Wind power plant cluster early warning system based on 5G and big data model

Abstract

The invention discloses a wind power plant cluster early warning system based on 5G and a big data model, which relates to the technical field of wind power management and comprises a data acquisition module, a 5G communication module, a data processing and analysis module, an early warning decision support module and a data interaction module. The data acquisition module is installed on each wind turbine generator set of a wind power plant and is used for acquiring operation data of the generator sets in real time, and the data comprises temperature, power and vibration; and the 5G communication module transmits the acquired data to the data processing and analyzing module at a high speed through a 5G network. The real-time performance and the accuracy of the early warning system are improved by realizing rapid transmission and processing of the data, the hidden information in the data is mined through a big data analysis technology, a scientific basis is provided for early warning decision making, whether the whole current cluster and the running state of each unit are abnormal or not can be rapidly judged, and the early warning efficiency is improved. Potential problems are found in advance, corresponding maintenance measures are taken, and it is ensured that the cluster can operate stably and efficiently.

Description

Technical Field

[0001] The present invention relates to the field of wind power management technology, and specifically to a wind farm fleet early warning system based on 5G and big data models. Background Art

[0002] There are many types of wind turbines in wind farms, the most common of which are horizontal-axis wind turbines and vertical-axis wind turbines. The rotor axis of a horizontal-axis wind turbine is parallel to the wind direction, and it can be further divided into different power levels and blade numbers. For example, the common three-blade, megawatt-level horizontal-axis wind turbine has a high wind energy conversion efficiency and is the most widely used type of wind farm. The rotor axis of a vertical-axis wind turbine is perpendicular to the wind direction, such as the Darrieus vertical-axis wind turbine. Although it accounts for a relatively small proportion of the overall proportion of wind farms, it also has unique advantages in some specific application scenarios. With the development of global wind energy resources With the continuous development and utilization of energy resources, the number and scale of wind farms have grown rapidly. However, the operating environment of wind turbines is complex and changeable, and is easily affected by weather and geographical factors, resulting in frequent equipment failures. At present, although wind farms have some monitoring and early warning methods, there are many shortcomings. The traditional monitoring system has slow data transmission rate and poor real-time performance. In addition, there is a lack of effective big data analysis models for in-depth mining and comprehensive analysis of massive operating data. It can often only judge a single parameter or simple indicator. The early warning results are biased and difficult to obtain a comprehensive analysis. In this regard, we propose a wind farm fleet early warning system based on 5G and big data models. Summary of the invention

[0003] In order to solve the above technical problems, a wind farm fleet early warning system based on 5G and big data models is provided. This technical solution solves the above-mentioned problems of high data transmission delay and lack of effective deep data mining and analysis.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is: a wind farm fleet early warning system based on 5G and big data model, including: a data acquisition module, a 5G communication module, a data processing and analysis module, an early warning decision support module and a data interaction module; The data acquisition module is installed on each wind turbine in the wind farm to collect the operating data of the turbine in real time, including temperature, power and vibration. The 5G communication module transmits the collected data to the data processing and analysis module at high speed through the 5G network; The data processing and analysis module is responsible for processing and analyzing the collected data. The data processing and analysis module includes data cleaning, feature extraction and model training. It obtains historical collected data, builds historical data sets, uses historical data sets to train models, learns the intrinsic relationship between various parameters under different operating conditions and fault characteristic patterns, inputs real-time collected data into the model, and determines whether there are abnormalities in the operating status of the current fleet as a whole and each unit. The early warning decision support module generates early warning information and gives corresponding maintenance suggestions based on the analysis results; The data interaction module is used for operation and maintenance personnel to view warning information and system status and conduct human-computer interaction.

[0005] Preferably, the data acquisition module is installed at various key locations of the wind turbine, including the gearbox, the stator and rotor of the generator, the heat sink and the control cabinet. The data acquisition module includes a temperature sensor, a power monitoring device and a vibration sensor. The power monitoring device is closely connected with the power generation system and the power transmission line of the wind turbine, and the power generation power changes caused by wind speed changes and performance fluctuations of the mechanical components of the unit are captured in real time. The vibration sensor is used to detect the vibration of the blades under wind load and to centrally organize and store the collected data.

