and the benchmark value. The health status curve C is generated by real-time monitoring. H ={H1,H2,...,H t}, showing the health status change trend of the device in time series. Secondly, set the health score threshold H threshold =50, the score is divided into three states: healthy (80-100), slightly abnormal (50-80) and seriously abnormal (<50). When the score is lower than the set threshold, the alarm mechanism is triggered, and the operation and maintenance personnel are notified through SMS, email or visual interface to handle it in time. In addition, the monitoring system supports the prediction and analysis of health status, combined with the score trend C H Early warning of potential risks helps formulate preventive maintenance strategies, thereby achieving timely intervention in equipment failures and comprehensive optimization of health management.

[0059] Embodiment 2

[0060] like Figure 2 As shown, Embodiment 2 of the present application provides a health status assessment system for multi-source heterogeneous power data, including:

[0061] Unified data module 21: building a unified data model for multi-source heterogeneous power data;

[0062] The main sources of power data include sensors, monitoring equipment and historical records. Sensor data is real-time monitoring information of power equipment operation, usually collected at a high frequency, and has strong time series characteristics; the status information recorded by monitoring equipment is relatively comprehensive, usually collected by cameras, smart meters, circuit breaker status sensors, etc. in the power system, with a low collection frequency, but the record of key states is relatively accurate and can provide time tags; historical record data comes from equipment maintenance logs and long-term archived operation data. The data timeliness is weak, but it is crucial for long-term trend analysis and model training. The method of building a unified data model for multi-source heterogeneous power data includes the following sub-steps:

[0063] Time alignment module: time aligns heterogeneous data;

[0064] There may be inconsistent timestamps, data delays or missing, and different time references between heterogeneous data. The time alignment method is used to unify asynchronous data from different sources into a standard time series for comprehensive analysis and to ensure that different data can be accurately associated at the same time point. Specifically, the formula is used: T aligned ={t1,t2,...,t n},t i =max(T sensor1 ,T sensor2 ,...) to time align heterogeneous data, where T aligned represents the aligned unified time series set, which consists of multiple time points {t1, t2, ..., t n}, the time points are arranged in ascending order, t i represents the i-th specific time point in the unified time series,

[0065] (T sensor1 ,T sensor2 ,...) represents the original timestamp sets of different data sources. Each set contains the timestamps of all sampling points of a data source. max means that among the timestamps of different data sources, the maximum timestamp at the current moment is taken as the aligned time point to ensure that the unified time point can cover all data sources.

[0066] Data processing module: data cleaning and standardization of heterogeneous data after time alignment;

[0067] After time alignment, the heterogeneous data is cleaned and standardized. First, the data integrity is checked and missing values ​​and outliers are processed. For missing values, interpolation is used to fill them to ensure the continuity of the time series; for outliers, statistical methods or machine learning algorithms are used to identify them, and abnormal data are removed or corrected according to needs. Subsequently, data from different sources are unified into the same numerical range and unit, and the data are mapped to a unified scale through normalization or standardization methods to eliminate dimensional differences between data. Finally, the cleaned and standardized data set T is generated. cleaned Provide high-quality input for subsequent feature extraction and model analysis to ensure data reliability and consistency.

[0068] Multi-dimensional feature data extraction module 22: extracting multi-dimensional data features according to a unified data model;

[0069] On the basis of a unified data model, we integrate multi-source heterogeneous power data (such as sensor data, monitoring equipment data, and historical record data), and use advanced data processing and analysis methods to extract key features based on their time series, laying the foundation for subsequent health status assessment and prediction. The specific method includes the following sub-steps:

[0070] Preliminary feature extraction module: Preliminary extraction of multidimensional data features;

[0071] Wavelet transform is used to extract short-term and long-term features of time series data, and the hidden patterns in the data are obtained through noise reduction and decomposition. Wavelet transform is a time-frequency analysis method that can analyze time series data in both the time domain and the frequency domain. The high-frequency part reflects the detailed features or mutation points in the data, and the low-frequency part retains the global trend or long-term change characteristics of the data. The data is denoised by wavelet transform to remove interference information, and the features on different time scales during the operation of power equipment are extracted to provide clear basic data for subsequent feature calculation and analysis. Specifically, the formula is used: Perform feature extraction, where X filtered represents the extracted features, represents the wavelet transform operation, where T cleaned represents the cleaned data set, and ψ represents the selected wavelet basis function.

[0072] Statistical feature calculation module: calculates the statistical features of preliminary feature data;

[0073] Statistical methods are used to analyze the extracted preliminary feature data, extract key features that describe data distribution and change patterns, and comprehensively characterize the operating status characteristics of the equipment (such as stability, volatility, and change trends) through statistical feature extraction, providing key reference indicators for health assessment. Specifically, the formula is used: Calculate the statistical characteristics of the characteristic data, where μ represents the mean, reflecting the central tendency of the data, σ represents the standard deviation, reflecting the volatility and dispersion of the data, and r represents the rate of change, reflecting the magnitude of the change in the data during the observation time. i ∈X filtered , i represents the subscript of the feature data, and its value ranges from 1 to n.

[0074] Importance assessment module: evaluates the importance of feature data based on statistical characteristics;

[0075] According to the statistical characteristics, the contribution of each feature to the model prediction results is evaluated, the importance of the features is quantified, the features with greater impact on health status assessment are screened out, the model input is optimized, and the accuracy and interpretability of the assessment are improved. Specifically, the formula is used:

[0076] F={μ,σ,r} for importance assessment, where φj represents the feature importance set, F represents the full set of statistical features, S represents the statistical feature subset, ν(F) represents the predicted value of the feature set, ν(S) represents the predicted value of the model when only the feature subset S is used, when φ j >0 means that feature j has a positive contribution to the prediction result of the model. j <0 means that feature j has a negative impact on the prediction results of the model.

