Fan fault diagnosis and online monitoring method based on time sequence parameter analysis
By acquiring real-time wind turbine timing parameter operation data, mapping correlation chains, and calculating abnormal threshold ranges, the problem of poor adaptability in wind turbine fault diagnosis is solved, and accurate online monitoring and fault level judgment are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-14
- Publication Date
- 2026-04-14
AI Technical Summary
Existing technologies rely on static threshold judgments for wind turbine fault diagnosis, resulting in poor adaptability and difficulty in matching the complex and ever-changing operating conditions of thermal power plants, leading to inaccurate diagnosis and missed diagnoses.
By acquiring real-time wind turbine timing parameter operation data, mapping the timing parameter correlation chain, obtaining abnormal impact data, and calculating the threshold range of abnormal operation data, online fault monitoring and level judgment can be achieved.
It improves the adaptability and accuracy of wind turbine fault diagnosis, enables adaptive adjustment to complex operating conditions, and reduces missed diagnoses.
Smart Images
Figure CN121854460A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wind turbine fault diagnosis technology, and specifically to a wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis. Background Technology
[0002] In thermal power plants, fans are the core power equipment for boiler forced and induced draft, desulfurization, and denitrification systems. Their operational reliability directly affects the safe and stable power generation and environmental compliance of the unit. Timing parameters are key configurations in electronic systems used to accurately control signal transmission, data processing, and event occurrence times, ensuring that operations are executed at the correct points in time and avoiding conflicts or data errors. A fan is a machine that uses input mechanical energy to increase gas pressure and discharge gas. It is a type of driven fluid machinery, and its time-series parameters include, but are not limited to, vibration, speed, and temperature. In existing technologies, the diagnosis of fan faults based on time-series parameters usually relies solely on static threshold ranges to judge abnormal conditions. This results in poor adaptability of the fan during operation, making it difficult to match the complex and ever-changing operating conditions of thermal power plants. Consequently, it cannot adaptively adjust to the operating conditions and may miss anomalies due to fixed thresholds. Therefore, the static threshold method judges only a single parameter independently, severing the inherent relationship between parameters, resulting in a strong one-sidedness in diagnosis, easy to miss diagnoses, and inaccurate diagnosis. Therefore, in order to improve the adaptability and comprehensiveness of wind turbine fault diagnosis, this invention provides a wind turbine fault diagnosis and online monitoring method based on time series parameter analysis. Summary of the Invention
[0003] The purpose of this invention is to provide a wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis to address the shortcomings in the prior art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for wind turbine fault diagnosis and online monitoring based on time-series parameter analysis, the method comprising the following steps: Step S1: Real-time acquisition of the time-series parameter operation data of the wind turbine at the corresponding time point, mapping the influence parameters of each time-series parameter in the time-series parameter operation data; and real-time acquisition of the influence data corresponding to the influence parameters; and anomaly analysis of the influence data to obtain abnormal influence data. Step S2: Obtain the standard threshold of the time series parameters, and then calculate the abnormal operation data threshold range of the time series parameters at the corresponding abnormal time points based on the obtained abnormal impact data threshold range; obtain the operation data threshold range of the time series parameters at the corresponding time points through the abnormal operation data threshold range; Step S3: Based on the operational data threshold range, perform online fault monitoring of the wind turbine to obtain fault operation data; Step S4: Obtain the abnormal timing parameter operation data of the fault operation data, and determine the abnormality level of the time point based on the abnormal timing parameter operation data to obtain the abnormality level time point.
[0005] Furthermore, the process of acquiring real-time operating data of wind turbine time parameters at corresponding time points and mapping the time parameter correlation chain of each time parameter in the operating data includes: The timing parameters of the structural components of the wind turbine are set according to the structural components, and the corresponding acquisition period is set. The timing parameter operation data at the corresponding time points are collected according to the acquisition period, and mapped with the preset area number of the corresponding structural component. The timing parameters include at least rotational speed, vibration and temperature. The influence parameters of the timing parameters corresponding to different structural components are obtained based on big data, and the influence position of the influence parameters is obtained. The influence position is mapped with the corresponding influence parameter, and then associated with the corresponding timing parameter to generate a timing parameter association chain.
