A wind turbine automatic maintenance method and system based on intelligent robots

Abstract

The application provides a kind of wind turbine automatic maintenance method and system based on intelligent robot, obtains fault mode from fault record and operation log and carries out classification, then determines the distribution of fault law related to environment according to fault feature set, determines the risk level of environment in combination with real-time environment data, when the risk level exceeds the preset level, the potential fault distribution characteristics are analyzed by fusing real-time operation data, real-time environment data and historical fault characteristics, the correlation degree of real-time monitoring data and potential fault distribution characteristics is analyzed, the maintenance path adjustment mechanism is triggered, and then the fault prediction ability of intelligent robot is updated, the maintenance time window is obtained, the maintenance time window is optimized according to real-time environment data and real-time operation data, the final maintenance execution scheme is determined, the maintenance execution scheme is adjusted in time according to environmental change and equipment running state, the maintenance efficiency of wind power equipment is improved, the reliability and operation efficiency of risk equipment are improved.

Description

TECHNICAL FIELD

[0001] The embodiment of the application relates to the technical field of wind power equipment maintenance, and particularly relates to a wind turbine automatic maintenance method and system based on an intelligent robot. BACKGROUND

[0002] The rapid development of the wind power industry is of great significance to the optimization of energy structure and the realization of sustainable development. As an important pillar of clean energy, the stable operation of wind turbines is directly related to the reliability of energy supply. However, how to ensure efficient maintenance of equipment in a complex operating environment has become a key issue that needs to be broken through in this field. At present, although many maintenance methods have been applied to wind power equipment management, due to the lack of comprehensive consideration of multi-dimensional influencing factors related to wind turbines during maintenance, many maintenance methods cannot flexibly adjust the maintenance scheme when facing special environmental conditions and complex operating states of different wind turbines, resulting in low maintenance efficiency. SUMMARY

[0003] Therefore, the embodiment of the application aims to provide a wind turbine automatic maintenance method and system based on an intelligent robot, which can timely adjust the maintenance execution scheme of wind power equipment and improve the maintenance efficiency of wind power equipment.

[0004] To achieve the above-mentioned purpose, the embodiment of the application provides a wind turbine automatic maintenance method based on an intelligent robot, comprising:

[0005] Obtain fault records and operation logs from a pre-established wind power equipment historical database, and classify and process fault modes in the fault records and the operation logs to obtain a fault feature set;

[0006] Perform correlation analysis on environmental features in the fault feature set, and determine an environment-related fault rule distribution according to the analysis result;

[0007] Obtain real-time environmental data, determine whether an environmental condition reaches a preset abnormal threshold according to the fault rule distribution and the real-time environmental data, and determine an environmental risk level according to the abnormal threshold;

[0008] If the environmental risk level exceeds a preset level, analyze real-time operation data, the real-time environmental data and historical fault features according to a preset evaluation rule to obtain potential fault distribution features of each wind power equipment;

[0009] Obtain real-time monitoring data of equipment efficiency of each wind power equipment, analyze the correlation degree between the real-time monitoring data and the potential fault distribution features to determine whether a maintenance path adjustment mechanism is triggered;

[0010] If the maintenance path adjustment mechanism is triggered, the fault prediction capability of the intelligent robot is updated according to the historical monitoring data and the real-time environmental data to obtain a maintenance time window for equipment maintenance;

[0011] Path planning and optimization are performed on the maintenance time window according to the real-time environmental data and the real-time operation data to determine a maintenance execution plan.

[0012] Furthermore, the fault modes in the fault record and the operation log are classified to obtain a set of fault characteristics, including:

[0013] The fault modes are divided and classified according to time periods to obtain a fault mode set; the frequency and duration of the fault in each time period are obtained from the fault mode set and features are extracted to obtain a preliminary feature set; if the feature value of any feature in the preliminary feature set exceeds a first preset threshold, an abnormal operation log related to the abnormal feature exceeding the first preset threshold is obtained; it is determined whether the abnormal operation log has a potential fault mode association to obtain a screening feature set; the abnormal distribution feature values ​​of the screening feature set are presented in multiple dimensions to obtain a fault feature set.

[0014] Furthermore, the environmental characteristics include changes in ambient temperature and wind speed. Correlation analysis is performed on the environmental characteristics in the fault feature set, and a distribution of fault patterns related to the environment is determined based on the analysis results, including:

[0015] The ambient temperature and wind speed changes are compared with historical environmental records, and the fault frequency data corresponding to the environmental characteristics exceeding the second preset threshold in the comparison results are extracted to obtain an environmental correlation data set; the humidity level data and air pressure fluctuation data in the environmental correlation data set are divided into time periods and environmental interference analysis is performed to obtain environmental interference distribution characteristics; based on the environmental interference distribution characteristics, a potential correlation between environmental dynamic changes and the frequency of fault occurrence is determined; the air pressure fluctuation data exceeding the third preset threshold in the potential correlation is compared with the equipment aging information; the relevant distribution of environmental interference is obtained from the comparison results, the influence of the environmental dynamic changes on the duration of fault occurrence is determined, and the fault rule distribution is obtained.

[0016] Furthermore, judging whether the environmental condition reaches a preset abnormality threshold based on the fault regularity distribution and the real-time environmental data, and determining the environmental risk level based on the abnormality threshold, includes:

[0017] Relevant data on ambient temperature, wind speed parameters and humidity levels are obtained from the real-time environmental data, and the relevant data are formatted to obtain an environmental condition data set; the ambient temperature or the wind speed parameters in the environmental condition data set are compared with the abnormal threshold, and data exceeding the abnormal threshold is marked as an abnormal point to obtain an environmental data set carrying an abnormal identification; a correlation analysis is performed on the abnormal point and the humidity level in the environmental data set carrying the abnormal identification, and the abnormal distribution in the environmental conditions is determined based on the analysis results and the fault law distribution; the relevant data are graded according to the abnormal distribution to obtain an environmental risk level.

[0018] Furthermore, if the environmental risk level exceeds a preset level, the real-time operating data, the real-time environmental data, and historical fault characteristics are analyzed using preset assessment rules to obtain potential fault distribution characteristics of each wind turbine, including:

[0019] The operating parameters in the real-time operating data are integrated with the environmental risk values ​​in the real-time environment to obtain an operating status data set; the operating parameters in the operating status data set are compared with a fourth preset threshold, and the operating parameters exceeding the fourth preset threshold are marked as abnormal state points to obtain an abnormal state data set; the abnormal state data set is matched and analyzed with the historical fault characteristics to obtain potential fault distribution characteristics.

