Wind power cluster preventive maintenance strategy intelligent generation method and system

Abstract

The invention relates to the field of wind power plants, in particular to a wind power cluster preventive maintenance strategy intelligent generation method and system. The method comprises the following steps: collecting multi-source heterogeneous data of a wind power cluster in real time, including SCADA system operation data, CMS monitoring data and environmental sensor data; performing standardization processing and abnormal value elimination on the collected data, and constructing a wind power equipment operation state database; a wind power equipment health state evaluation model is established based on a deep learning algorithm, the equipment operation state is analyzed in real time, and the potential fault risk is predicted; generating a preventive maintenance strategy by adopting a multi-objective optimization algorithm according to a fault risk prediction result in combination with historical maintenance records of the equipment and operation and maintenance resource scheduling information; and issuing the generated maintenance strategy to a wind power cluster operation and maintenance system to realize intelligent execution of preventive maintenance. The problems of high data fusion difficulty and low processing efficiency in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of wind farms, and in particular to a method and system for intelligently generating preventive maintenance strategies for wind power clusters. Background Art

[0002] With the rapid development of the wind power industry, the scale of wind power clusters continues to expand, and their operation, maintenance and management face numerous challenges. Traditional wind power equipment maintenance strategies rely primarily on regular inspections and post-fault repairs. This strategy is not only inefficient but also difficult to cope with the complex and changing wind power operating environment, easily leading to frequent equipment failures, increased maintenance costs and downtime. In addition, the operating data of wind power clusters is multi-source and heterogeneous, including SCADA system operating data, CMS monitoring data, and environmental sensor data. The real-time and accuracy of this data directly affect the assessment of equipment health status and the generation of maintenance strategies. Existing technologies have shortcomings in data fusion and processing, making it difficult to fully utilize multi-source data for accurate fault prediction and maintenance strategy optimization. At the same time, wind power clusters have limited maintenance resources. How to rationally allocate operation and maintenance resources, improve maintenance efficiency, and reduce maintenance costs is a problem that needs to be solved urgently. Therefore, there is an urgent need for an intelligent generation method and system for preventive maintenance strategies for wind power clusters that can solve the above problems. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for intelligently generating a preventive maintenance strategy for a wind power cluster, so as to solve the problems in the above-mentioned background technology.

[0004] To achieve the above objectives, the following technical solutions are adopted.

[0005] A method for intelligently generating a preventive maintenance strategy for a wind power cluster comprises the following steps: collecting multi-source heterogeneous data of a wind power cluster in real time, including SCADA system operation data, CMS monitoring data, and environmental sensor data; standardizing the collected data and eliminating outliers, and constructing a wind power equipment operation status database; establishing a wind power equipment health status assessment model based on a deep learning algorithm, analyzing the equipment operation status in real time, and predicting potential failure risks; generating a preventive maintenance strategy based on the failure risk prediction results, combined with historical equipment maintenance records and operation and maintenance resource scheduling information, using a multi-objective optimization algorithm; and sending the generated maintenance strategy to the wind power cluster operation and maintenance system to achieve intelligent execution of preventive maintenance.

[0006] Optionally, the standardized processing of multi-source heterogeneous data includes the following steps: time alignment and spatial matching of SCADA system data and CMS monitoring data to eliminate data acquisition delay and spatial deviation; using an adaptive weighted fusion algorithm to associate environmental sensor data with equipment operation data to construct a unified data feature matrix; and using data cleaning technology to remove noise and redundant information to improve data quality.

[0007] Optionally, the construction of the wind power equipment health status assessment model includes: establishing an equipment health status baseline curve based on historical fault data; extracting the dynamic characteristics of real-time operation data through a convolutional neural network, and comparing and analyzing it with the baseline curve; introducing equipment aging factors and operating condition correction coefficients, and dynamically adjusting the model output to reflect the actual health status of the equipment.

[0008] Optionally, the optimization objectives of the multi-objective optimization algorithm include: maximizing the overall operational reliability of the wind power cluster; minimizing maintenance costs and downtime; balancing operation and maintenance resource allocation efficiency and equipment maintenance priority; and determining the optimal maintenance strategy through Pareto frontier analysis.