[0006] Furthermore, the impact data corresponding to the impact parameters is acquired in real time; the process of obtaining abnormal impact data through anomaly analysis of the impact data includes: Each influencing location is obtained through a time-series parameter association chain, and the influencing locations are divided into several influencing location nodes. A data acquisition sensor is set at each influencing location node to collect the node influence data corresponding to the influencing location node. The node influence data is fused to obtain the influence data of the influence parameters corresponding to the influencing location, and mapped with the time point of the corresponding time-series parameter running data. Based on the threshold range of the impact data set for structural components, impact data that does not fall within the threshold range at the first time point is recorded as abnormal impact data; conversely, the impact data of each node of the impact data is analyzed, and the impact data of nodes that do not fall within the threshold range is recorded as abnormal node impact data. Then, the impact data corresponding to the impact location of the abnormal node impact data is recorded as abnormal impact data; otherwise, no processing is performed; and the time point corresponding to the abnormal impact data is marked as an abnormal time point.
[0007] Furthermore, the process of obtaining standard thresholds for time-series parameters, and then extrapolating the threshold range of abnormal operational data for the corresponding abnormal time points based on the threshold range of the impact data of abnormal data, includes: By obtaining the impact parameters of the abnormal impact data at abnormal time points, the corresponding time series parameters of the time series parameters are obtained, and then the running data of the time series parameters at the abnormal time points are obtained and marked as abnormal running data; the corresponding abnormal running data is recorded as the abnormal maximum running data threshold of the corresponding time series parameter; and so on, the abnormal maximum running data threshold of each time series parameter is obtained. Based on big data, standard thresholds for each time series parameter are obtained. The difference between the standard threshold and the maximum abnormal operating data threshold is determined. If the maximum abnormal operating data threshold is less than or equal to the standard threshold, the standard threshold is replaced with the maximum abnormal operating data threshold of the time series parameter corresponding to the abnormal time point. Otherwise, the standard threshold and the standard threshold form the range of abnormal operating data thresholds corresponding to the abnormal time series parameter.
[0008] Furthermore, the process of obtaining the threshold range of runtime data for the corresponding time point of the time series parameters through the threshold range of abnormal runtime data includes: The system schedules the execution data of time-series parameters in real time, corresponding to the time points of those parameters. It also obtains the influence data of the influencing parameters at those time points, and then calculates the ratio of this influence data to the abnormal influence data at the corresponding abnormal time points for the same time-series parameter. This ratio is denoted as the coefficient of variation. x; simultaneously obtain the time difference between the time point that affects the data and the abnormal time point corresponding to the time series parameters. ; Based on the structural components, set the time decay coefficient Q of each timing parameter as it changes over time, and set the influence coefficient k of each timing parameter on the affected parameter; obtain the time correction coefficient through the change coefficient and the influence correction coefficient through the influence coefficient; then obtain the operating data threshold range of the timing parameter at the corresponding time point based on the time correction coefficient and the influence correction coefficient.
[0009] Furthermore, the process of obtaining the time correction factor and the impact correction factor includes: like When, the corresponding impact correction factor is Conversely, the corresponding impact correction coefficient is... Where k is a natural number greater than 0; t represents a time point, j represents a time point number, s represents a timing parameter, and i represents the number of the timing parameter corresponding to the timing parameter running data; and i and j are both natural numbers greater than 0. The time correction factor obtained therefrom is Where Q is a natural number greater than 0.
[0010] Furthermore, the process of obtaining the threshold range of the runtime data for the corresponding time point based on the time correction factor and the impact correction factor includes: If the running data is less than the maximum abnormal running data threshold, the corresponding adjustment coefficient is recorded as follows: Where y represents the standard threshold corresponding to the time series parameter, and ci is the running data corresponding to si; Conversely, the corresponding adjustment coefficient is Where 'a' represents the adjustment coefficient; The threshold value of the adjustment operation data at the corresponding time point of the time series parameter is obtained based on the adjustment coefficient at the corresponding time point of the time series parameter. , recorded as Among them, C si This is represented as the maximum abnormal running data threshold corresponding to the timing parameter si; If the adjusted running data threshold is greater than or equal to the standard threshold, then the running data threshold range of the time point corresponding to the time sequence parameter is marked as the standard threshold y; Conversely, the operating data threshold range corresponding to the timing parameters is marked as... .
