Renewable energy access-oriented power system data monitoring and management system

Abstract

The invention discloses a renewable energy access-oriented power system data monitoring and management system, and relates to a power system data monitoring and management system. According to the invention, the model training module constructs a power system fault prediction model by using a support vector machine algorithm, and can effectively recognize the boundary between normal operation and a potential fault state. Potential faults can be predicted in advance through the model obtained through training, so that preventive maintenance is realized, loss caused by sudden faults is reduced, and reliability and safety of a power system are improved. And the fault detection module monitors real-time power system data by using a trained model, and can give an alarm in time once detecting that a data mode is matched with a fault feature. The real-time monitoring capability enables operation and maintenance personnel to respond quickly and take corresponding measures, thereby avoiding fault expansion, ensuring stable operation of the power system, shortening fault processing time and reducing maintenance cost.

Description

Technical Field

[0001] The present invention relates to a power system data monitoring system, and particularly to a power system data monitoring and management system for renewable energy access. Background Art

[0002] Data monitoring of renewable energy power systems is an important means to achieve comprehensive monitoring of the power generation processes of renewable energy sources such as wind power and solar energy. This system installs various sensors, collectors, and transmission devices to collect the operation data of power generation equipment in real time, including key information such as current, voltage, power, frequency, and ambient temperature, and uses data analysis and processing technologies to evaluate power generation efficiency, equipment status, energy quality, etc. The data monitoring system can not only improve the reliability and stability of renewable energy power systems, but also provide a scientific basis for power grid dispatching, fault diagnosis, and maintenance, thereby optimizing the energy structure and promoting the sustainable and healthy development of the renewable energy power industry.

[0003] Existing systems cannot accurately identify the key features affecting power system faults, resulting in low data analysis efficiency and prone to overfitting problems. At the same time, traditional methods cannot effectively capture the complex nonlinear relationships in power systems, resulting in inaccurate fault prediction. Summary of the Invention

[0004] The purpose of the present invention is to provide a power system data monitoring and management system for renewable energy access. This system uses the support vector machine algorithm to construct a power system fault prediction model. The model obtained through training can predict potential faults in advance, thereby realizing preventive maintenance, reducing losses caused by sudden faults, and improving the reliability and safety of power systems.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A power system data monitoring and management system for renewable energy access, the system includes: a renewable energy equipment data collection module, a power grid data collection module, an environmental data collection module, a data preprocessing module, a data storage module, a data analysis module, a visualization display module, and a system management module;

[0007] The data analysis module is internally provided with a feature selection module, a model training module, a fault detection module, and a fault location module;

[0008] The renewable energy equipment data collection module, the power grid data collection module, and the environmental data collection module serve as the front-end perception layer of the system, responsible for collecting various relevant data in real time and transmitting these data to the data processing module;

[0009] The data processing module cleans, transforms, and normalizes the received raw data to ensure data quality and consistency, and then sends the processed data to the data storage module;

[0010] The data storage module includes a real-time database and a historical database, responsible for data storage, backup, and recovery, and providing data support for the data analysis module at the same time;

[0011] The data analysis module deeply analyzes the stored data, including status monitoring, fault diagnosis, predictive analysis, and optimization suggestions. The analysis results will be fed back to the visualization display module and the alarm management module;

[0012] The visualization display module visually displays real-time data, historical data, and analysis results in the form of charts and curves, providing users with convenient means for data viewing and analysis;

[0013] The system management module runs through the entire system, responsible for user management, log management, and system configuration, ensuring the security, stability, and configurability of the system.

[0014] In a preferred embodiment, the sensor network of the renewable energy device data acquisition module is deployed on the photovoltaic panel array. Each photovoltaic panel is equipped with a micro-inverter, and the inverter is built-in with current sensors and voltage sensors for real-time monitoring of output current and voltage;

[0015] The grid data acquisition module installs high-precision smart meters and remote terminal units at various key nodes of the grid, including substations and feeder outlets; the smart meters are responsible for real-time measurement of grid current, voltage, active power, and reactive power parameters and outputting them through digital signals; the RTU is responsible for periodically reading data from the smart meters and transmitting the data to the monitoring center through the fiber optic communication network; the RTU also has an event recording function, which can record grid abnormal events and send the event information together with the time stamp for real-time monitoring and analysis of the grid status;

[0016] The detailed configuration of the environmental data acquisition module includes: installing multiple environmental monitoring stations at different locations in the renewable energy power generation field. Each monitoring station includes anemometers, wind vanes, thermometers, hygrometers, and barometer sensors; these sensors are connected to the data collector by wired or wireless means. The data collector is responsible for periodically reading sensor data and performing preliminary data verification; the collected environmental data is transmitted to the data storage server through the LoRa or Wi-Fi network. The environmental data analysis software on the server side will evaluate the impact of environmental factors on the renewable energy power generation efficiency based on these data and provide a basis for adjusting the power generation strategy.