[0020] Furthermore, real-time monitoring data of the equipment efficiency of each wind power equipment is obtained, and the correlation between the real-time monitoring data and the potential fault distribution characteristics is analyzed to determine whether to trigger a maintenance path adjustment mechanism, including:

[0021] Obtain a dynamic data stream of the equipment efficiency, analyze the efficiency index based on the dynamic data stream; compare the efficiency index with a fifth preset threshold, mark the efficiency index below the fifth preset threshold as an abnormal fluctuation point, and obtain an abnormal fluctuation data set; compare the abnormal fluctuation point with the potential fault distribution characteristics to obtain a correlation degree; classify the operating status in the real-time monitoring data according to the correlation degree; compare the matching degree of the operating status at each level and the maintenance path with a sixth preset threshold, and trigger a maintenance path adjustment mechanism if it is lower than the sixth preset threshold.

[0022] Furthermore, after comparing the matching degree between the operating status and the maintenance path at each level with a sixth preset threshold and triggering a maintenance path adjustment mechanism if the matching degree is lower than the sixth preset threshold, the method further includes:

[0023] If the matching degree is lower than a sixth preset threshold, a maintenance path adjustment mechanism is triggered to obtain an adjustment requirement identifier; the maintenance path is dynamically updated according to the adjustment requirement identifier to generate an updated path scheme.

[0024] Further, if the maintenance path adjustment mechanism is triggered, the fault prediction capability of the intelligent robot is updated according to historical monitoring data and the real-time environment data to obtain a maintenance time window for equipment maintenance, including:

[0025] The fault prediction capability is updated and processed through the historical monitoring data and the real-time environment data to obtain a prediction capability improvement scheme; the historical monitoring data and the real-time environment data are compared and fused according to the prediction capability improvement scheme to obtain a data feature distribution; the data feature distribution is abnormity point screened through a seventh preset threshold, data exceeding the seventh preset threshold is marked as a high-risk prediction point to obtain a high-risk prediction point set; the fault prediction capability is optimized and adjusted according to the high-risk prediction point set to obtain an adjusted prediction capability parameter; a time window range matched with current environment data is obtained according to the prediction capability parameter to determine a maintenance time window.

[0026] Further, the maintenance time window is path planning optimized according to the real-time environment data and the real-time running data to determine a maintenance execution scheme, including:

[0027] The real-time environment data and the real-time running data are analyzed to obtain corresponding environment dynamic changes and equipment running states, the environment dynamic changes and the running states are compared in multiple dimensions to obtain a preliminary evaluation result of environment adaptability; the maintenance time window and equipment state data are matched according to the preliminary evaluation result, if the matching degree is lower than an eighth preset threshold, a path planning scheme is locally corrected to obtain a path optimization result; a range matched with the maintenance time window is obtained from the path optimization result, the real-time environment data and the real-time environment data are abnormity value detected; if abnormity data exceeding a ninth preset threshold is detected, the path optimization result is adjusted to obtain a final path planning scheme; a maintenance execution scheme adapted to the real-time environment data is generated according to the final path planning scheme.

[0028] To achieve the above object, the embodiment of the present application further provides a wind turbine automatic maintenance system based on an intelligent robot, including:

[0029] A classification module is configured to obtain fault records and running logs from a pre-established wind power equipment historical database, and to classify fault modes in the fault records and the running logs to obtain a fault feature set.

[0030] A first analysis module is configured to perform correlation analysis on the environmental features in the fault feature set and determine the distribution of fault patterns related to the environment based on the analysis results;

[0031] a judgment module, configured to obtain real-time environmental data, judge whether the environmental condition reaches a preset abnormality threshold based on the fault regularity distribution and the real-time environmental data, and determine the environmental risk level based on the abnormality threshold;

[0032] A second analysis module is configured to analyze the real-time operating data, the real-time environmental data, and historical fault characteristics using preset evaluation rules to obtain potential fault distribution characteristics of each wind turbine if the environmental risk level exceeds a preset level;

[0033] a third analysis module, configured to obtain real-time monitoring data of the equipment efficiency of each wind power equipment, analyze the correlation between the real-time monitoring data and the potential fault distribution characteristics, and determine whether to trigger a maintenance path adjustment mechanism;

[0034] An updating module, configured to update the fault prediction capability of the intelligent robot based on historical monitoring data and the real-time environmental data when the maintenance path adjustment mechanism is triggered, to obtain a maintenance time window for performing equipment maintenance;

[0035] An optimization module is used to perform path planning optimization on the maintenance time window according to the real-time environmental data and the real-time operation data, and determine a maintenance execution plan.

[0036] Compared with the prior art, the present invention has the following technical effects:

[0037] (1) The present invention can identify multi-dimensional fault mode types by obtaining and classifying fault modes from fault records and operation logs, and then determine the fault pattern distribution related to the environment based on the fault feature set, reflecting the correlation between environmental conditions and faults. The environmental risk level is determined in combination with real-time environmental data, which can effectively improve the accuracy of environmental risk assessment, and then analyze the influence of environmental factors on wind power equipment, which is helpful for the subsequent formulation of targeted maintenance measures based on the environmental risk level, ensuring the comprehensiveness and accuracy of fault analysis;

[0038] (2) The present invention integrates real-time operation data, real-time environmental data and historical fault characteristics to analyze the potential fault distribution characteristics. This correlation analysis combines historical fault characteristics and real-time data to explore deep fault risk factors, and combines multi-dimensional data of equipment operation and real-time environment to ensure the comprehensiveness of data analysis;

[0039] (3) The present invention combines the fault prediction capability of the intelligent robot with historical monitoring data and real-time environmental data to obtain a maintenance time window. The integration of historical monitoring data and real-time environmental data can improve the accuracy of the prediction. The determination of the maintenance time window is combined with real-time environmental data, which helps to reasonably arrange maintenance time and ensure more reasonable resource allocation.

[0040] (4) The present invention optimizes the maintenance time window based on real-time environmental data and real-time operating data, determines the final maintenance execution plan, and can timely adjust the maintenance execution plan according to environmental changes and equipment operating status, thereby improving the maintenance efficiency of wind power equipment, improving the reliability and operating efficiency of wind power equipment, reducing the maintenance cost of wind power equipment, and extending the service life of wind power equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of a first embodiment of a method for automatic maintenance of a wind turbine generator system based on an intelligent robot according to the present invention.

[0042] Figure 2 This is a schematic diagram of program modules of a second embodiment of an automatic maintenance system for wind turbines based on an intelligent robot according to the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is 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 intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0044] Example 1

[0045] See Figure 1 , shows a flowchart of the steps of an intelligent robot-based automatic wind turbine maintenance method according to Embodiment 1 of the present invention. It should be understood that the flowchart in this method embodiment does not limit the order in which the steps are executed. The following example description uses a computer device as the execution subject. The details are as follows.

[0046] Step S100: Acquire fault records and operation logs from a pre-established wind power equipment history database, classify the fault modes in the fault records and operation logs, and obtain a fault feature set.