[0009] Optionally, the generation of the preventive maintenance strategy also includes: dynamically adjusting the maintenance time window according to the wind farm's power generation plan and grid scheduling requirements; prioritizing maintenance tasks during low wind speed or low load periods to reduce power generation losses; and optimizing the execution path of maintenance tasks by combining drone inspections and remote diagnosis results.

[0010] Optionally, the method further includes a maintenance effect feedback step: recording the execution results of each maintenance task and the subsequent operating status of the equipment; comparing and analyzing the actual maintenance effect with the output of the prediction model; and iteratively optimizing the health status assessment model and maintenance strategy generation algorithm based on the analysis results.

[0011] Optionally, the maintenance effect feedback step specifically includes: establishing a maintenance effect evaluation index system, including fault recurrence rate, maintenance time, resource consumption, etc.; determining the weak links of the maintenance strategy through statistical analysis; and dynamically updating model parameters and maintenance strategy library based on feedback data.

[0012] Optionally, the method further includes a collaborative early warning step: integrating the real-time data processing results of the edge computing nodes to generate a global device health status heat map; when risk warnings appear on multiple devices at the same time, triggering a collaborative maintenance mechanism to optimize the resource scheduling plan.

[0013] The intelligent generation system of preventive maintenance strategies for wind power clusters includes: a data acquisition module for acquiring multi-source heterogeneous data of wind power clusters in real time; a data processing module for data standardization and feature extraction; a health assessment module for operating status prediction and fault risk assessment; a strategy generation module for outputting maintenance strategies based on a multi-objective optimization algorithm; an execution control module for issuing strategies and coordinating operation and maintenance resources; and a feedback optimization module for achieving closed-loop optimization of maintenance effects.

[0014] Optionally, the data processing module further includes: a data cleaning unit for removing noise and outliers; a feature fusion unit for constructing a unified data feature matrix; the health assessment module integrates a deep learning model to support dynamic health status scoring and risk level classification.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] The present application provides a method for intelligently generating preventive maintenance strategies for wind power clusters. By real-time collection of multi-source heterogeneous data from wind power clusters, including SCADA system operation data, CMS monitoring data, and environmental sensor data, it solves the problems of high difficulty in data fusion and low processing efficiency in the existing technology. The method performs standardization processing on the collected data and eliminates outliers, constructs a wind power equipment operation status database, and establishes a wind power equipment health status assessment model based on a deep learning algorithm to analyze the equipment operation status in real time and predict potential failure risks. Based on the failure risk prediction results, combined with the equipment's historical maintenance records and operation and maintenance resource scheduling information, a multi-objective optimization algorithm is used to generate a preventive maintenance strategy, and the generated maintenance strategy is sent to the wind power cluster operation and maintenance system to achieve intelligent execution of preventive maintenance. This method not only improves the accuracy and reliability of equipment health status assessment, but also optimizes the allocation of maintenance resources, reducing maintenance costs and downtime.

[0017] This application improves and supplements the standardized processing of multi-source heterogeneous data, the construction of a wind turbine equipment health assessment model, the optimization objectives of a multi-objective optimization algorithm, the generation of preventive maintenance strategies, maintenance effect feedback, and collaborative early warning. These improvements not only enhance the intelligence and practicality of maintenance strategies, but also strengthen the system's dynamic adaptability and closed-loop optimization capabilities, enabling it to better adapt to the complex and changing wind turbine operating environment and providing strong technical support for the efficient operation and maintenance of wind turbine clusters. DETAILED DESCRIPTION

[0018] The present invention will be described in detail below with reference to the embodiments. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0019] The following detailed description is an exemplary description, which is intended to provide further detailed description of the present invention. Unless otherwise indicated, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present invention.

[0020] This invention provides a method and system for intelligently generating preventive maintenance strategies for wind turbine clusters. This method aims to achieve efficient operation and preventive maintenance of wind turbine clusters through real-time data acquisition, deep learning algorithms, multi-objective optimization algorithms, and feedback optimization. The following describes specific implementations of this method and system.