[0006] Preferably, after the 5G communication module obtains the data collected and stored by the data acquisition module, it preprocesses the acquired data, and the preprocessing includes data integrity verification, error correction and format conversion operations. The preprocessed data is sent to the 5G network environment for transmission in the form of wireless signals, and the data is uploaded to the data processing and analysis module.

[0007] Preferably, data cleaning in the data processing and analysis module is based on mean interpolation, median interpolation and mode difference interpolation methods to clean the data, and the feature extraction method is based on statistical analysis methods to extract features; the model adopted is a support vector machine model, and a basic model structure is constructed, wherein the historical collected data includes normal operating data and fault case data, and the collected data is constructed into a data set, and the relationship between each parameter under different operating conditions is learned by the model, and the historical data set is divided into a training set, a verification set and a test set. In the training set, for data samples marked as normal operating conditions, the model learns that the temperature parameters are stable within a reasonable range under normal circumstances. For data samples marked as fault conditions, the model analyzes the specific change characteristics of each parameter under each fault condition and the combination rules between them; the test set is input into the model for training, and the test results are output. The output results of the model are verified by the verification set, and the parameters of the model are optimized through the verification results to complete the training of the model.

[0008] Preferably, the support vector machine model determines that the operating state is abnormal by: Where w is the weight vector, b is the bias term, x is the unit feature vector collected in real time, including temperature, power and vibration parameters, f(x) is the output result, and T is the inner product operation; the collected data is input into the model, and the predicted result is obtained after output to judge the operating status of the wind farm group. When the output result of f(x) is +1, it indicates that the wind farm group is in normal operating state. When the output result of f(x) is -1, it indicates that the wind farm group is in abnormal condition.

[0009] Preferably, the determination of the weight w is obtained by solving the optimization problem corresponding to the support vector machine model, and the solution form is: where x i With y i are the category labels corresponding to the feature vectors of the i-th sample in the training data set, and the goal is to minimize the bi-norm of the weight vector w. , satisfying the constraints , where i=1, 2, ..., n, and the weight w is solved by the above formula to obtain the weight value.

[0010] Preferably, the early warning decision support module analyzes the abnormal conditions of the unit's operating status based on the results output by the data processing and analysis module. If the analysis results show that the vibration amplitude of a wind turbine unit is far beyond the normal range and the power output shows obvious unstable fluctuations, accompanied by a continuous increase in the temperature of key parts, the early warning decision support module will quickly identify the potential fault hazards corresponding to this series of abnormal manifestations, and then generate accurate and targeted early warning information based on preset rules and algorithms.

[0011] Preferably, when an abnormality is detected and an early warning message is sent, the suggestions given include immediately inspecting the gearbox to check for poor lubrication and gear wear; conducting a comprehensive and detailed inspection of the blades, focusing on checking whether there are cracks and visible damage due to wear on the blade surface; replacing the corresponding lubricating oil. If the wear is serious, it is necessary to prepare suitable gear spare parts for replacement, and the gearbox needs to be re-debugged after replacement; the early warning decision support module is based on historical operation and maintenance experience and past failure case data, combined with the specific abnormal conditions and potential failure characteristics reflected by the current analysis results, to provide practical maintenance suggestions for operation and maintenance personnel.

[0012] Preferably, the data interaction module receives the warning message transmitted from the warning decision support module in real time, and presents it to the operation and maintenance personnel in an eye-catching manner, highlighting it with warning boxes of different colors, and the color distinction is determined according to the urgency of the warning.