[0017] In a preferred embodiment, the data preprocessing module uses a moving average filtering algorithm to reduce the influence of random noise by calculating the average value of data within a certain time window. The calculation formula for moving average filtering is as follows:

[0018]

[0019] Where: y[n] is the filtered output value at the nth time point; x[n] is the original data value at the nth time point; N is the size of the moving average window, that is, the number of adjacent data points considered; i is the index within the window, ranging from 0 to N - 1.

[0020] In a preferred embodiment, the data storage module uses the MySQL relational database management system to store and manage the collected data; multiple tables are designed in the database, and each table corresponds to different types of data, including the operation data table of renewable energy equipment, the power grid parameter data table, and the environmental monitoring data table; each table contains fields including timestamp, device ID, and measurement value; the data storage module optimizes the data writing and query efficiency by writing stored procedures and triggers. At the same time, a strategy of regular full backup and real-time incremental backup is adopted to ensure data security.

[0021] In a preferred embodiment, the feature selection module screens out features that have an important impact on fault prediction and diagnosis from the preprocessed data; the feature selection and optimization module uses the ReliefF algorithm. First, for each feature Xi, the system traverses all samples to find the nearest neighbor of the same class and different class samples for each sample; then, it calculates the feature differences diff(Xi,j,H) and diff(Xi,j,M), where diff(Xi,j,H) represents the difference of feature Xi among samples of the same class, and diff(Xi,j,M) represents the difference among samples of different classes; then, according to the formula:

[0022] W(Xi) = ∑ j＝1 m (diff(Xi,j,H) - diff(Xi,j,M)) to calculate the feature weight, where W(Xi) represents the weight of feature Xi, and m represents the number of samples; finally, the system selects features with higher weights as the input feature set for model training.

[0023] In a preferred embodiment, the model training module uses the support vector machine algorithm to construct a fault prediction model; then it finds the optimal classification hyperplane through training data, that is, solves the optimal parameter β so that the distance from each support vector to the separation hyperplane is not less than 1 and there is no misjudgment; during the training process, the system uses the cross-validation method to optimize the model parameters. After training is completed, the model is used for classification prediction of new data to identify whether there is a fault in the power equipment;

[0024] Among them, the kernel function adopts the Gaussian kernel, and the calculation formula is:

[0025]

[0026] Among them, K(x_i,x_j) represents the Gaussian kernel function, x_i and x_j represent samples, and σ represents the kernel function parameter; through this formula, the system maps the input data to a high-dimensional space, searches for the optimal classification hyperplane, and realizes fault classification.

[0027] In a preferred embodiment, the fault detection module uses the modulus maximum detection algorithm based on wavelet transform for fault detection, specifically including:

[0028] S1. Extraction of wavelet coefficient modulus maximum:

[0029] Calculation formula: M_{j,k}=\max(|W_{j,k}|)

[0030] Among them:

[0031] M_{j,k} is the wavelet coefficient modulus maximum at scale j and position k;

[0032] W_{j,k} is the wavelet coefficient, representing the signal characteristics at scale j and position k;

[0033] S2. Screening of modulus maximum, removing the modulus maximum caused by noise and other non-faults, and retaining possible fault characteristics; by setting a threshold, screening out the modulus maximum exceeding the threshold;

[0034] The calculation formula is:

[0035] Parameter definition:

[0036] T_{j} is the threshold at scale j;

[0037] σj is the noise standard deviation estimate at scale j;

[0038] N is the length of the signal;

[0039] log is the natural logarithm;

[0040] The screening condition is M_{j,k}>T_{j}

[0041] S3. Fault feature recognition, identifying whether the screened modulus maximum belongs to the fault feature; by analyzing the distribution, amplitude, and duration characteristics of the modulus maximum, judging whether it meets the conditions of the fault feature;

[0042] [A_{threshold}) represents the set fault amplitude threshold;

[0043] $D_{threshold}$ represents the set fault duration threshold;

[0044] Judgment condition: $M_{j,k}>A_{threshold}$;

[0045] And $D_{j,k}>D_{threshold}$;