[0077] Health status assessment module 23: constructing a health status assessment model based on the extracted data features;

[0078] According to the above extracted data features, a health status assessment model is constructed. j Weight different features to highlight the impact of key features on health status and weaken the role of secondary features; use time series to extract value X filtered , real-time reflection of the operating status of the equipment, enhancing the real-time and responsiveness of the evaluation model; using the benchmark value X baseline As a basis for comparison, measure the degree of deviation between the current state of the equipment and the normal state to ensure that the evaluation has a clear reference system. Specifically, the formula is used:

[0079] Construct a health status assessment model, where H represents the health score, ranging from 0 to 100. The higher the value, the healthier the device operation status. j represents the importance of feature j, represents the eigenvalue of feature j, Indicates the reference value of feature j.

[0080] Monitoring and alarm module 24: Establishing a real-time health status monitoring and alarm mechanism based on the health status assessment model;

[0081] According to the health status assessment model, a real-time health status monitoring and alarm mechanism is established. First, the time series characteristic data of equipment operation is continuously collected, and the health score H of the current equipment is dynamically calculated using the health score formula in combination with the characteristic contribution value and the benchmark value. The health status curve C is generated by real-time monitoring. H ={H1,H2,...,H t}, showing the health status change trend of the device in time series. Secondly, set the health score threshold H threshold =50, the score is divided into three states: healthy (80-100), slightly abnormal (50-80) and seriously abnormal (<50). When the score is lower than the set threshold, the alarm mechanism is triggered, and the operation and maintenance personnel are notified through SMS, email or visual interface to handle it in time. In addition, the monitoring system supports the prediction and analysis of health status, combined with the score trend C HEarly warning of potential risks helps formulate preventive maintenance strategies, thereby achieving timely intervention in equipment failures and comprehensive optimization of health management.

[0082] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for assessing the health status of multi-source heterogeneous power data, characterized in that: include: S110, constructing a unified data model for multi-source heterogeneous power data; S120, extracting multidimensional data features according to a unified data model; S130, constructing a health status assessment model based on the extracted data features; S140. Establish a real-time health status monitoring and alarm mechanism based on the health status assessment model.

2. A method for evaluating the health status of multi-source heterogeneous power data according to claim 1, characterized in that: The method for constructing a unified data model for multi-source heterogeneous power data includes the following sub-steps: Time-align heterogeneous data; Perform data cleaning and standardization on heterogeneous data after time alignment.

3. A method for evaluating the health status of multi-source heterogeneous power data according to claim 2, characterized in that: According to the unified data model, the method for extracting multidimensional data features includes the following sub-steps: Preliminary extraction of multidimensional data features; Calculate statistical characteristics of preliminary feature data; According to the statistical characteristics, the importance of feature data is evaluated.

4. A method for evaluating the health status of multi-source heterogeneous power data according to claim 1, characterized in that: Different features are weighted by feature importance to highlight the impact of key features on health status and weaken the role of secondary features. Time series extraction values ​​are used to reflect the operating status of the equipment in real time. The benchmark value is used as a comparison basis to measure the degree of deviation between the current status of the equipment and the normal status, thereby constructing a health status assessment model.

5. The method for health status assessment of multi-source heterogeneous power data according to claim 1, characterized in that: Based on the health status assessment model, a real-time health status monitoring and alarm mechanism is established to continuously collect the time series characteristic data of equipment operation, and combine the characteristic contribution value and baseline value to dynamically calculate the health score of the current equipment using the health score formula. Real-time monitoring generates a health status curve to display the health status change trend of the equipment in the time series.

6. A health status assessment system for multi-source heterogeneous power data, characterized in that: include: Unified data module: build a unified data model for multi-source heterogeneous power data; Multi-dimensional feature data extraction module: extract multi-dimensional data features based on a unified data model; Health status assessment module: constructs a health status assessment model based on the extracted data features; Monitoring and alarm module: Establish real-time health status monitoring and alarm mechanism based on the health status assessment model.

7. A health status assessment system for multi-source heterogeneous power data as claimed in claim 6, characterized in that: The method for constructing a unified data model for multi-source heterogeneous power data includes the following sub-steps: Time-align heterogeneous data; Perform data cleaning and standardization on heterogeneous data after time alignment.

8. A health status assessment system for multi-source heterogeneous power data as claimed in claim 6, characterized in that: According to the unified data model, the method for extracting multidimensional data features includes the following sub-steps: Preliminary extraction of multidimensional data features; Calculate statistical characteristics of preliminary feature data; According to the statistical characteristics, the importance of feature data is evaluated.

9. A health status assessment system for multi-source heterogeneous power data as claimed in claim 6, characterized in that: Different features are weighted by feature importance to highlight the impact of key features on health status and weaken the role of secondary features. Time series extraction values ​​are used to reflect the operating status of the equipment in real time. The benchmark value is used as a comparison basis to measure the degree of deviation between the current status of the equipment and the normal status, thereby constructing a health status assessment model.

10. A health status assessment system for multi-source heterogeneous power data according to claim 6, characterized in that: Based on the health status assessment model, a real-time health status monitoring and alarm mechanism is established to continuously collect the time series characteristic data of equipment operation, and combine the characteristic contribution value and baseline value to dynamically calculate the health score of the current equipment using the health score formula. Real-time monitoring generates a health status curve to display the health status change trend of the equipment in the time series.

Images