[0013] Preferably, the data interaction module presents the generated warning information and system status in the form of a report, and the report is generated in a text form and a chart form combined with each other.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention utilizes the high rate and low latency characteristics of the 5G network to achieve rapid transmission and processing of data, improve the real-time and accuracy of the early warning system, and through big data analysis technology, mine the implicit information in the data to provide a scientific basis for early warning decisions. The real-time collected data is input into the trained model, which can quickly determine whether there are any abnormalities in the operating status of the current fleet as a whole and each unit, discover potential problems in advance and take corresponding maintenance measures to ensure that the fleet can operate stably and efficiently. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a framework diagram of the wind farm fleet early warning system of the present invention. DETAILED DESCRIPTION

[0016] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0017] Reference Figure 1 As shown, a wind farm fleet early warning system based on 5G and big data model includes: Data acquisition module, 5G communication module, data processing and analysis module, early warning decision support module and data interaction module; The data acquisition module is installed on each wind turbine in the wind farm to collect the operating data of the turbine in real time, including temperature, power and vibration. The 5G communication module transmits the collected data to the data processing and analysis module at high speed through the 5G network; The data processing and analysis module is responsible for processing and analyzing the collected data. The data processing and analysis module includes data cleaning, feature extraction and model training. It obtains historical collected data, builds historical data sets, uses historical data sets to train models, learns the intrinsic relationship between various parameters under different operating conditions and fault characteristic patterns, inputs real-time collected data into the model, and determines whether there are abnormalities in the operating status of the current fleet as a whole and each unit. The early warning decision support module generates early warning information and gives corresponding maintenance suggestions based on the analysis results; The data interaction module is used for operation and maintenance personnel to view warning information and system status and conduct human-computer interaction.

[0018] This application highlights the essential characteristics of unit operation through feature extraction, focuses on core information, and helps operation and maintenance personnel and analysis models to grasp the key status of the unit more efficiently. By using historical data sets to train the model, the model learns the complex internal relationship between various parameters under different operating conditions and various fault characteristic modes, so that when faced with real-time collected data, the model can accurately determine whether there are abnormalities in the operating status of the current fleet as a whole and each unit; Based on the analysis results output by the data processing and analysis module, it is possible to quickly identify abnormal conditions in the operation of the unit and generate corresponding early warning information. Once a potential fault risk is detected, such as the vibration amplitude of a unit continuously exceeding the normal range and the power fluctuating unstably, the early warning decision support module will immediately issue an early warning to remind the operation and maintenance personnel to pay attention in time, thus shortening the fault discovery time to the greatest extent, avoiding further deterioration of the fault, and reducing the loss of power generation and the increase in maintenance costs caused by downtime for maintenance; The warning information and system status are displayed to the operation and maintenance personnel in a clear and friendly interface, so that they can understand the overall operation status of the wind farm fleet and the specific status of each unit at a glance.

[0019] The data acquisition module is installed in various key parts of the wind turbine, including the gearbox, stator and rotor of the generator, heat sink and control cabinet. The data acquisition module includes temperature sensors, power monitoring devices and vibration sensors. The power monitoring device is closely connected with the power generation system and power transmission line of the wind turbine, and can capture the changes in power generation caused by changes in wind speed and performance fluctuations of the mechanical components of the unit in real time; the vibration sensor is used to detect the vibration of the blades under the action of wind load, and the collected data is centrally organized and stored.

[0020] This application installs different types of sensors at various key locations to collect multi-dimensional data including temperature, power, and vibration, providing a rich information basis for in-depth fault diagnosis. Changes in a single parameter are sometimes difficult to accurately determine the cause of a fault, but the combined change trends and interrelationships of multiple parameters can more accurately analyze the root cause of the fault.

[0021] After the 5G communication module obtains the data collected and stored by the data acquisition module, it preprocesses the acquired data. The preprocessing includes data integrity verification, error correction and format conversion operations. The preprocessed data is sent to the 5G network environment for transmission in the form of wireless signals, and the data is uploaded to the data processing and analysis module.

[0022] In this application, integrity check determines the number of data records that should be generated in a specific time period or under specific conditions based on the data generation logic and expectations, and compares the actual number of records collected with the expected number to determine whether the data is complete; Error correction is done by building a regression model, training the model with normal data, and then using the model to predict the suspected abnormal data points. If the difference between the predicted value and the actual value is too large, the point is determined to be an outlier, and the predicted value can be used to replace the actual outlier for error correction; The format conversion operation converts the data format by changing the data type and precision.