[0047] In this embodiment, the wind power equipment includes a wind turbine generator set. Through a pre-established historical database of wind power equipment, fault records and operation log data are obtained, and the fault modes of different time periods are preliminarily classified. The fault records and operation log data are classified by time period to obtain a fault mode set. Based on the fault mode set, the frequency and duration of the fault in each time period are obtained, and the fault records and log data are subjected to feature extraction based on the frequency and duration data to determine a preliminary fault feature set. If the feature values ​​of certain features in the preliminary feature set exceed a first preset threshold, the features exceeding the first preset threshold are subjected to a secondary screening, and the operation logs associated with the features of the first preset threshold are obtained from the historical database to determine whether there is a potential fault mode association, thereby obtaining a screening feature set. The screening feature set is presented in multiple dimensions, and for the feature values ​​with abnormal distribution in the presentation results, the corresponding equipment operating parameters are obtained from the historical database to determine the final fault feature set.

[0048] Furthermore, step S100 further includes:

[0049] Fault modes are divided into time periods and fault classifications are performed to obtain a fault mode set; the frequency and duration of faults in each time period are obtained from the fault mode set and features are extracted to obtain a preliminary feature set; if the feature value of any feature in the preliminary feature set exceeds a first preset threshold, abnormal operation logs related to abnormal features exceeding the first preset threshold are obtained; it is determined whether the abnormal operation logs have potential fault mode associations to obtain a filtered feature set; the abnormal distribution feature values ​​of the filtered feature set are presented in multiple dimensions to obtain a fault feature set.

[0050] In this example, in the wind turbine equipment fault analysis scenario, fault records and operation log data in the historical database are divided by time period. Data within a year can be segmented by month, resulting in fault records for 12 time periods. Fault patterns can then be preliminarily categorized into various types, such as bearing failure, gear wear, and electrical failure. Within each time period, bearing failures may be more frequent in winter, while electrical failures are more frequent in summer. This categorization helps identify potential correlations between faults and seasonal factors.

[0051] Specifically, the frequency and duration of faults in each time period are extracted, and if the bearing fault frequency is 5 times and the average duration is 2 hours, and the electrical fault frequency is 3 times and the duration is 1.5 hours. Through statistical analysis, the distribution characteristics of frequency and duration are extracted to form a preliminary feature set. If the bearing fault frequency exceeds the preset threshold of 4 times, further analysis is required: retrieve the relevant operation logs from the historical database, and if the equipment operating temperature is generally high when the bearing fault occurs frequently, there may be potential fault mode correlation factors, and then a more accurate fault feature set is selected. After feature extraction, if the feature values of some features are abnormal, such as the proportion of fault modes with a duration of more than 3 hours is too high, secondary screening is required. By comparison, it is found that such fault modes are mostly related to wind speed changes, and combined with operating parameters such as wind speed data, the final fault feature set is determined, which can effectively eliminate noise data and improve feature accuracy.

[0052] Specifically, the screened feature set can be presented in multiple dimensions to generate a scatter plot of fault frequency and time, temperature. Assuming that the scatter plot shows that the fault frequency is abnormally concentrated in winter, the device operating parameters such as speed and lubrication state in the corresponding time period are extracted from the historical database, and it is found that low temperature leads to insufficient lubrication, and the final fault feature set is determined. Visualization can intuitively reveal the abnormal distribution and improve analysis efficiency, not only accurately positioning the fault features, but also providing data support for subsequent maintenance. The final fault feature set can be used to predict fault trends, optimize device operation strategies, reduce downtime, and significantly improve the reliability and economic benefits of wind power equipment.

[0053] Step S102, correlation analysis is performed on the environmental features in the fault feature set, and the environmental-related fault rule distribution is determined according to the analysis result.

[0054] In this embodiment, according to the environmental temperature and wind speed change data obtained from the pre-established fault feature set, the environmental temperature and wind speed change are compared with the historical environmental records, and the corresponding fault frequency data is extracted from the data exceeding the second preset threshold in the comparison result to obtain a preliminary environmental correlation data set. By dividing the humidity level and air pressure fluctuation data in the environmental correlation data set by time period, the distribution characteristics of environmental interference in the time period are obtained to determine the potential correlation mode between environmental dynamic changes and fault frequency. If the air pressure fluctuation data in some time period in the potential correlation mode exceeds the third preset threshold, the air pressure fluctuation data exceeding the third preset threshold is compared and screened with the device aging information to obtain the related distribution of environmental interference, and the influence law of environmental dynamic changes on fault duration is determined. According to the influence law, the environmental temperature and the wind speed change data are presented in multiple dimensions, the features of abnormal distribution in the presentation result are extracted, and the fault distribution law under the influence of the environment is determined.

[0055] Further, the environmental features include environmental temperature and wind speed variation, and step S102 further includes:

[0056] The environmental temperature and wind speed variation are compared with the historical environmental records, the fault frequency data corresponding to the environmental features exceeding the second preset threshold in the comparison result are extracted, to obtain an environmental correlation dataset; the humidity level data and the air pressure fluctuation data in the environmental correlation dataset are divided by time period and subjected to environmental interference analysis, to obtain environmental interference distribution features; a potential correlation between the environmental dynamic change and the frequency of fault occurrence is determined according to the environmental interference distribution features; the air pressure fluctuation data exceeding the third preset threshold in the potential correlation is compared with the equipment aging information; the relevant distribution of environmental interference is obtained from the comparison result, the influence law of the environmental dynamic change on the duration of fault occurrence is determined, and a fault law distribution is obtained.

[0057] In the embodiment, in the scenario of wind power equipment fault analysis, the environmental temperature and wind speed variation data are matched and analyzed with the historical records from the pre-established fault feature set. If the environmental temperature data range of a wind power plant within a year is-10 to 35 degrees Celsius, and the wind speed variation range is 3 to 25 meters per second, it is found after comparison with the historical environmental records that the fault frequency significantly increases when the temperature is below 0 degrees Celsius and the wind speed exceeds 20 meters per second, exceeding the preset threshold 2 times per month. The fault frequency information of the corresponding time period is extracted to form a preliminary environmental correlation dataset, laying a foundation for subsequent analysis. The humidity level and air pressure fluctuation data in the preliminary environmental correlation dataset can be divided by month, and the distribution features of environmental interference in each time period are analyzed. If the humidity level in July of a certain wind power plant is above 85%, and the air pressure fluctuation range is within 5 hundred pascals, the fault frequency is 30% higher than that in other months, it can be preliminarily judged that high humidity may interfere with equipment operation, and the potential correlation mode between environmental dynamic change and fault frequency is determined.

[0058] Specifically, if the air pressure fluctuation data of some time periods in the potential correlation mode exceeds the third preset threshold, such as the air pressure fluctuation reaches 8 hundred pascals in a certain month, far exceeding the normal range, the air pressure fluctuation data is compared with the equipment aging information. It can be known from the analysis of the comparison result that the fault duration of the equipment with high aging degree is prolonged by an average of 1.5 hours when the air pressure fluctuates dramatically. The comparison is helpful to determine the influence law of environmental interference on fault duration, and provides a basis for subsequent maintenance strategy.