[0011] Furthermore, the process of acquiring fault operation data for online fault monitoring of wind turbines based on operational data threshold ranges includes: The timing parameters are used to obtain the corresponding time point operation data based on the timing parameter operation data. The operation data is then compared with the corresponding operation data threshold range to obtain the fault operation data.
[0012] Furthermore, the process of obtaining abnormal time-series parameters of fault operation data, and determining the abnormality level of time points based on these abnormal time-series parameters, includes: The time series parameter running data containing faulty running data is recorded as abnormal time series parameter running data; the number of faults in the abnormal time series parameter running data is obtained; and then the ratio of the faulty data to the running data in the abnormal time series parameter running data is obtained, which is recorded as the fault ratio. The abnormality level time point is obtained based on the failure ratio, and then the abnormality level time point is processed.
[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention acquires the time-series parameter operation data of the wind turbine at corresponding time points in real time, maps the time-series parameter correlation chain of each time-series parameter in the time-series parameter operation data; and acquires the impact data corresponding to the influencing parameters in real time; performs anomaly analysis on the impact data to obtain abnormal impact data; acquires the standard threshold of the time-series parameters, and then calculates the abnormal operation data threshold range of the time-series parameters corresponding to abnormal time points based on the threshold range of the impact data of the abnormal impact data; acquires the operation data threshold range of the time-series parameters corresponding to the abnormal operation data threshold range; and then performs online fault monitoring of wind turbine faults based on the operation data threshold range to obtain fault operation data; acquires abnormal time-series parameter operation data of fault operation data, and judges the anomaly level of the time point based on the abnormal time-series parameter operation data to obtain the anomaly level time point; effectively improving the adaptability and accuracy of wind turbine anomaly judgment. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.
[0015] Figure 1 This is a flowchart of the present invention. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] Please see Figure 1 As shown, a method for wind turbine fault diagnosis and online monitoring based on time-series parameter analysis is presented, the method comprising the following steps: Step S1: Real-time acquisition of the time-series parameter operation data of the wind turbine at the corresponding time point, mapping the influence parameters of each time-series parameter in the time-series parameter operation data; and real-time acquisition of the influence data corresponding to the influence parameters; and anomaly analysis of the influence data to obtain abnormal influence data. Step S2: Obtain the standard threshold of the time series parameters, and then calculate the abnormal operation data threshold range of the time series parameters at the corresponding abnormal time points based on the obtained abnormal impact data threshold range; obtain the operation data threshold range of the time series parameters at the corresponding time points through the abnormal operation data threshold range; Step S3: Based on the operational data threshold range, perform online fault monitoring of the wind turbine to obtain fault operation data; Step S4: Obtain the abnormal timing parameter operation data of the fault operation data, and determine the abnormality level of the time point based on the abnormal timing parameter operation data to obtain the abnormality level time point.
[0018] Step S1 requires further clarification. The process of acquiring the time-series parameter operation data of the wind turbine at the corresponding time point in real time and mapping the time-series parameter correlation chain of each time-series parameter in the time-series parameter operation data includes: The timing parameters of the structural components of the wind turbine are set according to the structural components, and the corresponding acquisition period is set. The timing parameter operation data at the corresponding time points are collected according to the acquisition period, and mapped with the preset area number of the corresponding structural component. The timing parameters include at least rotational speed, vibration and temperature. The influence parameters of the timing parameters corresponding to different structural components are obtained based on big data, and the influence position of the influence parameters is obtained. The influence position is mapped with the corresponding influence parameter, and then associated with the corresponding timing parameter to generate a timing parameter association chain. In the above embodiments, it should be further explained that the structural components are used to represent the main structural components of the fan, including but not limited to impellers, casings, air inlets, supports, motors, pulleys, couplings, silencers, bearings, etc.; different structural components are located in different positions on the fan; and different sets of timing parameters are set for different structural components. For example, timing parameters such as vibration, impeller speed, blade stress, and impeller inlet and outlet pressure difference are set on the impeller; timing parameters such as motor speed, stator temperature, rotor temperature, and motor winding insulation resistance are set on the motor; furthermore, the influencing parameters are used to indicate how changes in the timing parameters will affect other parameters of the fan, excluding the timing parameters themselves; for example, changes in vibration on the impeller will affect the impeller... Uneven dust adheres to the impeller, thus the dust on the impeller becomes a vibration parameter affecting the impeller's corresponding time-series parameter set, with the affected location being inside the impeller. Similarly, changes in bearing speed affect the flowability of the bearing's internal grease. Abnormal grease flowability, in turn, exacerbates fluctuations in bearing speed. Therefore, bearing grease flowability is a time-series parameter affecting bearing speed, with the affected location being the bearing's internal lubrication cavity. Furthermore, an increase in motor stator temperature accelerates the aging rate of the insulation layer surrounding the motor windings. The degree of insulation aging affects the stator's heat dissipation efficiency; therefore, the degree of insulation aging is a time-series parameter affecting motor stator temperature, with the affected location being the area surrounding the motor stator windings.