[0021] First, the data acquisition module collects multi-source, heterogeneous data from the wind turbine cluster in real time, including SCADA system operating data, CMS monitoring data, and environmental sensor data. This data comes from a wide range of sources, covering the wind turbine cluster's operating environment and equipment status, and comprehensively reflecting the wind turbine cluster's operational status. SCADA system operating data primarily includes key operating parameters such as wind turbine power, speed, and temperature. This data is collected in real time through the wind turbine's SCADA system, with a sampling frequency that can be set based on actual needs, such as once per second. CMS monitoring data is acquired from vibration sensors installed on key wind turbine components. These sensors monitor component vibration in real time, typically at a high sampling frequency of hundreds of times per second to capture subtle changes in vibration signals. Environmental sensor data, including wind speed, direction, temperature, and air pressure, is collected from meteorological monitoring stations installed at the wind farm. This data is crucial for analyzing the wind turbine cluster's operating environment and predicting potential failures.

[0022] The collected multi-source heterogeneous data differs in time and space, so preprocessing is required. Specifically, SCADA system data and CMS monitoring data are time-aligned and spatially matched to eliminate data acquisition delays and spatial deviations. An adaptive weighted fusion algorithm is used to associate environmental sensor data with equipment operation data to construct a unified data feature matrix. Data cleaning techniques are used to remove noise and redundant information to improve data quality. For example, a sliding average filter algorithm can be used to filter vibration data to remove abnormal vibration signals caused by sensor failure or external interference. At the same time, Z-score standardization is used to eliminate dimensional differences, convert the eigenvalues of different data sources into dimensionless standardized values, and generate standardized feature vectors with timestamps. A dynamic time warping algorithm is used to align data streams with different sampling frequencies, construct a multidimensional input sequence that is consistent in time and space, and ensure the temporal consistency of different data sources.

[0023] A wind power equipment health status assessment model is established based on a deep learning algorithm to analyze the equipment's operating status in real time and predict potential failure risks. Specifically, a baseline curve for equipment health status is established based on historical failure data. The dynamic features of real-time operating data are extracted through a convolutional neural network (CNN) and compared with the baseline curve for analysis. Equipment aging factors and operating condition correction coefficients are introduced to dynamically adjust the model output to reflect the actual health status of the equipment. The equipment aging factor reflects the performance degradation of the equipment over time, while the operating condition correction coefficient takes into account the impact of operating conditions on the equipment's health status, such as wind speed and load. Through these calibration measures, it is ensured that the health status assessment model can accurately reflect the real-time health status of the equipment.

[0024] The construction of the health status assessment model includes the following steps:

[0025] Data preprocessing: Standardize the collected multi-source heterogeneous data and eliminate outliers to build a wind power equipment operation status database.

[0026] Feature extraction: Convolutional neural networks are used to extract dynamic features of real-time operating data and generate key feature vectors that can reflect the health status of the equipment.

[0027] Model training: Establish a baseline curve for equipment health status based on historical fault data. Train the health status assessment model by comparing and analyzing real-time feature vectors with the baseline curve.

[0028] Model calibration: Introducing equipment aging factors and operating condition correction coefficients to dynamically adjust model outputs to ensure that the model can accurately reflect the actual health status of the equipment.

[0029] Based on the failure risk prediction results, combined with historical equipment maintenance records and O&M resource scheduling information, a multi-objective optimization algorithm is used to generate a preventive maintenance strategy. The optimization objectives include maximizing the overall operational reliability of the wind turbine cluster, minimizing maintenance costs and downtime, and balancing O&M resource allocation efficiency with equipment maintenance priorities. The optimal maintenance strategy is determined through Pareto front analysis, generating the optimal solution for the multi-objective optimization. Specifically, multi-objective optimization algorithms such as genetic algorithms and particle swarm optimization can be used, combined with Pareto front analysis, to generate an optimal operation strategy that balances multiple optimization objectives.

[0030] The generation of a preventive maintenance strategy involves the following steps:

[0031] Fault risk prediction: Based on the health status assessment model, it analyzes the equipment operating status in real time and predicts potential fault risks.

[0032] Optimization goal setting: Setting the optimization goals of the multi-objective optimization algorithm, including maximizing overall operational reliability, minimizing maintenance costs and downtime, and balancing operation and maintenance resource allocation efficiency with equipment maintenance priorities.