[0059] In this embodiment, ambient temperature and wind speed change data can be presented in multiple dimensions based on their influencing patterns. A scatter plot is generated to illustrate the relationship between temperature, wind speed, and fault distribution. This scatter plot shows that when the temperature is below -5 degrees Celsius and the wind speed is above 18 meters per second, the fault distribution shows a concentrated trend and a significant abnormality. By extracting this abnormal distribution pattern, further analysis can be performed to determine the impact of environmental factors on the equipment, visually demonstrating the correlation between environmental conditions and faults, helping to quickly locate key issues.

[0060] Specifically, the correlation between ambient temperature and wind speed changes, as core factors, and fault frequency and duration was gradually analyzed, ensuring comprehensiveness and accuracy. By combining specific numerical values ​​and distribution characteristics during time series segmentation and data comparison, the system effectively identifies key points of environmental interference, providing data support for optimizing wind turbine operation. It also helps identify potential risk points in advance and enhances the equipment's adaptability in complex environments.

[0061] Step S104: Acquire real-time environmental data, determine whether the environmental conditions have reached a preset abnormal threshold based on the fault pattern distribution and the real-time environmental data, and determine the environmental risk level based on the abnormal threshold.

[0062] In this embodiment, data related to ambient temperature, wind speed parameters, and humidity levels are obtained from real-time environmental data collected by sensors. This data is formatted to produce a consolidated environmental condition dataset. The environmental condition dataset is then compared against preset thresholds. If the ambient temperature or wind speed parameter exceeds the abnormal threshold, it is marked as an outlier, resulting in a set of environmental data with an abnormality indicator. Correlation analysis is performed between the outliers and humidity levels to obtain dynamically changing distribution characteristics and determine the abnormal distribution of environmental conditions. The combined impact of the ambient temperature, wind speed parameters, and humidity levels in the abnormal distribution is then graded to determine the final classification result of the environmental risk level.

[0063] Furthermore, step S104 further includes:

[0064] Relevant data on ambient temperature, wind speed parameters, and humidity levels are obtained from real-time environmental data, and the relevant data are formatted to obtain an environmental condition data set. The ambient temperature or wind speed parameters in the environmental condition data set are compared with the abnormal threshold, and the data exceeding the abnormal threshold is marked as an abnormal point to obtain an environmental data set with an abnormal mark. Correlation analysis is performed between the abnormal points and the humidity levels in the environmental data set with the abnormal mark, and the abnormal distribution of the environmental conditions is determined based on the analysis results and the distribution of fault patterns. The relevant data is graded according to the abnormal distribution to obtain the environmental risk level.

[0065] In this embodiment, data on ambient temperature, wind speed parameters, and humidity levels can be acquired through a sensor network to record environmental changes within a wind farm in real time. For example, on a winter day at a wind farm, the ambient temperature was -8 degrees Celsius, the wind speed reached 22 meters per second, and the humidity was 80%. These relevant data are formatted to form a unified environmental condition dataset, facilitating subsequent analysis. Anomalies are identified within the environmental condition dataset. For example, if the preset ambient temperature threshold is -5 degrees Celsius or the wind speed threshold is 20 meters per second, the collected temperature and wind speed data are automatically marked as anomalies when they exceed these thresholds. This creates a dataset of environmental data with anomaly markers, helping to quickly identify environmental conditions that may affect equipment operation.

[0066] Specifically, the potential relationship between anomalies and humidity levels was analyzed. In the environmental dataset with anomaly markers, fluctuations in wind turbine operating conditions were more frequent when the temperature was below -5 degrees Celsius and the humidity level exceeded 75%. By analyzing the distribution characteristics of dynamic changes, the distribution of anomalies was determined. For example, in low-temperature, high-humidity environments, certain wind turbine components may be more susceptible to impact, providing important evidence for subsequent risk assessments.

[0067] In this embodiment, the combined impact of ambient temperature, wind speed parameters, and humidity levels in abnormal distribution conditions is graded. Environmental risk levels are divided into low, medium, and high. If the temperature is -10 degrees Celsius, the wind speed is 25 meters per second, and the humidity is 85% during a certain period of time, the combined impact is rated as a high risk level. Meanwhile, if the temperature is 0 degrees Celsius, the wind speed is 10 meters per second, and the humidity is 60% during another period of time, it may be rated as a low risk level. This intuitively reflects the degree of impact of different environmental conditions on equipment, helping to formulate targeted response measures.

[0068] In this embodiment, the collaborative integration of data, comparison, classification, and grading effectively improves the accuracy of environmental risk assessments. By ensuring data consistency through integration, screening outliers through comparison, and exploring correlations between environmental factors, risk management is then graded. This not only helps identify potential environmental threats but also provides data support for wind turbine operation and maintenance, optimizing resource allocation and extending equipment life.

[0069] Step S106: If the environmental risk level exceeds the preset level, the real-time operation data, real-time environmental data and historical fault characteristics are analyzed according to preset evaluation rules to obtain the potential fault distribution characteristics of each wind power equipment.

[0070] In this embodiment, operating parameters and environmental risk values ​​are obtained from real-time monitoring equipment and integrated to produce a consolidated operating status dataset. The operating status dataset is then compared item by item with preset thresholds. If an operating parameter exceeds a fourth preset threshold, it is marked as an abnormal state point, thereby determining an abnormal state dataset. The abnormal state point is then matched and analyzed with historical fault characteristics to obtain a dynamically changing fault relevance distribution and determine the distribution characteristics of potential faults. Based on the distribution characteristics of potential faults, a probability calculation is performed on the comprehensive impact of the device status to produce a potential fault probability assessment result for the current device.

[0071] Furthermore, step S106 further includes:

[0072] The operating parameters in the real-time operating data and the environmental risk values ​​in the real-time environment are integrated and processed to obtain an operating status data set; the operating parameters in the operating status data set are compared with a fourth preset threshold, and the operating parameters exceeding the fourth preset threshold are marked as abnormal state points to obtain an abnormal state data set; the abnormal state data set is matched and analyzed with historical fault characteristics to obtain potential fault distribution characteristics.

[0073] In this embodiment, in the wind turbine operation monitoring scenario, real-time monitoring equipment is used to obtain key operating parameters and environmental risk values. Operating parameters such as rotational speed and vibration frequency, as well as environmental risk factors such as wind speed and temperature, are integrated to form a unified operating status dataset, encompassing the real-time status of the equipment, facilitating further analysis of potential issues. For example, if a wind turbine's rotational speed is 1800 rpm, its vibration frequency is 5.2 Hz, and its ambient wind speed is 23 m / s and its temperature is -7°C at a certain moment, this integration forms a structured operating status dataset, providing a basis for subsequent comparisons. The operating status dataset is then compared item by item against a fourth preset threshold to identify abnormal status points. For example, if the rotational speed threshold is 2000 rpm, the vibration frequency threshold is 6.0 Hz, the wind speed threshold is 20 m / s, and the temperature threshold is -5°C, if the rotational speed is 1800 rpm and within the thresholds, but the wind speed of 23 m / s and the temperature of -7°C both exceed the thresholds, this data point is marked as an abnormal status point, forming the abnormal status dataset. It helps to quickly locate key issues that may affect equipment operation and provide accurate data support for subsequent analysis.