[0046] S4. Fault detection decision. Based on the identified fault characteristics, make a fault detection decision. If there is a modulus maximum value that meets the fault characteristics, it is judged as a fault; otherwise, it is judged as no fault;

[0047] Decision rule: If The fault characteristic conditions are met, then the fault detection result = fault;

[0048] Otherwise, the fault detection result = no fault

[0049] The fault detection sub-module realizes the detection of electrical faults through steps such as extracting the modulus maximum value of wavelet coefficients, screening the modulus maximum value, identifying fault characteristics, and making a fault detection decision; the calculation formula and parameter definition are adjusted and optimized according to specific applications and signal characteristics; this algorithm utilizes the multi-scale analysis ability of wavelet transform to effectively detect fault characteristics in electrical signals and provide accurate information for subsequent fault location.

[0050] In a preferred embodiment, the fault location module uses the modulus maximum value location algorithm based on wavelet transform to be responsible for determining the specific location of the fault according to the fault characteristic information provided by the fault detection sub-module. The specific steps include:

[0051] S1. Determination of fault characteristic points

[0052] Purpose: Determine the modulus maximum value points screened by the fault detection sub-module as fault characteristic points;

[0053] Input: The screened modulus maximum value $M_{j,k}$ and its corresponding positions $k$ and scales $j$;

[0054] S2. Estimation of fault propagation speed. Estimate the fault propagation speed in the electrical system for calculating the fault location. According to the physical parameters and configuration of the electrical system, estimate the fault propagation speed. The calculation formula is:

[0055]

[0056] Where:

[0057] $v$ is the fault propagation speed;

[0058] $L'$ is the equivalent inductance of the electrical system;

[0059] $C'$ is the equivalent capacitance of the electrical system;

[0060] S3. Fault location calculation: Calculate the fault location based on the fault characteristic points and the fault propagation speed; utilize the time-frequency location characteristics of wavelet transform and combine with the fault propagation speed to calculate the fault location; the calculation formula is: d = v·Δt

[0061] where d is the fault location, that is, the distance from the fault point to the detection point;

[0062] Δt is the time delay of the fault characteristic point in time, that is, the difference between the time corresponding to the modulus maximum point and the fault occurrence time;

[0063] If there are multiple detection points, fuse the location results of multiple points to improve the location accuracy; adopt weighted average or other fusion algorithms and combine the location results of multiple detection points;

[0064] The weighted average calculation formula is:

[0065] where:

[0066] dfinal is the finally determined fault location;

[0067] d_i is the fault location calculated by the i-th detection point;

[0068] w_i is the weight of the i-th detection point, which can be set according to factors such as signal quality and the importance of the detection point.

[0069] In a preferred embodiment, develop a Web-based monitoring platform for visualization display, using HTML5, CSS3 and JavaScript front-end technologies, and combine with a chart library to create dynamic charts and graphics; users can access this platform through a browser to view the real-time data monitoring screen, including line charts showing the change in power generation, bar charts showing the comparison of power generation efficiency of different devices, and pie charts showing the energy distribution.

[0070] In a preferred embodiment, the detailed functions of the system management module include: providing a background management interface that allows the system administrator to create, delete user accounts and assign permissions; permission control ensures that users at different levels can access corresponding levels of data and functions; the log management function records all user operations and system events for easy tracking and auditing; the fault alarm system automatically detects system anomalies through preset rules and notifies the administrator through multiple channels; the system backup function is executed automatically regularly, and the backup content includes user data, system configuration and key business data to support the rapid recovery of the system in case of a failure.

[0071] The beneficial effects of the present invention are:

[0072] 1. The feature selection module of the present invention uses the ReliefF algorithm to screen out features that have an important impact on power system fault prediction and diagnosis from the preprocessed data, significantly improving the efficiency and accuracy of data analysis. The ReliefF algorithm can take into account the interactions between features and screen out those characteristics that can stably reflect the state of the power system under different operating conditions, thereby reducing the data dimension, lowering the computational complexity, and also reducing the risk of overfitting. This step ensures that the most relevant information is used in the subsequent model training and fault detection processes, improving the performance of the entire system.