[0033] Strategy generation: Combining historical equipment maintenance records and operation and maintenance resource scheduling information, a multi-objective optimization algorithm is used to generate preventive maintenance strategies.

[0034] Dynamic Adjustment: Based on the wind farm's power generation plan and grid dispatch requirements, maintenance time windows are dynamically adjusted, prioritizing maintenance tasks during periods of low wind speed or low load to minimize power generation losses. The execution path of maintenance tasks is optimized by combining drone inspections and remote diagnostic results.

[0035] The method also includes a maintenance effectiveness feedback step, recording the results of each maintenance task and the subsequent equipment operating status. Actual maintenance effectiveness is compared and analyzed with the predictive model output. Based on the analysis results, the health status assessment model and maintenance strategy generation algorithm are iteratively optimized. Specifically, a maintenance effectiveness evaluation indicator system is established, including failure recurrence rate, maintenance time, and resource consumption. Statistical analysis is used to identify weak links in the maintenance strategy, and the model parameters and maintenance strategy library are dynamically updated based on the feedback data.

[0036] The method also includes a collaborative early warning step, integrating real-time data processing results from edge computing nodes to generate a global device health status heat map. When risk warnings are issued simultaneously for multiple devices, a collaborative maintenance mechanism is triggered to optimize resource scheduling. Specifically, edge computing nodes process device operating data in real time to generate a global device health status heat map. When risk warnings are issued simultaneously for multiple devices, a collaborative maintenance mechanism is triggered to optimize resource scheduling and ensure efficient utilization of maintenance resources.

[0037] The present invention also provides a wind power cluster preventive maintenance strategy intelligent generation system, which includes the following modules:

[0038] Data Acquisition Module: This module is used to acquire real-time, multi-source, heterogeneous data from the wind power cluster, including SCADA system operating data, CMS monitoring data, and environmental sensor data. This module uses multiple sensors and data interfaces to ensure data real-time and integrity.

[0039] Data processing module: This module is used for data standardization and feature extraction. It includes a data cleaning unit for removing noise and outliers, and a feature fusion unit for constructing a unified data feature matrix. These units work together to ensure high data quality and consistency.

[0040] Health Assessment Module: This module is used for operating status prediction and fault risk assessment. It integrates a deep learning model to support dynamic health status scoring and risk classification. It uses a convolutional neural network to extract dynamic features from real-time operating data and compares and analyzes them against baseline curves to generate device health status assessment results.

[0041] Strategy Generation Module: This module generates maintenance strategies based on a multi-objective optimization algorithm. This module combines historical equipment maintenance records with operational resource scheduling information to generate preventive maintenance strategies. Pareto frontier analysis is used to determine the optimal maintenance strategy, ensuring its scientific and practical validity.

[0042] Execution Control Module: This module distributes maintenance policies and coordinates maintenance resources. This module distributes generated maintenance policies to the wind farm cluster operation and maintenance system, enabling intelligent execution of preventive maintenance. By dynamically adjusting maintenance windows and optimizing execution paths for maintenance tasks, it reduces power generation losses and improves maintenance efficiency.

[0043] Feedback Optimization Module: This module is used to achieve closed-loop optimization of maintenance effectiveness. This module records the execution results of each maintenance task and the subsequent equipment operating status. It compares and analyzes actual maintenance results with the output of the prediction model. Based on these results, it iteratively optimizes the health assessment model and maintenance strategy generation algorithm. Using a maintenance effectiveness evaluation indicator system, it dynamically updates model parameters and the maintenance strategy library, ensuring continuous optimization and improvement of the system.

[0044] Through the collaborative operation of the above modules, the present invention enables intelligent generation of preventive maintenance strategies for wind power clusters, improving the accuracy and reliability of equipment health assessments, optimizing the allocation of maintenance resources, and reducing maintenance costs and downtime. The system not only monitors the operating status of equipment in real time but also uses intelligent algorithms to predict potential failures, enabling proactive maintenance measures and reducing both failure rates and repair costs. Furthermore, by optimizing data processing and model updating, the system reduces model lag, improves the system's real-time performance and response speed, and provides strong technical support for the intelligent operation and maintenance of wind power clusters.