[0074] Specifically, the abnormal state point is matched and analyzed with the historical fault feature to obtain a dynamically changing fault correlation degree distribution, and a potential fault distribution feature is obtained. The historical data shows that when the wind speed exceeds 20 meters per second and the temperature is lower than-5 degrees, the probability of failure of the bearing part of the equipment is relatively high. By matching the current abnormal state point with the historical fault feature, it is judged that the current equipment may have the risk of bearing overload, which can reveal the distribution feature of potential failure and provide an important reference for predictive maintenance.

[0075] In this embodiment, the probability of failure distribution feature is calculated by the influencing factors related to the equipment state, and the potential failure probability evaluation result of each wind power equipment is obtained; a maintenance warning signal is triggered according to the potential failure probability evaluation result, so that a maintenance scheme can be prepared in advance according to the maintenance warning signal. The potential failure probability evaluation result of the current equipment is obtained by probabilistically calculating the comprehensive influence of the distribution feature of the potential failure on the equipment state. When the wind speed and temperature abnormal state point continues to appear, and the historical matching shows that the bearing failure probability is 70%, the potential failure probability of the current equipment is calculated as 65% by comprehensively considering other parameters such as vibration frequency. This intuitively reflects the current risk level of the equipment, which helps to develop a maintenance plan in advance and reduce the possibility of sudden failure.

[0076] From multiple aspects, the collaborative application of data integration, comparison, correlation and grading in this embodiment can effectively improve the reliability of wind power equipment operation monitoring. When data integration is performed, multi-dimensional information is obtained by sensors to ensure data comprehensiveness; in the comparison link, multi-level threshold values are set to adapt to different operation scenarios; in the correlation analysis, historical environmental data and real-time operation data are combined to mine deep risk factors; and in the grading evaluation, probability results are provided to support decision-making. From core monitoring to risk prediction, a complete logical chain is formed, which provides a strong guarantee for long-term stable operation of the equipment.

[0077] In step S108, real-time monitoring data of the equipment efficiency of each wind power equipment is obtained, and the correlation degree between the real-time monitoring data and the potential failure distribution feature is analyzed to determine whether to trigger the maintenance path adjustment mechanism.

[0078] In this embodiment, dynamic data flow of the equipment efficiency is obtained from the real-time monitoring equipment, the efficiency index is analyzed according to the dynamic data flow, and the efficiency index is compared with the fifth preset threshold value one by one. If the efficiency index is lower than the fifth preset threshold value, it is marked as an abnormal fluctuation point, and an abnormal fluctuation data set is obtained. The abnormal fluctuation point is compared and processed with the fault distribution feature to obtain the dynamic correlation degree distribution between them, and the correlation degree between the equipment efficiency fluctuation and the potential failure is determined. The running state is classified and divided through the correlation degree, the matching degree of each level of running state and the maintenance path is evaluated, and it is judged whether to trigger the maintenance path adjustment mechanism.

[0079] Furthermore, step S108 further includes:

[0080] Obtain a dynamic data stream of equipment efficiency, and analyze the efficiency index based on the dynamic data stream; compare the efficiency index with a fifth preset threshold, mark the efficiency index below the five preset thresholds as an abnormal fluctuation point, and obtain an abnormal fluctuation data set; compare the abnormal fluctuation point with the potential fault distribution characteristics to obtain the degree of correlation; classify the operating status in the real-time monitoring data according to the degree of correlation; compare the matching degree of each level of operating status and the maintenance path with a sixth preset threshold, and if it is lower than the sixth preset threshold, trigger the maintenance path adjustment mechanism.

[0081] In this embodiment, in the scenario of wind power equipment operation monitoring, the dynamic data stream of equipment efficiency is continuously obtained through real-time monitoring equipment, and the efficiency index is compared one by one with the fifth preset threshold. The efficiency index of a certain wind power equipment is 85%, and the preset threshold is 90%. Since it is lower than the threshold, it is marked as an abnormal fluctuation point, forming an abnormal fluctuation data set, which helps to quickly identify potential risk points of efficiency decline. The abnormal fluctuation point can be compared with the historical fault distribution characteristics to obtain the dynamic correlation distribution and the degree of correlation. Historical data shows that when the efficiency index is lower than 90% and lasts for more than 2 hours, the possibility of partial failure of the equipment blade is higher. By comparing the current efficiency fluctuation with the historical characteristics, it is judged that the current equipment may have the risk of blade damage, and the degree of correlation is assessed as medium to high, providing data support for subsequent risk grading.

[0082] In this embodiment, after determining the degree of correlation, the operating status is classified into different levels, and the matching degree between each level of status and the maintenance path is evaluated. The operating status can be divided into three levels. Level 2 corresponds to high-correlation faults with efficiency fluctuations, and the matching maintenance path is to increase the inspection frequency. If the matching degree is 75%, which is lower than the preset threshold of 80%, the maintenance path adjustment mechanism is triggered and an adjustment need indicator is generated, helping to ensure that the maintenance path matches the actual risk.

[0083] In this embodiment, when comparing efficiency indicators and flagging anomalies, additional analysis can be conducted in conjunction with environmental factors. For example, if the efficiency is below 90% and the ambient wind speed reaches 25 m / s, comprehensive consideration of the impact of wind speed on efficiency can further confirm the cause of the abnormal fluctuation and help more accurately locate the root cause of the problem.

[0084] Furthermore, after step S108, the method further includes:

[0085] If the matching degree is lower than a sixth preset threshold, the maintenance path adjustment mechanism is triggered to obtain an adjustment requirement identifier; the maintenance path is dynamically updated according to the adjustment requirement identifier to generate an updated path plan.

[0086] In this embodiment, the maintenance path is dynamically updated according to the adjustment demand identifier, and the updated path scheme is generated by combining the monitoring frequency and the data integration result, and it is judged whether it meets the comprehensive demand of equipment efficiency and failure probability. The maintenance path is dynamically updated according to the adjustment demand identifier, and a new scheme is generated by combining the monitoring frequency and the data integration result. For example, if the original path is to inspect once a week, the updated path is to inspect once every three days, and the blade part is checked first, and the data collection frequency is increased to once an hour. It is further judged whether the new path meets the comprehensive demand of equipment efficiency and failure probability, such as confirming that the new path can effectively cover the risk points and can improve the pertinence of maintenance through simulation analysis.

[0087] Specifically, in the maintenance path update, personnel scheduling optimization can also be introduced. If the new path needs to increase the inspection frequency, the shortest route can be planned according to the available time of personnel and the location of equipment to ensure the efficiency of resource utilization, reduce time waste, and improve response speed. The operating state can also be classified, and a multi-level early warning mechanism can be set. For example, if the efficiency fluctuation correlation degree is medium-high, a yellow early warning is issued to remind relevant personnel to pay attention, and if it further deteriorates, a red early warning is upgraded, which helps to intervene in advance and reduce the possibility of failure.