[0073] 2. The model training module of the present invention uses the support vector machine algorithm to construct a power system fault prediction model. This algorithm has advantages in dealing with nonlinear problems and classification problems and can effectively identify the boundary between normal operations and potential fault states. The model obtained through training can predict potential faults in advance, thereby realizing preventive maintenance, reducing losses caused by sudden faults, and improving the reliability and safety of the power system. The fault detection module uses the trained model to monitor real-time power system data. Once it detects that the data pattern matches the fault characteristics, it can issue an alarm in a timely manner. This real-time monitoring ability enables operation and maintenance personnel to respond quickly, take corresponding measures, avoid the expansion of faults, and ensure the stable operation of the power system. The fault location module determines the specific location of the fault through a weighted average calculation formula. By assigning weights to different detection points in this step, the system can more accurately locate the fault source, reduce misdiagnosis and misoperations, shorten the fault handling time, and reduce the maintenance cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 is the overall system block diagram of the present invention;

[0075] Figure 2 is the system block diagram of the data analysis module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0076] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present 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 only used to explain the present invention and are not used to limit the present invention.

[0077] Referring to Figure 1-2 , a power system data monitoring and management system for renewable energy access, the system includes: a renewable energy device data acquisition module, a power grid data acquisition module, an environmental data acquisition module, a data preprocessing module, a data storage module, a data analysis module, a visualization display module, and a system management module;

[0078] The internal settings of the data analysis module include a feature selection module, a model training module, a fault detection module, and a fault location module;

[0079] The renewable energy device data acquisition module, the power grid data acquisition module, and the environmental data acquisition module, as the front-end perception layer of the system, are responsible for collecting various relevant data in real time and transmitting this data to the data processing module.

[0080] The data processing module cleans, transforms, and normalizes the received raw data to ensure the quality and consistency of the data, and then sends the processed data to the data storage module.

[0081] The data storage module includes a real-time database and a historical database, and is responsible for data storage, backup, and recovery. At the same time, it provides data support for the data analysis module.

[0082] The data analysis module deeply analyzes the stored data, including status monitoring, fault diagnosis, predictive analysis, and optimization suggestions. The analysis results will be fed back to the visualization display module and the alarm management module.

[0083] The visualization display module intuitively displays real-time data, historical data, and analysis results in the form of charts, curves, etc., providing users with convenient means for data viewing and analysis.

[0084] The system management module runs through the entire system and is responsible for user management, log management, and system configuration to ensure the security, stability, and configurability of the system.

[0085] The sensor network of the renewable energy device data acquisition module is deployed on the photovoltaic panel array. Each photovoltaic panel is equipped with a micro-inverter, and the inverter is built-in with current sensors and voltage sensors for real-time monitoring of the output current and voltage. In addition, temperature sensors are installed on the back of the photovoltaic panel to monitor the working temperature changes. The data acquisition card is connected to the sensors, collects data through the RS-485 communication protocol, and uses time synchronization technology to ensure the accuracy of all sensor data acquisition. The data acquisition card also has edge computing capabilities, can perform preliminary noise reduction and compression processing on the data, and then upload the processed data to the central server through the 4G network to ensure the real-time and reliability of the data;

[0086] The power grid data acquisition module installs high-precision smart meters and remote terminal units (RTUs) at various key nodes of the power grid, such as substations and feeder outlets. The smart meters are responsible for measuring parameters of the power grid in real time, such as current, voltage, active power, and reactive power, and outputting them through digital signals. The RTU is responsible for periodically reading data from the smart meters and transmitting the data to the monitoring center through an optical fiber communication network. The RTU also has an event recording function, which can record abnormal events of the power grid, such as tripping and overload, and send the event information together with the timestamp for real-time monitoring and analysis of the power grid status;

[0087] The detailed configuration of the environmental data acquisition module includes: installing multiple environmental monitoring stations at different locations in the renewable energy power generation plant. Each monitoring station contains sensors such as anemometers, wind vanes, thermometers, hygrometers, and barometers. These sensors are connected to the data collector by wired or wireless means. The data collector is responsible for regularly reading the sensor data and performing preliminary data verification. The collected environmental data is transmitted to the data storage server through the LoRa or Wi-Fi network. The environmental data analysis software on the server side will evaluate the impact of environmental factors on the renewable energy power generation efficiency based on this data and provide a basis for adjusting the power generation strategy.

[0088] The data preprocessing module uses the moving average filtering algorithm to reduce the influence of random noise by calculating the average value of data within a certain time window. The calculation formula for moving average filtering is as follows:

[0089]

[0090] Where: y[n] is the filtered output value at the nth time point. x[n] is the original data value at the nth time point. N is the size of the moving average window, that is, the number of adjacent data points considered. i is the index within the window, ranging from 0 to N - 1.