[0023] Data cleaning in the data processing and analysis module is based on mean interpolation, median interpolation and mode difference interpolation methods to clean the data, and the feature extraction method is based on statistical analysis methods to extract features; the model used is the support vector machine model, and a basic model structure is constructed. The historical collected data includes normal operating data and fault case data. The collected data is constructed into a data set, and the relationship between various parameters under different operating conditions is learned through the model. The historical data set is divided into a training set, a validation set and a test set. In the training set, the data samples marked as normal operating status, the model learns that the temperature parameters are stable within a reasonable range under normal circumstances. For data samples marked as fault status, the model analyzes the specific change characteristics of each parameter under each fault state and the combination rules between them; the test set is input into the model for training, and the test results are output. The output results of the model are verified by the validation set. The parameters of the model are optimized through the verification results to complete the training of the model.

[0024] Data cleaning and feature extraction in this application are both existing technologies and will not be elaborated on here. The historical collected data includes normal operating data and fault case data, which enables the model to be exposed to the performance of the wind farm fleet under different conditions. By learning the relationship between various parameters under normal operating conditions, the model can understand how parameters such as temperature, power, and vibration cooperate with each other and are maintained within a reasonable range when the unit is working stably. Under normal circumstances, the temperature parameter is stable within a reasonable range, the power will change according to the expected law with external factors such as wind speed, and the vibration amplitude will remain at a low level without abnormal fluctuations. After the model learns these normal parameter association patterns, it can use this as a reference to determine whether subsequent real-time data is abnormal.

[0025] The support vector machine model determines the abnormal running status expression as follows: Where w is the weight vector, b is the bias term, x is the unit feature vector collected in real time, including temperature, power and vibration parameters, f(x) is the output result, and T is the inner product operation; the collected data is input into the model, and the predicted result is obtained after output to judge the operating status of the wind farm group. When the output result of f(x) is +1, it indicates that the wind farm group is in normal operating state. When the output result of f(x) is -1, it indicates that the wind farm group is in abnormal condition.

[0026] The judgment of the w weight is obtained by solving the optimization problem corresponding to the support vector machine model, and the solution form is: where x i With y i are the category labels corresponding to the feature vectors of the i-th sample in the training data set, and the goal is to minimize the bi-norm of the weight vector w. , satisfying the constraints , where i=1, 2, ..., n, and the weight w is solved by the above formula to obtain the weight value.

[0027] The early warning decision support module analyzes the abnormal conditions of the unit's operating status based on the results output by the data processing and analysis module. If the analysis results show that the vibration amplitude of a wind turbine unit is far beyond the normal range and the power output shows obvious unstable fluctuations, accompanied by a continuous increase in the temperature of key parts, the early warning decision support module will quickly identify the fault hazards corresponding to this series of abnormal manifestations, and then generate accurate and targeted early warning information based on preset rules and algorithms.

[0028] In the present application, the wind farm fleet is in continuous operation, with numerous units and complex operating conditions. Manual real-time monitoring and fault detection often involve certain difficulties and lags. The early warning decision support module can automatically and quickly analyze the operating status of the units based on the results output by the data processing and analysis module. Once abnormal manifestations such as vibration amplitude far exceeding the normal range, unstable power output fluctuations, and continuous temperature increases in key parts are found, these changes can be captured immediately. The working mechanism of the early warning decision support module helps wind farms to shift from the traditional post-maintenance mode to preventive maintenance. By continuously monitoring the operating status of the units, potential fault hazards can be discovered in advance and early warnings can be issued, allowing operation and maintenance personnel to intervene and handle the fault before it actually occurs or when it is in the early stages.