[0088] Step S110, if the maintenance path adjustment mechanism is triggered, the failure prediction ability of the intelligent robot is updated according to the historical monitoring data and real-time environment data, and a maintenance time window for equipment maintenance is obtained.

[0089] In this embodiment, the failure prediction ability is updated and processed by the intelligent diagnosis framework, the integrated information of the historical monitoring data and the real-time environment data is obtained, and a preliminary prediction ability improvement scheme is determined. The historical monitoring data and the real-time environment data are deeply compared according to the prediction ability improvement scheme, and the fused data feature distribution is obtained. The data feature distribution is screened for abnormal points by a seventh preset threshold, and if the data feature distribution exceeds the seventh preset threshold, it is marked as a high-risk prediction point, and a high-risk prediction point set is obtained. The support vector machine (SVM) is used to optimize and adjust the prediction ability according to the high-risk prediction point set, and the adjusted prediction ability parameter is obtained. The time window range matched with the current environment data is obtained through the maintenance time window and the prediction ability parameter, and the optimized maintenance time window is determined. The maintenance time window is a continuous time interval for managing automatic maintenance tasks. Maintenance can be performed during the business low peak period to reduce the impact on business.

[0090] Further, step S110 further comprises:

[0091] The fault prediction capability is updated and processed through historical monitoring data and real-time environmental data to obtain a prediction capability improvement plan; the historical monitoring data and real-time environmental data are compared and integrated according to the prediction capability improvement plan to obtain the data feature distribution; the data feature distribution is screened for abnormal points through the seventh preset threshold, and the data exceeding the seventh preset threshold is marked as a high-risk prediction point to obtain a high-risk prediction point set; the fault prediction capability is optimized and adjusted according to the high-risk prediction point set to obtain adjusted prediction capability parameters; the time window range that matches the current environmental data is obtained according to the prediction capability parameters to determine the maintenance time window.

[0092] In this embodiment, a pre-defined intelligent diagnostic framework integrates historical monitoring data and real-time environmental data to gradually enhance fault prediction capabilities. A data acquisition system regularly collects equipment operating parameters, such as speed and power output, from historical monitoring data. This data is combined with real-time environmental data, such as wind speed and temperature, to form a comprehensive dataset that is fed into the intelligent diagnostic framework for improved and updated predictive capabilities. For example, if historical data indicates a 10% drop in power output for a particular wind turbine when wind speeds exceed 20 m / s, and real-time monitoring and environmental data also show similar trends, a preliminary assessment of the potential for a potential fault can be made, providing a more comprehensive understanding of the equipment's status.

[0093] In this example, a deep comparison of historical monitoring data and real-time environmental data was performed. Time series analysis was used to overlay and analyze efficiency fluctuation data from the past year with environmental data from the current week. This fusion revealed a strong correlation between efficiency drops and sudden increases in wind speed within a certain time period. This characteristic distribution was then extracted, and the hidden correlations were then mined to obtain the data characteristic distribution.

[0094] In this embodiment, the efficiency drop of more than 5% or the abnormal fluctuation of wind speed can be preset as the seventh preset threshold. If the efficiency of a certain wind power equipment drops by 7% at a certain point in time, it is marked as a high-risk prediction point and included in the high-risk prediction point set, which helps to quickly lock in the key points that need attention. The optimization and adjustment of the high-risk prediction point set can be carried out by calibrating the parameters of the prediction ability through the support vector machine model. If the model analysis finds that the efficiency drop is highly correlated with blade wear, the prediction parameters are adjusted to increase the weight of attention to the blade status. This optimization method can improve the accuracy of the prediction.

[0095] In the dynamic planning of the maintenance time window in this embodiment, a time range of 48 hours suitable for maintenance can be determined according to real-time environmental data, such as a prediction that the wind speed will be low in the next three days. For example, if the original time window is a fixed day every week, the optimization adjusts the time window to be flexibly arranged according to environmental predictions, which helps to reasonably arrange the maintenance time. If the matching degree of the optimized maintenance time window and the maintenance path is lower than a preset value of 80%, such as only 70%, the path adjustment mechanism is used to re-plan. The original path covers a wide range but is less efficient, and after adjustment, the new path configuration prioritizes covering high-risk equipment areas, which can ensure more reasonable resource allocation. A specific plan is generated by combining real-time environmental data and the maintenance time window to obtain a maintenance execution scheme. For example, the new scheme requires completing the inspection of key equipment within the next 24 hours and adjusting the inspection order according to the wind speed prediction, which helps to ensure the executability of the scheme.

[0096] Step S112, path planning optimization of the maintenance time window is performed according to real-time environmental data and real-time operation data to determine a maintenance execution scheme.

[0097] In this embodiment, the preliminary evaluation result of the environmental adaptability is determined by comprehensively comparing the distribution characteristics of multi-dimensional factors of the environmental dynamic changes and the equipment operation states according to the maintenance time window and the real-time environmental data. The equipment state complex data, i.e., the equipment state data, is deeply matched with the maintenance time window in the preliminary evaluation result. If the matching degree is lower than an eighth preset threshold, the path planning scheme is locally corrected to obtain a path optimization result. The range of the time window matching is obtained from the path optimization result, and the real-time environmental data corresponding to the environmental dynamic changes is subjected to outlier detection. If the detected outliers exceed a ninth preset threshold, the path optimization result is adjusted again to obtain a final path planning scheme. A maintenance execution scheme adapted to the real-time environmental data is generated according to the final path planning scheme, and an execution arrangement that matches the maintenance time window is determined.

[0098] Further, step S112 further includes:

[0099] The real-time environmental data and the real-time operation data are analyzed to obtain corresponding environmental dynamic changes and equipment operation states. The environmental dynamic changes and the equipment operation states are compared in multiple dimensions to obtain a preliminary evaluation result of the environmental adaptability. The equipment state data is matched with the maintenance time window according to the preliminary evaluation result. If the matching degree is lower than an eighth preset threshold, the path planning scheme is locally corrected to obtain a path optimization result. The range of the time window matching is obtained from the path optimization result, and the real-time environmental data and the real-time operation data are subjected to outlier detection. If outliers exceeding a ninth preset threshold are detected, the path optimization result is adjusted to obtain a final path planning scheme. A maintenance execution scheme adapted to the real-time environmental data is generated according to the final path planning scheme.