[0091] The data storage module uses the MySQL relational database management system to store and manage the collected data. Multiple tables are designed in the database, and each table corresponds to different types of data, such as renewable energy equipment operation data tables, power grid parameter data tables, and environmental monitoring data tables, etc. Each table contains fields such as timestamp, device ID, measurement value, etc. The data storage module optimizes the data writing and query efficiency by writing stored procedures and triggers. At the same time, it adopts the strategies of regular full backup and real-time incremental backup to ensure data security. In addition, the database also realizes read-write separation and load balancing to cope with the storage and access requirements of a large amount of data.

[0092] The feature selection module screens out the features that have an important impact on fault prediction and diagnosis from the preprocessed data; the feature selection and optimization module uses the ReliefF algorithm. First, for each feature Xi, the system traverses all samples to find the nearest neighbor samples of the same class and different classes for each sample; then, it calculates the feature differences diff(Xi,j,H) and diff(Xi,j,M), where diff(Xi,j,H) represents the difference of feature Xi among samples of the same class, and diff(Xi,j,M) represents the difference among samples of different classes; next, according to the formula:

[0093] W(Xi) = ∑ j＝1 m (diff(Xi,j,H) - diff(Xi,j,M)) to calculate the feature weight, where W(Xi) represents the weight of feature Xi, and m represents the number of samples; finally, the system selects the features with higher weights as the input feature set for model training.

[0094] The model training module uses the support vector machine algorithm to construct a fault prediction model; then it finds the optimal classification hyperplane through the training data, that is, solves the optimal parameter β so that the distance from each support vector to the separation hyperplane is not less than 1 and there is no misjudgment; during the training process, the system uses the cross-validation method to optimize the model parameters. After training, the model is used for the classification prediction of new data to identify whether there is a fault in the power equipment;

[0095] Among them, the kernel function uses the Gaussian kernel, and the calculation formula is:

[0096]

[0097] Among them, K(x_i,x_j) represents the Gaussian kernel function, x_i and x_j represent samples, and σ represents the kernel function parameter; through this formula, the system maps the input data to a high-dimensional space, finds the optimal classification hyperplane, and realizes fault classification.

[0098] The fault detection module uses the modulus maximum detection algorithm based on wavelet transform for fault detection, specifically including:

[0099] S1. Extraction of wavelet coefficient modulus maximum:

[0100] Calculation formula: M_{j,k} = \max(|W_{j,k}|)

[0101] Among them:

[0102] M_{j,k} is the wavelet coefficient modulus maximum at scale j and position k.

[0103] W_{j,k} is the wavelet coefficient, representing the signal feature at scale j and position k.

[0104] S2. Modulus maximum screening, removing the modulus maxima caused by noise and other non-faults, and retaining the possible fault features. By setting a threshold, the modulus maxima exceeding the threshold are screened out.

[0105] The calculation formula is:

[0106] Parameter definition:

[0107] T_{j} is the threshold at scale j.

[0108] σj is the estimated standard deviation of noise at scale j.

[0109] N is the length of the signal.

[0110] log is the natural logarithm.

[0111] The screening condition is M_{j,k}>T_{j}

[0112] S3. Fault feature recognition, identifying whether the screened modulus maxima belong to fault features. By analyzing the features such as the distribution, amplitude, and duration of the modulus maxima, it is judged whether the conditions of fault features are met.

[0113] [A_{threshold}) represents the set fault amplitude threshold.

[0114] D_{threshold} represents the set fault duration threshold.

[0115] Judgment condition: M_{j,k}>A_{threshold};

[0116] And D_{j,k}>D_{threshold};

[0117] S4. Fault detection decision, making a fault detection decision according to the identified fault features. If there are modulus maxima that meet the fault feature conditions, it is judged as a fault; otherwise, it is judged as no fault.

[0118] Decision rule: If The fault feature conditions are met, then the fault detection result = fault;

[0119] Otherwise, the fault detection result = no fault

[0120] The fault detection sub-module realizes the detection of electrical faults through steps such as extracting the modulus maxima of wavelet coefficients, screening the modulus maxima, identifying fault features, and making fault detection decisions. The calculation formula and parameter definition are adjusted and optimized according to specific applications and signal characteristics. This algorithm utilizes the multi-scale analysis ability of wavelet transform to effectively detect fault features in electrical signals and provides accurate information for subsequent fault location.

[0121] The fault location module is responsible for determining the specific fault location based on the fault feature information provided by the fault detection sub-module using the modulus maximum location algorithm based on wavelet transform. The specific steps are as follows:

[0122] S1. Fault feature point determination

[0123] Purpose: To determine the modulus maximum points screened by the fault detection sub-module as fault feature points.