[0029] When an abnormality is detected and an early warning message is sent, the suggestions given include immediately inspecting the gearbox to check for poor lubrication and gear wear; conducting a comprehensive and detailed inspection of the blades, focusing on checking whether there are cracks and visible damage on the blade surface; replacing the corresponding lubricating oil. If the wear is serious, prepare suitable gear spare parts for replacement, and re-debug the gearbox after replacement; the early warning decision support module is based on historical operation and maintenance experience and past failure case data, combined with the specific abnormal conditions and potential failure characteristics reflected in the current analysis results, to provide operation and maintenance personnel with practical maintenance suggestions.

[0030] The maintenance suggestions given by this application warning decision support module are very specific and targeted. They clearly point out the components that need to be checked. The suggestions generated based on past experience and data analysis are all effective operating steps that have been verified in practice or summarized based on the laws of failure occurrence. This prevents operation and maintenance personnel from making unnecessary attempts or repeated inspections, and each operation is more likely to move in the direction of solving the problem.

[0031] The data interaction module receives the warning messages transmitted by the warning decision support module in real time, and presents them to the operation and maintenance personnel in a striking manner, highlighting them with warning boxes of different colors. The color distinction is based on the urgency of the warning.

[0032] In the complex and numerous operation and maintenance working environment of a wind farm, the operation and maintenance personnel need to pay attention to multiple units and numerous system operating parameters and status information at the same time. The data interaction module presents the warning message in a striking manner, which can quickly stand out from the numerous information and catch the attention of the operation and maintenance personnel at the first time; the warning message is presented in a visual form of a color-coded warning box, which conforms to the characteristics of human visual cognition and makes it easier for the operation and maintenance personnel to understand and remember the relevant information.

[0033] The data interaction module presents the generated warning information and system status in the form of a report, and the report is generated in a combination of text and chart forms.

[0034] The text of this application can provide an accurate and detailed description of the warning information and system status. For the warning information, it can provide a detailed description of the specific number of the unit where the abnormality occurred, the time when the abnormality occurred, the specific abnormal manifestations, and the possible corresponding fault hazards, and attach preliminary recommended measures based on system analysis. When describing the system status, it can also use text to clearly list the operating parameters of each unit, the current working mode, and environmental related information, so as to facilitate the operation and maintenance personnel to fully and deeply understand the specific situation and provide rich details for subsequent analysis and decision-making.

[0035] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention, and these changes and improvements fall within the scope of the present invention claimed.

Claims

1. A wind farm fleet early warning system based on 5G and big data model, characterized in that: include: Data acquisition module, 5G communication module, data processing and analysis module, early warning decision support module and data interaction module; The data acquisition module is installed on each wind turbine in the wind farm to collect the operating data of the turbine in real time, including temperature, power and vibration. The 5G communication module transmits the collected data to the data processing and analysis module at high speed through the 5G network; The data processing and analysis module is responsible for processing and analyzing the collected data. The data processing and analysis module includes data cleaning, feature extraction and model training. It obtains historical collected data, builds historical data sets, uses historical data sets to train models, learns the intrinsic relationship between various parameters under different operating conditions and fault characteristic patterns, inputs real-time collected data into the model, and determines whether there are abnormalities in the operating status of the current fleet as a whole and each unit. The early warning decision support module generates early warning information and gives corresponding maintenance suggestions based on the analysis results; The data interaction module is used for operation and maintenance personnel to view warning information and system status and conduct human-computer interaction.

2. According to claim 1, a wind farm fleet early warning system based on 5G and big data model is characterized in that: The data acquisition module is installed in various key parts of the wind turbine, including the gearbox, the stator and rotor of the generator, the heat sink and the control cabinet. The data acquisition module includes temperature sensors, power monitoring devices and vibration sensors. The power monitoring device is closely connected with the power generation system and power transmission line of the wind turbine, and can capture the changes in power generation caused by changes in wind speed and performance fluctuations of mechanical components of the unit in real time. The vibration sensor is used to detect the vibration of the blades under wind load and to centrally organize and store the collected data.