[0019] It should be further explained that the impact data corresponding to the impact parameters is acquired in real time; the process of performing anomaly analysis on the impact data to obtain abnormal impact data includes: Each influencing location is obtained through a time-series parameter association chain, and the influencing locations are divided into several influencing location nodes. A data acquisition sensor is set at each influencing location node to collect the node influence data corresponding to the influencing location node. The node influence data is fused to obtain the influence data of the influence parameters corresponding to the influencing location, and mapped with the time point of the corresponding time-series parameter running data. Based on the threshold range of the impact data set for structural components, impact data that does not fall within the threshold range at the first time point is recorded as abnormal impact data; conversely, the impact data of each node of the impact data is analyzed, and the impact data of nodes that do not fall within the threshold range is recorded as abnormal node impact data. Furthermore, the impact data corresponding to the impact location of the abnormal node impact data is also recorded as abnormal impact data; otherwise, no processing is performed; and the time point corresponding to the abnormal impact data is marked as an abnormal time point. It should be further explained that the influencing location is usually a region or a surface. Therefore, several influencing location nodes are set, and corresponding acquisition sensors are set at these nodes to collect node influence data. The influence data of the influencing location is obtained through data fusion of the node influence data. The data fusion includes, but is not limited to, the mean, which can be obtained through other algorithms, such as the median filtering algorithm, which is suitable for scenarios with instantaneous impulse noise and can remove abnormally changing node influence data, thus improving the stability of the fused data; the Kalman filtering algorithm, which, for dynamically changing influence parameters, achieves smooth fusion and dynamic tracking of node influence data through a prediction-update iteration process, etc., will not be elaborated here.
[0020] Step S2 requires further refinement. The process of obtaining standard thresholds for time-series parameters and then extrapolating the abnormal operational data threshold range for the corresponding abnormal time points based on the impact data threshold range of the abnormal impact data includes: By obtaining the impact parameters of the abnormal impact data at abnormal time points, the corresponding time series parameters are obtained in the time series parameter running data. Then, the running data of the time series parameter at the abnormal time point is obtained and marked as abnormal running data. The corresponding abnormal running data is recorded as the abnormal maximum running data threshold of the corresponding time series parameter. The abnormal maximum running data threshold of each time series parameter is obtained in this way. Based on big data, standard thresholds for each time series parameter are obtained. The difference between the standard threshold and the maximum abnormal operating data threshold is determined. If the maximum abnormal operating data threshold is less than or equal to the standard threshold, the standard threshold is replaced with the maximum abnormal operating data threshold of the time series parameter corresponding to the abnormal time point. Otherwise, the standard threshold and the standard threshold form the range of abnormal operating data thresholds corresponding to the abnormal time series parameter. In the above embodiments, it should be further explained that for abnormal impact data, there may be different or the same corresponding abnormal time points. The impact parameter of the abnormal impact data at the abnormal time point is the time-series parameter corresponding to the time-series parameter running data; where the time-series parameter running number is denoted as {tj, c1, c2, ..., ci}; ci is the running data, and the time-series parameter corresponding to ci is si; the impact parameter corresponding to the time-series parameter si is ri; the impact data of the impact parameter is Rri; t is the time point, and j and i are both natural numbers that are not 0; if Rr1 is abnormal impact data, then the corresponding abnormal time point is obtained as t1, and the corresponding impact parameter is obtained as r1, and then... The corresponding time-series parameter running data is obtained as s1; the corresponding abnormal running data is obtained as c1; then the maximum abnormal running data threshold c is marked as c1; and the standard threshold corresponding to s1 is y; if c1≤y, then the corresponding abnormal running data range is [c1, y]; otherwise, c1 is marked as the maximum abnormal running data threshold; first, the first abnormal time point of the time-series parameter is obtained through the abnormal impact data, and the abnormal running data threshold range of the abnormal running data at the first abnormal time point is used to lock the time point and running data corresponding to the initial time-series parameter triggered by the abnormal impact, so as to prepare for the change of the threshold range corresponding to the subsequent changes of the time-series parameter at different times.