[0124] Input: The screened modulus maximum M_{j,k} and its corresponding position k and scale j.

[0125] S2. Fault propagation speed estimation. Estimate the fault propagation speed in the electrical system for calculating the fault location. According to the physical parameters and configuration of the electrical system, estimate the fault propagation speed. The calculation formula is:

[0126]

[0127] Where:

[0128] v is the fault propagation speed.

[0129] L' is the equivalent inductance of the electrical system.

[0130] C' is the equivalent capacitance of the electrical system.

[0131] S3. Fault location calculation. Calculate the fault location based on the fault feature points and the fault propagation speed. Utilize the time-frequency localization characteristic of wavelet transform and combine with the fault propagation speed to calculate the fault location. The calculation formula is: d = v·Δt

[0132] Where d is the fault location, that is, the distance from the fault point to the detection point.

[0133] Δt is the time delay of the fault feature point, that is, the difference between the time corresponding to the modulus maximum point and the fault occurrence time.

[0134] If there are multiple detection points, fuse the location results of multiple points to improve the location accuracy. Adopt weighted average or other fusion algorithms to combine the location results of multiple detection points.

[0135] The weighted average calculation formula is:

[0136] Where:

[0137] dfinal is the finally determined fault location.

[0138] d_i is the fault location calculated by the i-th detection point.

[0139] $w_i$ is the weight of the $i$-th detection point, which can be set according to factors such as signal quality and the importance of the detection point.

[0140] The specific implementation of the visualization display module involves the following aspects: developing a Web-based monitoring platform, using front-end technologies such as HTML5, CSS3, and JavaScript, and combining chart libraries (such as Highcharts or ECharts) to create dynamic charts and graphs. Users can access this platform through a browser to view the real-time data monitoring screen, including line charts showing the change in power generation, bar charts showing the comparison of power generation efficiency of different devices, pie charts showing the energy distribution, etc. In addition, the platform also integrates GIS functions, which can mark the geographical locations of renewable energy devices on the map, and view detailed information by clicking on the device icons, realizing the geographical spatial visualization of data.

[0141] The detailed functions of the system management module include: providing a background management interface that allows system administrators to create, delete, and assign permissions for user accounts. Permission control ensures that users at different levels can access corresponding levels of data and functions. The log management function records all user operations and system events, facilitating tracking and auditing. The fault alarm system automatically detects system anomalies through preset rules and notifies the administrator through multiple channels (such as email, SMS, audible and visual alarms, etc.). The system backup function is executed automatically at regular intervals, and the backup content includes user data, system configurations, and key business data to support the rapid recovery of the system in case of failures.

[0142] In the present invention, the feature selection module uses the ReliefF algorithm to screen out features that have an important impact on power system fault prediction and diagnosis from the preprocessed data, significantly improving the efficiency and accuracy of data analysis. The ReliefF algorithm can take into account the interactions between features and screen out those characteristics that can stably reflect the state of the power system under different operating conditions, thereby reducing the data dimension, lowering the computational complexity, and also reducing the risk of overfitting. This step ensures that the most relevant information is used in the subsequent model training and fault detection processes, improving the performance of the entire system.

[0143] In the present invention, the model training module constructs a power system fault prediction model using the support vector machine algorithm. This algorithm has advantages in dealing with non-linear problems and classification problems and can effectively identify the boundary between normal operations and potential fault states. The model obtained through training can predict potential faults in advance, thereby realizing preventive maintenance, reducing losses caused by sudden faults, and improving the reliability and safety of the power system. The fault detection module monitors the real-time power system data using the trained model. Once it detects that the data pattern matches the fault characteristics, it can issue an alarm in a timely manner. This real-time monitoring ability enables the operation and maintenance personnel to respond quickly, take corresponding measures, avoid the expansion of faults, and ensure the stable operation of the power system. The fault location module determines the specific location of the fault through a weighted average calculation formula. By assigning weights to different detection points in this step, the system can more accurately locate the fault source, reduce misdiagnosis and misoperations, shorten the fault handling time, and reduce the maintenance cost.