3. According to claim 1, a wind farm fleet early warning system based on 5G and big data model is characterized in that: After the 5G communication module obtains the data collected and stored by the data acquisition module, it preprocesses the acquired data. The preprocessing includes data integrity verification, error correction and format conversion operations. The preprocessed data is sent to the 5G network environment for transmission in the form of wireless signals, and the data is uploaded to the data processing and analysis module.

4. According to claim 1, a wind farm fleet early warning system based on 5G and big data model is characterized in that: Data cleaning in the data processing and analysis module is based on mean interpolation, median interpolation and mode difference interpolation methods to clean the data, and the feature extraction method is based on statistical analysis methods to extract features; the model used is the support vector machine model, and a basic model structure is constructed. The historical collected data includes normal operating data and fault case data. The collected data is constructed into a data set, and the relationship between various parameters under different operating conditions is learned through the model. The historical data set is divided into a training set, a validation set and a test set. In the training set, the data samples marked as normal operating status, the model learns that the temperature parameters are stable within a reasonable range under normal circumstances. For data samples marked as fault status, the model analyzes the specific change characteristics of each parameter under each fault state and the combination rules between them; the test set is input into the model for training, and the test results are output. The output results of the model are verified by the validation set. The parameters of the model are optimized through the verification results to complete the training of the model.

5. The wind farm fleet early warning system based on 5G and big data model according to claim 1 is characterized in that: The support vector machine model determines the abnormal running status expression as follows: Where w is the weight vector, b is the bias term, x is the unit feature vector collected in real time, including temperature, power and vibration parameters, f(x) is the output result, and T is the inner product operation; the collected data is input into the model, and the predicted result is obtained after output to judge the operating status of the wind farm group. When the output result of f(x) is +1, it indicates that the wind farm group is in normal operating state. When the output result of f(x) is -1, it indicates that the wind farm group is in abnormal condition.

6. The wind farm fleet early warning system based on 5G and big data model according to claim 5 is characterized in that: The judgment of the w weight is obtained by solving the optimization problem corresponding to the support vector machine model, and the solution form is: where x i With y i are the category labels corresponding to the feature vectors of the i-th sample in the training data set, and the goal is to minimize the bi-norm of the weight vector w. , satisfying the constraints , where i=1, 2, ..., n, and the weight w is solved by the above formula to obtain the weight value.

7. The wind farm fleet early warning system based on 5G and big data model according to claim 1 is characterized in that: The early warning decision support module analyzes the abnormal conditions of the unit's operating status based on the results output by the data processing and analysis module. If the analysis results show that the vibration amplitude of a wind turbine unit is far beyond the normal range and the power output shows obvious unstable fluctuations, accompanied by a continuous increase in the temperature of key parts, the early warning decision support module will quickly identify the fault hazards corresponding to this series of abnormal manifestations, and then generate accurate and targeted early warning information based on preset rules and algorithms.

8. The wind farm fleet early warning system based on 5G and big data model according to claim 7 is characterized in that: When an abnormality is detected and an early warning message is sent, the suggestions given include immediately checking the gearbox to check for poor lubrication and gear wear; conducting a comprehensive and detailed inspection of the blades, focusing on checking whether there are cracks and visible damage on the blade surface; Replace the corresponding lubricating oil. If the wear is serious, prepare suitable gear spare parts for replacement. After replacement, the gearbox needs to be re-debugged; the early warning decision support module is based on historical operation and maintenance experience and past failure case data, combined with the specific abnormal conditions and potential failure characteristics reflected in the current analysis results, to provide practical maintenance suggestions for operation and maintenance personnel.

9. The wind farm fleet early warning system based on 5G and big data model according to claim 1 is characterized in that: The data interaction module receives the warning messages transmitted by the warning decision support module in real time, and presents them to the operation and maintenance personnel in a striking manner, highlighting them with warning boxes of different colors. The color distinction is based on the urgency of the warning.

10. The wind farm fleet early warning system based on 5G and big data model according to claim 1, characterized in that: The data interaction module presents the generated warning information and system status in the form of a report, and the report is generated in a combination of text and chart forms.

Images