[0021] It should be further explained that the process of obtaining the threshold range of runtime data for the corresponding time point of the time series parameters through the threshold range of abnormal runtime data includes: The system schedules the execution data of time-series parameters in real time, corresponding to the time points of those parameters. It also obtains the influence data of the influencing parameters at those time points, and then calculates the ratio of this influence data to the abnormal influence data at the corresponding abnormal time points for the same time-series parameter. This ratio is denoted as the coefficient of variation. Simultaneously, obtain the time difference between the time points that affect the data and the abnormal time points corresponding to the time series parameters. ; Based on the structural components, set the time decay coefficient Q of each timing parameter as it changes over time, and set the influence coefficient k of each timing parameter on the affected parameter; obtain the time correction coefficient through the change coefficient and the influence correction coefficient through the influence coefficient; The process of obtaining the time correction factor and the impact correction factor includes: like When the time is right, the corresponding impact correction factor is: Conversely, the corresponding impact correction coefficient is: Where k is a natural number greater than 0; t represents a time point, j represents a time point number, s represents a timing parameter, and i represents the number of the timing parameter corresponding to the timing parameter running data; and i and j are both natural numbers greater than 0. The time correction factor obtained therefrom is Where Q is a natural number greater than 0.
[0022] In the above embodiments, it should be further explained that, since the abnormal impact data is the first abnormal data, the impact data corresponding to the time series parameter at the time point may be greater than or equal to the abnormal impact data or less than or equal to the abnormal impact data; wherein, the time difference is positively correlated with the impact coefficient of the time series parameter on the impact parameter. Since the time correction coefficient is always positively calculated by the exponential formula, that is, the value range is 0 to positive infinity; and it is positively correlated with the time difference, therefore, the larger the time difference, the greater the impact of the time series parameter on the impact parameter; for example, when the wind turbine has just started, the impact of the corresponding impact parameter is almost 0, and thus the time correction coefficient is also almost 0; furthermore, the value range of the time correction coefficient is (0, k], the larger the coefficient of variation, the closer the time correction coefficient is to k, representing a higher impact intensity; wherein, when Δx=1, If , it means there is no additional intensity correction.
[0023] Determine the difference between the running data at the corresponding time point of the time series parameter and the threshold of the maximum abnormal running data. Small, and then obtain the adjustment coefficient of the time series parameter at the corresponding time point through the time correction coefficient and the influence correction coefficient; If the running data is less than the maximum abnormal running data threshold, the corresponding adjustment coefficient is recorded as follows: Where y represents the standard threshold corresponding to the time series parameter, and ci is the running data corresponding to si; Conversely, the corresponding adjustment coefficient is Where 'a' represents the adjustment coefficient; The threshold value of the adjustment operation data at the corresponding time point of the time series parameter is obtained based on the adjustment coefficient at the corresponding time point of the time series parameter. , recorded as Among them, C si This is represented as the maximum abnormal running data threshold corresponding to the timing parameter si; If the adjusted running data threshold is greater than or equal to the standard threshold, then the running data threshold range of the time point corresponding to the time sequence parameter is marked as the standard threshold y; Conversely, the operating data threshold range corresponding to the timing parameters is marked as... .
[0024] In the above embodiments, it should be further explained that the threshold range of the time series parameters is dynamically updated in real time by using time difference and change coefficient, which accurately adapts to the dynamic correlation between time series parameters and influencing parameters, so that the threshold range evolves with time and the influence intensity is adjusted in real time, effectively improving the timeliness and accuracy of the judgment of the operating data.