[0144] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0145] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A power system data monitoring and management system for renewable energy access, characterized by: The system includes: a renewable energy equipment data acquisition module, a power grid data acquisition module, an environmental data acquisition module, a data preprocessing module, a data storage module, a data analysis module, a visualization display module and a system management module; The data analysis module is internally provided with a feature selection module, a model training module, a fault detection module and a fault location module; The renewable energy equipment data acquisition module, the power grid data acquisition module and the environmental data acquisition module serve as the front-end perception layer of the system, and are responsible for real-time acquisition of various relevant data and transmission of these data to the data processing module; The data processing module cleans, converts and normalizes the received raw data to ensure the quality and consistency of the data, and then sends the processed data to the data storage module; The data storage module includes a real-time database and a historical database, which are responsible for data storage, backup and recovery, and provide data support for the data analysis module; The data analysis module conducts in-depth analysis of the stored data, including status monitoring, fault diagnosis, predictive analysis and optimization suggestions, and the analysis results will be fed back to the visualization display module and the alarm management module; The visualization module intuitively displays real-time data, historical data and analysis results in the form of charts and curves, providing users with convenient data viewing and analysis methods; The system management module runs through the entire system and is responsible for user management, log management and system configuration to ensure the security, stability and configurability of the system.

2. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The sensor network of the renewable energy equipment data acquisition module is deployed on a photovoltaic panel array, each photovoltaic panel is equipped with a micro inverter, and the inverter has built-in current sensors and voltage sensors for real-time monitoring of output current and voltage; The power grid data acquisition module is equipped with high-precision smart meters and telemetry terminal units at various key nodes of the power grid, including substations and feeder outlets; the smart meters are responsible for real-time measurement of the current, voltage, active power and reactive power parameters of the power grid, and output them through digital signals; the RTU is responsible for periodically reading data from the smart meters and transmitting the data to the monitoring center through the optical fiber communication network; the RTU also has an event recording function, which can record abnormal events of the power grid and send the event information together with the timestamp to facilitate real-time monitoring and analysis of the power grid status; The detailed configuration of the environmental data acquisition module includes: installing multiple environmental monitoring stations at different locations of the renewable energy power plant, each monitoring station includes an anemometer, a wind vane, a thermometer, a hygrometer and a barometer sensor; these sensors are connected to the data collector via wired or wireless means, and the data collector is responsible for regularly reading the sensor data and performing preliminary data verification; the collected environmental data is transmitted to the data storage server via the LoRa or Wi-Fi network, and the environmental data analysis software on the server side will evaluate the impact of environmental factors on the efficiency of renewable energy power generation based on these data, and provide a basis for adjusting the power generation strategy.

3. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The data preprocessing module uses a moving average filtering algorithm to reduce the impact of random noise by calculating the average value of data within a certain time window. The calculation formula of the moving average filtering is as follows: Where: y[n] is the filtered output value at the nth time point; x[n] is the original data value at the nth time point; N is the size of the moving average window, that is, the number of adjacent data points considered; i is the index within the window, from 0 to N-1.

4. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The data storage module uses the MySQL relational database management system to store and manage the collected data; multiple tables are designed in the database, each table corresponds to different types of data, including renewable energy equipment operation data table, power grid parameter data table and environmental monitoring data table; each table contains fields including timestamp, device ID, and measurement value; The data storage module optimizes data writing and query efficiency by writing stored procedures and triggers. At the same time, it adopts regular full backup and real-time incremental backup strategies to ensure data security.

5. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The feature selection module selects features that have an important impact on fault prediction and diagnosis from the preprocessed data; the feature selection and optimization module uses the ReliefF algorithm. First, for each feature Xi, the system traverses all samples and finds the nearest neighbor samples of the same type and samples of different types for each sample; then, the feature differences diff(Xi,j,H) and diff(Xi,j,M) are calculated, where diff(Xi,j,H) represents the difference between samples of the same type of feature Xi, and diff(Xi,j,M) represents the difference between samples of different types; then, according to the formula: W(Xi)=∑ j＝1 m (diff(Xi,j,H)-diff(Xi,j,M)) calculates the feature weight, where W(Xi) represents the weight of feature Xi and m represents the number of samples. Finally, the system selects features with higher weights as the input feature set for model training.

6. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The model training module uses the support vector machine algorithm to build a fault prediction model; then the optimal classification hyperplane is found through training data, that is, the optimal parameter β is solved so that the distance from each support vector to the separation hyperplane is not less than 1, and there is no misjudgment; during the training process, the system uses a cross-validation method to optimize the model parameters. After the training is completed, the model is used for classification prediction of new data to identify whether there is a fault in the power equipment; The kernel function uses Gaussian kernel, and the calculation formula is: Among them, K(x_i,x_j) represents the Gaussian kernel function, x_i and x_j represent samples, and σ represents the kernel function parameter; through this formula, the system maps the input data to a high-dimensional space, finds the optimal classification hyperplane, and realizes fault classification.

7. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The fault detection module uses a modulus maximum detection algorithm based on wavelet transform to perform fault detection, specifically including: S1. Wavelet coefficient modulus maximum extraction: Calculation formula: M_{j,k}=\max(|W_{j,k}|) in: M_{j,k} is the maximum modulus of the wavelet coefficients at scale j and position k; W_{j,k} is the wavelet coefficient, which represents the signal characteristics at scale j and position k; S2. Modulus maximum value screening, remove the modulus maximum values ​​caused by noise and other non-faults, and retain possible fault characteristics; by setting a threshold, filter out the modulus maximum values ​​that exceed the threshold; The calculation formula is: Parameter definition: T_{j} is the threshold of scale j; σj is the estimate of the noise standard deviation at scale j; N is the length of the signal; log is the natural logarithm; The screening condition is M_{j,k}>T_{j} S3. Fault feature identification, identifying whether the filtered modulus maximum value belongs to the fault feature; by analyzing the distribution, amplitude and duration characteristics of the modulus maximum value, judging whether the conditions of the fault feature are met; [A_{threshold}) represents the set fault amplitude threshold; D_{threshold} represents the set fault duration threshold; Judgment condition: M_{j,k}>A_{threshold}; And D_{j,k}>D_{threshold}; S4. Fault detection decision: make a fault detection decision based on the identified fault characteristics. If there is a modulus maximum value that satisfies the fault characteristics, it is judged as a fault; otherwise, it is judged as no fault; Decision rule: If If the fault characteristic conditions are met, the fault detection result = fault; Otherwise, fault detection result = no fault The fault detection submodule detects electrical faults by extracting the modulus maxima of wavelet coefficients, screening the modulus maxima, identifying fault features and making fault detection decision steps; the calculation formula and parameter definitions are adjusted and optimized according to the specific application and signal characteristics; the algorithm uses the multi-scale analysis capability of wavelet transform to effectively detect fault features in electrical signals and provide accurate information for subsequent fault location.

8. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The fault location module uses a modulus maximum location algorithm based on wavelet transform to determine the specific location of the fault according to the fault feature information provided by the fault detection submodule. The specific steps include: S1. Determination of fault characteristic points Purpose: To determine the modulus maximum point selected by the fault detection submodule as the fault feature point; Input: the filtered modulus maximum M_{j,k} and its corresponding position k and scale j; S2. Fault propagation speed estimation, estimates the fault propagation speed in the electrical system, which is used to calculate the fault location. According to the physical parameters and configuration of the electrical system, the fault propagation speed is estimated. The calculation formula is: in: v is the fault propagation speed; L' is the equivalent inductance of the electrical system; C' is the equivalent capacitance of the electrical system; S3. Fault location calculation: Calculate the fault location based on the fault feature points and the fault propagation speed; use the time-frequency location characteristics of wavelet transform and the fault propagation speed to calculate the fault location; the calculation formula is: d = v·Δt Where d is the fault location, that is, the distance between the fault point and the detection point; Δt is the time delay of the fault characteristic point, that is, the difference between the time corresponding to the modulus maximum point and the time when the fault occurs; If there are multiple detection points, fuse the positioning results of multiple points to improve positioning accuracy; use weighted average or other fusion algorithms to combine the positioning results of multiple detection points; The weighted average calculation formula is: in: dfinal is the final determined fault location; d_i is the fault location calculated at the i-th detection point; w_i is the weight of the i-th detection point, which can be set according to the signal quality and the importance of the detection point.

9. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The visualization module develops a web-based monitoring platform, using HTML5, CSS3 and JavaScript front-end technologies, combined with a chart library to create dynamic charts and graphs; Users can access the platform through a browser to view real-time data monitoring screens, including line graphs showing changes in power generation, bar graphs showing power generation efficiency comparisons of different devices, and pie charts showing energy distribution.

10. The power system data monitoring and management system for renewable energy access according to claim 1, characterized in that: The detailed functions of the system management module include: providing a background management interface that allows system administrators to create, delete and assign permissions to user accounts; permission control ensures that users of different levels can access data and functions of corresponding levels; the log management function records all user operations and system events for easy tracking and auditing; the fault alarm system automatically detects system anomalies through preset rules and notifies administrators through multiple channels; the system backup function is automatically executed regularly, and the backup content includes user data, system configuration and key business data to support rapid recovery of the system when a fault occurs.

Images