[0045] It is understood from common technical knowledge that the present invention may be implemented by other embodiments that do not depart from its spirit or essential features. Therefore, the embodiments disclosed above are, in all respects, merely illustrative and not exclusive. All modifications within the scope of the present invention or equivalent to the scope of the present invention are intended to be encompassed by the present invention.

Claims

1. A method for intelligently generating a preventive maintenance strategy for a wind power cluster, characterized in that: The method includes the following steps: real-time collection of multi-source heterogeneous data of wind power clusters, including SCADA system operation data, CMS monitoring data and environmental sensor data; standardization of the collected data and elimination of outliers to build a wind power equipment operation status database; establishment of a wind power equipment health status assessment model based on a deep learning algorithm to analyze the equipment operation status in real time and predict potential failure risks; based on the failure risk prediction results, combined with the equipment's historical maintenance records and operation and maintenance resource scheduling information, a multi-objective optimization algorithm is used to generate a preventive maintenance strategy; the generated maintenance strategy is sent to the wind power cluster operation and maintenance system to realize the intelligent execution of preventive maintenance.

2. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The standardized processing of multi-source heterogeneous data includes the following steps: time alignment and spatial matching of SCADA system data and CMS monitoring data to eliminate data acquisition delay and spatial deviation; using an adaptive weighted fusion algorithm to associate environmental sensor data with equipment operation data to construct a unified data feature matrix; and using data cleaning technology to remove noise and redundant information to improve data quality.

3. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The construction of the wind power equipment health status assessment model includes: establishing an equipment health status baseline curve based on historical fault data; extracting the dynamic characteristics of real-time operating data through a convolutional neural network and comparing and analyzing it with the baseline curve; introducing equipment aging factors and operating condition correction coefficients to dynamically adjust the model output to reflect the actual health status of the equipment.

4. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The optimization objectives of the multi-objective optimization algorithm include: maximizing the overall operational reliability of the wind power cluster; minimizing maintenance costs and downtime; balancing the efficiency of operation and maintenance resource allocation and equipment maintenance priority; and determining the optimal maintenance strategy through Pareto frontier analysis.

5. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The generation of the preventive maintenance strategy also includes: dynamically adjusting the maintenance time window based on the wind farm's power generation plan and grid scheduling needs; prioritizing maintenance tasks during low wind speed or low load periods to reduce power generation losses; and optimizing the execution path of maintenance tasks by combining drone inspections and remote diagnosis results.

6. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The method also includes a maintenance effect feedback step: recording the execution results of each maintenance task and the subsequent operating status of the equipment; comparing and analyzing the actual maintenance effect with the output of the prediction model; and iteratively optimizing the health status assessment model and maintenance strategy generation algorithm based on the analysis results.

7. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 6, characterized in that: The maintenance effect feedback step specifically includes: establishing a maintenance effect evaluation index system, including fault recurrence rate, maintenance time, resource consumption, etc.; determining the weak links of the maintenance strategy through statistical analysis; and dynamically updating model parameters and maintenance strategy library based on feedback data.

8. The method for intelligently generating a preventive maintenance strategy for a wind power cluster according to claim 1, characterized in that: The method also includes a collaborative early warning step: integrating the real-time data processing results of edge computing nodes to generate a global device health status heat map; when risk warnings appear on multiple devices at the same time, triggering a collaborative maintenance mechanism to optimize resource scheduling plans.

9. Intelligent generation system of wind power cluster preventive maintenance strategy, characterized by: include: Data acquisition module, used to obtain multi-source heterogeneous data of wind power clusters in real time; Data processing module, used for data standardization and feature extraction; Health assessment module, used for operation status prediction and failure risk assessment; Strategy generation module, used to output maintenance strategies based on multi-objective optimization algorithms; The execution control module is used to issue strategies and coordinate operation and maintenance resources; the system also includes a feedback optimization module for achieving closed-loop optimization of maintenance effects.

10. The wind power cluster preventive maintenance strategy intelligent generation system according to claim 9, characterized in that: The data processing module further includes: a data cleaning unit for removing noise and outliers; a feature fusion unit for constructing a unified data feature matrix; the health assessment module integrates a deep learning model to support dynamic health status scoring and risk level classification.