[0025] Step S3 needs further clarification. The process of acquiring fault operation data for online fault monitoring of wind turbines based on operational data threshold range includes: The timing parameters are used to obtain the corresponding time point operation data based on the timing parameter operation data. The operation data is then compared with the corresponding operation data threshold range to obtain the fault operation data. If the corresponding running data is greater than or equal to the standard threshold y or there is a running data threshold range, the corresponding running data is recorded as the fault running data of the timing parameter at the time point; otherwise, no processing is performed.
[0026] Step S4: Obtain abnormal timing parameters of the fault operation data, and determine the abnormality level of the time point based on the abnormal timing parameters. The process of obtaining the abnormality level time point includes: The time series parameter running data containing faulty running data is recorded as abnormal time series parameter running data; the number of faults in the abnormal time series parameter running data is obtained; and then the ratio of the faulty data to the running data in the abnormal time series parameter running data is obtained, which is recorded as the fault ratio. The abnormality level time point is obtained based on the failure ratio, and then the abnormality level time point is processed. If the failure ratio is greater than or equal to 0.25, the time point with the corresponding abnormal level will be recorded as the first-level abnormal level time point. If the failure ratio is greater than 0.25 and less than or equal to 0.5, the time point corresponding to the abnormality level will be recorded as the second-level abnormality level time point. If the failure ratio is greater than 0.5 and less than or equal to 0.75, the time point corresponding to the abnormality level is recorded as the third-level abnormality time point. If the failure ratio is greater than 0.75 and less than or equal to 1, then the time point corresponding to the abnormality level is recorded as the fourth-level abnormality time point.
[0027] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A method for wind turbine fault diagnosis and online monitoring based on time-series parameter analysis, characterized in that, The method includes the following steps: Step S1: Real-time acquisition of the time-series parameter operation data of the wind turbine at the corresponding time point, mapping the influence parameters of each time-series parameter in the time-series parameter operation data; and real-time acquisition of the influence data corresponding to the influence parameters; and anomaly analysis of the influence data to obtain abnormal influence data. Step S2: Obtain the standard threshold of the time series parameters, and then calculate the abnormal operation data threshold range of the time series parameters at the corresponding abnormal time points based on the obtained abnormal impact data threshold range; obtain the operation data threshold range of the time series parameters at the corresponding time points through the abnormal operation data threshold range; Step S3: Based on the operational data threshold range, perform online fault monitoring of the wind turbine to obtain fault operation data; Step S4: Obtain the abnormal timing parameter operation data of the fault operation data, and determine the abnormality level of the time point based on the abnormal timing parameter operation data to obtain the abnormality level time point.
2. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 1, characterized in that, The process of acquiring real-time operating data of wind turbine time series parameters at corresponding time points and mapping the influencing parameters of each time series parameter in the operating data includes: The timing parameters of the structural components of the wind turbine are set according to the structural components, and the corresponding acquisition period is set. The timing parameter operation data at the corresponding time points are collected according to the acquisition period, and mapped with the preset area number of the corresponding structural component. The timing parameters include at least rotational speed, vibration and temperature. The influence parameters of the timing parameters corresponding to different structural components are obtained based on big data, and the influence position of the influence parameters is obtained. The influence position is mapped with the corresponding influence parameter, and then associated with the corresponding timing parameter to generate a timing parameter association chain.
3. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 2, characterized in that, Real-time acquisition of impact data corresponding to the influencing parameters; The process of obtaining anomalous impact data through anomaly analysis includes: Each influencing location is obtained through a time-series parameter association chain, and the influencing locations are divided into several influencing location nodes. A data acquisition sensor is set at each influencing location node to collect the node influence data corresponding to the influencing location node. The node influence data is fused to obtain the influence data of the influence parameters corresponding to the influencing location, and mapped with the time point of the corresponding time-series parameter running data. Based on the threshold range of the impact data set for structural components, impact data that does not fall within the threshold range at the first time point is recorded as abnormal impact data; conversely, the impact data of each node of the impact data is analyzed, and the impact data of nodes that do not fall within the threshold range is recorded as abnormal node impact data. Then, the impact data corresponding to the impact location of the abnormal node impact data is recorded as abnormal impact data; otherwise, no processing is performed; and the time point corresponding to the abnormal impact data is marked as an abnormal time point.
4. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 3, characterized in that, The process of obtaining standard thresholds for time-series parameters, and then extrapolating the threshold range of abnormal operational data for the corresponding abnormal time points based on the threshold range of the obtained abnormal impact data, includes: By obtaining the impact parameters of the abnormal impact data at abnormal time points, the corresponding time series parameters of the time series parameters are obtained, and then the running data of the time series parameters at the abnormal time points are obtained and marked as abnormal running data; the corresponding abnormal running data is recorded as the abnormal maximum running data threshold of the corresponding time series parameter; and so on, the abnormal maximum running data threshold of each time series parameter is obtained. Based on big data, standard thresholds for each time series parameter are obtained. The difference between the standard threshold and the maximum abnormal operating data threshold is determined. If the maximum abnormal operating data threshold is less than or equal to the standard threshold, the standard threshold is replaced with the maximum abnormal operating data threshold of the time series parameter corresponding to the abnormal time point. Otherwise, the standard threshold and the standard threshold form the range of abnormal operating data thresholds corresponding to the abnormal time series parameter.
5. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 4, characterized in that, The process of obtaining the threshold range of runtime data for a time point corresponding to a time series parameter by using the threshold range of abnormal runtime data includes: The system schedules the execution data of time-series parameters in real time, corresponding to the time points of those parameters. It also obtains the influence data of the influencing parameters at those time points, and then calculates the ratio of this influence data to the abnormal influence data at the corresponding abnormal time points for the same time-series parameter. This ratio is denoted as the coefficient of variation. Simultaneously, obtain the time difference between the time points that affect the data and the abnormal time points corresponding to the time series parameters. ; Based on the structural components, set the time decay coefficient Q of each timing parameter as it changes over time, and set the influence coefficient k of each timing parameter on the affected parameter; obtain the time correction coefficient through the change coefficient and the influence correction coefficient through the influence coefficient; then obtain the operating data threshold range of the timing parameter at the corresponding time point based on the time correction coefficient and the influence correction coefficient.
6. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 5, characterized in that, The process of obtaining the time correction factor and the impact correction factor includes: like When, the corresponding impact correction factor is Conversely, the corresponding impact correction coefficient is... Where k is a natural number greater than 0; t represents a time point, j represents a time point number, s represents a timing parameter, and i represents the number of the timing parameter corresponding to the timing parameter running data; and i and j are both natural numbers greater than 0. The time correction factor obtained therefrom is Where Q is a natural number greater than 0.
7. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 6, characterized in that, The process of obtaining the threshold range of runtime data for time points corresponding to time series parameters based on time correction factors and impact correction factors includes: If the running data is less than the maximum abnormal running data threshold, the corresponding adjustment coefficient is recorded as follows: Where y represents the standard threshold corresponding to the time series parameter, and ci is the running data corresponding to si; Conversely, the corresponding adjustment coefficient is Where 'a' represents the adjustment coefficient; The threshold value of the adjustment operation data at the corresponding time point of the time series parameter is obtained based on the adjustment coefficient at the corresponding time point of the time series parameter. , recorded as Among them, C si This is represented as the maximum abnormal running data threshold corresponding to the timing parameter si; If the adjusted running data threshold is greater than or equal to the standard threshold, then the running data threshold range of the time point corresponding to the time sequence parameter is marked as the standard threshold y; Conversely, the operating data threshold range corresponding to the timing parameters is marked as... .
8. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 7, characterized in that, The process of acquiring fault operation data for online fault monitoring of wind turbines based on operational data threshold ranges includes: The timing parameters are used to obtain the corresponding time point operation data based on the timing parameter operation data. The operation data is then compared with the corresponding operation data threshold range to obtain the fault operation data.
9. The wind turbine fault diagnosis and online monitoring method based on time-series parameter analysis according to claim 8, characterized in that, The process of obtaining abnormal time-series parameters of fault operation data, and determining the abnormality level of time points based on the abnormal time-series parameters, includes: The time series parameter running data containing faulty running data is recorded as abnormal time series parameter running data; the number of faults in the abnormal time series parameter running data is obtained; and then the ratio of the faulty data to the running data in the abnormal time series parameter running data is obtained, which is recorded as the fault ratio. The abnormality level time point is obtained based on the failure ratio, and then the abnormality level time point is processed.