[0100] For example, in order to implement the optimized maintenance time window and dynamic path adjustment algorithm, a data acquisition system is used to monitor the dynamic changes in the environment and the status of equipment in real time. In an industrial park, there are 10 pieces of equipment that require maintenance. The operating data of each piece of equipment includes temperature, vibration frequency, and energy consumption. The acquisition frequency is once an hour, and the data is stored in a cloud database. The temperature threshold is set to 80 degrees Celsius, and the vibration frequency threshold is set to 5 Hz. If the eighth preset threshold is exceeded, a maintenance demand signal is triggered. A path planning model is constructed based on multi-dimensional factors, taking environmental factors and equipment status into consideration. For example, if the probability of rainfall in the weather forecast is abnormal, and the probability is greater than the ninth preset threshold of 60%, the maintenance time window needs to be adjusted. If the historical data analysis shows that the probability of failure of a certain equipment is 0.25, which is greater than the ninth preset threshold of 0.2, the maintenance time window needs to be adjusted. The ninth preset threshold is adjusted according to the abnormal data and can be a set of threshold data. The optimal route was calculated using a modified Dijkstra algorithm. A distance matrix was set between the maintenance vehicle's starting point and each device, such as 3 kilometers from device A to device B and 5 kilometers from device B to device C. The Dijkstra algorithm, designed to minimize total distance and time, calculated a total route length of 28 kilometers, with an estimated travel time of 4.5 hours. Real-time traffic data was incorporated to dynamically adjust the route. For example, if a sudden congestion on a certain road section increased travel time by 30 minutes, traffic information was retrieved through an API and the route was recalculated to avoid the congested section. This adjustment increased the route length to 30 kilometers, but reduced the total travel time to 4.2 hours, resulting in the final route plan. When determining the maintenance execution plan, the system takes into account the time window constraint, limiting daily maintenance time to within 6 hours, and priority sorting, prioritizing equipment with a high probability of failure. This generates the final maintenance execution plan. For example, equipment B with a failure probability of 0.25 is prioritized for maintenance and placed first in the maintenance time window. The remaining equipment is arranged in order of distance optimization to ensure that the total time is controlled within 5.8 hours. The system automatically completes the closed-loop processing from data collection to plan generation, logically forming a complete chain from environmental perception to path optimization, ensuring maximum maintenance efficiency and resource utilization.

[0101] Example 2

[0102] Please continue reading Figure 2, showing a program module schematic diagram of a second embodiment of an automatic maintenance system for wind turbines based on an intelligent robot according to the present invention. In this embodiment, an automatic maintenance system 20 for wind turbines based on an intelligent robot may include or be divided into one or more program modules, one or more program modules are stored in a storage medium, and are executed by one or more processors to complete the present invention and realize the above-mentioned automatic maintenance method for wind turbines based on an intelligent robot. The program module referred to in the embodiment of the present invention refers to a series of computer program instruction segments that can perform specific functions, and is more suitable for describing the execution process of an automatic maintenance system 20 for wind turbines based on an intelligent robot in a storage medium than the program itself. The following description will specifically introduce the functions of each program module of this embodiment:

[0103] The classification module 200 is configured to obtain fault records and operation logs from a pre-established wind power equipment history database, classify the fault modes in the fault records and operation logs, and obtain a set of fault characteristics.

[0104] The first analysis module 202 is configured to perform correlation analysis on the environmental features in the fault feature set, and determine the distribution of fault patterns related to the environment based on the analysis results.

[0105] The judgment module 204 is used to obtain real-time environmental data, judge whether the environmental conditions have reached a preset abnormal threshold according to the fault regular distribution and the real-time environmental data, and determine the environmental risk level according to the abnormal threshold.

[0106] The second analysis module 206 is configured to analyze the real-time operation data, the real-time environmental data and historical fault characteristics according to preset evaluation rules to obtain potential fault distribution characteristics of each wind power equipment if the environmental risk level exceeds a preset level.

[0107] The third analysis module 208 is configured to obtain real-time monitoring data of the equipment efficiency of each wind power equipment, analyze the correlation between the real-time monitoring data and the potential fault distribution characteristics, and determine whether to trigger a maintenance path adjustment mechanism.

[0108] The updating module 210 is used to update the fault prediction capability of the intelligent robot according to the historical monitoring data and the real-time environmental data if the maintenance path adjustment mechanism is triggered, so as to obtain a maintenance time window for equipment maintenance.

[0109] The optimization module 212 is configured to optimize the path planning of the maintenance time window according to the real-time environmental data and the real-time operation data, and determine a maintenance execution plan.

[0110] An embodiment of the present invention provides an automatic maintenance system for wind turbines based on intelligent robots. The system obtains and classifies fault modes from fault records and operation logs, determines the distribution of fault patterns related to the environment based on a set of fault characteristics, and determines the environmental risk level in combination with real-time environmental data. When the risk level exceeds a preset level, the system integrates real-time operation data, real-time environmental data and historical fault characteristics to analyze the potential fault distribution characteristics, analyzes the degree of correlation between real-time monitoring data and potential fault distribution characteristics, triggers a maintenance path adjustment mechanism, and then updates the fault prediction capability of the intelligent robot to obtain a maintenance time window. The maintenance time window is optimized based on real-time environmental data and real-time operation data, and the final maintenance execution plan is determined. The present invention can timely adjust the maintenance execution plan according to environmental changes and equipment operating status, thereby improving the maintenance efficiency of wind power equipment, improving the reliability and operating efficiency of wind power equipment, reducing the maintenance cost of wind power equipment, and extending the service life of wind power equipment.

[0111] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0112] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method.

[0113] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A wind turbine automatic maintenance method based on an intelligent robot, characterized in that: include: Obtaining fault records and operation logs from a pre-established wind power equipment history database, classifying and processing the fault modes in the fault records and the operation logs to obtain a set of fault characteristics; Performing correlation analysis on the environmental features in the fault feature set, and determining the distribution of fault patterns related to the environment based on the analysis results; Acquire real-time environmental data, determine whether environmental conditions have reached a preset abnormality threshold based on the fault regularity distribution and the real-time environmental data, and determine an environmental risk level based on the abnormality threshold; If the environmental risk level exceeds a preset level, the real-time operating data, the real-time environmental data and historical fault characteristics are analyzed using preset evaluation rules to obtain potential fault distribution characteristics of each wind power equipment; Acquiring real-time monitoring data of equipment efficiency of each wind power equipment, and analyzing the correlation between the real-time monitoring data and the potential fault distribution characteristics to determine whether to trigger a maintenance path adjustment mechanism; If the maintenance path adjustment mechanism is triggered, the fault prediction capability of the intelligent robot is updated according to the historical monitoring data and the real-time environmental data to obtain a maintenance time window for equipment maintenance; Optimize the path planning of the maintenance time window according to the real-time environmental data and the real-time operation data, and determine a maintenance execution plan; The process of obtaining real-time monitoring data of the equipment efficiency of each wind power equipment and analyzing the correlation between the real-time monitoring data and the potential fault distribution characteristics to determine whether to trigger the maintenance path adjustment mechanism includes: Acquiring a dynamic data stream of the equipment efficiency, and analyzing efficiency indicators based on the dynamic data stream; Comparing the efficiency index with a fifth preset threshold, marking the efficiency index below the fifth preset threshold as an abnormal fluctuation point, and obtaining an abnormal fluctuation data set; Comparing the abnormal fluctuation point with the potential fault distribution characteristics to obtain a correlation degree; Classifying the operating status in the real-time monitoring data according to the degree of association; The matching degree between the operating status of each level and the maintenance path is compared with a sixth preset threshold, and if it is lower than the sixth preset threshold, the maintenance path adjustment mechanism is triggered.

2. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1 is characterized in that: Classify the fault modes in the fault record and the operation log to obtain a set of fault characteristics, including: Dividing the fault mode into time periods and classifying the faults to obtain a set of fault modes; Obtaining the frequency and duration of the fault in each time period from the fault mode set and performing feature extraction to obtain a preliminary feature set; If the feature value of any feature in the preliminary feature set exceeds a first preset threshold, obtaining an abnormal operation log related to the abnormal feature exceeding the first preset threshold; Determine whether the abnormal operation log has potential failure mode association, and obtain a screening feature set; The abnormal distribution characteristic values ​​of the screening feature set are presented in multiple dimensions to obtain a fault feature set.

3. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: The environmental characteristics include changes in ambient temperature and wind speed. Correlation analysis is performed on the environmental characteristics in the fault feature set, and a distribution of fault patterns related to the environment is determined based on the analysis results, including: Comparing the ambient temperature and the wind speed changes with historical environmental records, extracting fault frequency data corresponding to environmental features exceeding a second preset threshold in the comparison results, to obtain an environment-related data set; Dividing the humidity level data and the air pressure fluctuation data in the environment-related data set into time periods and performing environmental interference analysis to obtain environmental interference distribution characteristics; Determining a potential correlation between environmental dynamic changes and the frequency of fault occurrence based on the environmental interference distribution characteristics; Comparing the air pressure fluctuation data exceeding a third preset threshold in the potential association relationship with the device aging information; The relevant distribution of environmental interference is obtained from the comparison results, the influence pattern of the dynamic change of the environment on the duration of the fault occurrence is determined, and the fault pattern distribution is obtained.

4. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: Determining whether the environmental condition reaches a preset abnormality threshold based on the fault regularity distribution and the real-time environmental data, and determining the environmental risk level based on the abnormality threshold, including: Acquiring relevant data of ambient temperature, wind speed parameters and humidity level from the real-time environmental data, and formatting the relevant data to obtain an environmental condition data set; Comparing the ambient temperature or the wind speed parameter in the environmental condition data set with the abnormality threshold, marking data exceeding the abnormality threshold as an abnormal point, and obtaining an environmental data set carrying an abnormality mark; Performing correlation analysis on the abnormal point and the humidity level in the environmental data set carrying the abnormality identifier, and determining the abnormal distribution in the environmental condition based on the analysis result and the fault regular distribution; The relevant data is graded according to the abnormal distribution to obtain an environmental risk grade.

5. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: If the environmental risk level exceeds a preset level, the real-time operating data, the real-time environmental data, and historical fault characteristics are analyzed using preset assessment rules to obtain potential fault distribution characteristics of each wind turbine, including: Integrating the operating parameters in the real-time operating data with the environmental risk values ​​in the real-time environment to obtain an operating status data set; comparing the operating parameters in the operating status data set with a fourth preset threshold, marking the operating parameters exceeding the fourth preset threshold as abnormal state points, to obtain an abnormal state data set; The abnormal state data set is matched and analyzed with the historical fault characteristics to obtain potential fault distribution characteristics.

6. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: After comparing the matching degree between the operating status and the maintenance path at each level with a sixth preset threshold and triggering a maintenance path adjustment mechanism if the matching degree is lower than the sixth preset threshold, the method further includes: If the matching degree is lower than a sixth preset threshold, a maintenance path adjustment mechanism is triggered to obtain an adjustment requirement identifier; The maintenance path is dynamically updated according to the adjustment requirement identifier to generate an updated path plan.

7. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: If the maintenance path adjustment mechanism is triggered, the fault prediction capability of the intelligent robot is updated based on the historical monitoring data and the real-time environmental data to obtain a maintenance time window for equipment maintenance, including: The fault prediction capability is updated by using the historical monitoring data and the real-time environmental data to obtain a prediction capability improvement plan; Comparing and fusing the historical monitoring data with the real-time environmental data according to the prediction capability improvement plan to obtain data feature distribution; Screening the data feature distribution for abnormal points using a seventh preset threshold, marking data exceeding the seventh preset threshold as high-risk prediction points, and obtaining a high-risk prediction point set; Optimizing and adjusting the fault prediction capability according to the high-risk prediction point set to obtain adjusted prediction capability parameters; A time window range matching the current environmental data is obtained based on the prediction capability parameter to determine a maintenance time window.

8. The automatic maintenance method for wind turbines based on intelligent robots according to claim 1, characterized in that: Optimizing the path planning of the maintenance time window according to the real-time environmental data and the real-time operation data to determine a maintenance execution plan includes: Analyze the real-time environmental data and the real-time operating data to obtain corresponding dynamic environmental changes and equipment operating status, perform multi-dimensional comparison of the dynamic environmental changes and the operating status, and obtain a preliminary assessment result of environmental adaptability; matching the maintenance time window with the equipment status data according to the preliminary evaluation result, and if the matching degree is lower than an eighth preset threshold, performing a partial correction on the path planning scheme to obtain a path optimization result; Obtaining a matching range of the maintenance time window from the path optimization result, and performing outlier detection on the real-time environment data and the real-time operation data; If abnormal data exceeding a ninth preset threshold is detected, the path optimization result is adjusted to obtain a final path planning solution; A maintenance execution plan adapted to the real-time environmental data is generated according to the final path planning plan.

9. An automatic maintenance system for wind turbines based on intelligent robots, characterized in that: Used to implement the method according to any one of claims 1 to 8, comprising: A classification module is used to obtain fault records and operation logs from a pre-established wind power equipment history database, classify the fault modes in the fault records and operation logs, and obtain a set of fault characteristics; A first analysis module is configured to perform correlation analysis on the environmental features in the fault feature set and determine the distribution of fault patterns related to the environment based on the analysis results; a judgment module, configured to obtain real-time environmental data, judge whether the environmental condition reaches a preset abnormality threshold based on the fault regularity distribution and the real-time environmental data, and determine the environmental risk level based on the abnormality threshold; A second analysis module is configured to analyze the real-time operating data, the real-time environmental data, and historical fault characteristics using preset evaluation rules to obtain potential fault distribution characteristics of each wind turbine if the environmental risk level exceeds a preset level; a third analysis module, configured to obtain real-time monitoring data of the equipment efficiency of each wind power equipment, and analyze the correlation between the real-time monitoring data and the potential fault distribution characteristics to determine whether to trigger a maintenance path adjustment mechanism; An updating module, configured to update the fault prediction capability of the intelligent robot based on historical monitoring data and the real-time environmental data when the maintenance path adjustment mechanism is triggered, to obtain a maintenance time window for performing equipment maintenance; An optimization module is used to perform path planning optimization on the maintenance time window according to the real-time environmental data and the real-time operation data, and determine a maintenance execution plan.

Images