A fan tower health management method based on a modal model and a storage medium

Abstract

The present application relates to the field of wind power equipment structural health monitoring, and discloses a wind turbine tower health management method based on a modal model. Vibration modal, environmental and operation data are collected by deploying a sensor array, three types of working conditions, namely, shutdown, low wind speed and rated wind speed, are divided in an initial operation period, a wind speed weighted average is used to generate a reference modal frequency and a mode shape matrix to establish a health baseline. In the operation stage, single working condition data are screened through a dynamic time window, a real-time modal frequency deviation ratio β and a mode shape matrix MAC value are calculated, and a fault risk level is output in combination with a multi-stage early warning threshold. The present application solves the problems of high false alarm rate and inaccurate damage positioning of traditional methods, and significantly improves the accuracy and efficiency of wind turbine tower structure safety monitoring.

Description

A method and storage medium for wind turbine tower health management based on modal modeling Technical Field

[0001] This invention relates to the technology of monitoring the structural health of electrical equipment, specifically to a method for dynamic assessment and early warning of the health status of wind turbine towers based on vibration modal analysis. Background Technology

[0002] Wind power, as a core representative of clean and renewable energy, has developed rapidly in recent years. With the increase in the number of wind turbines and the accumulation of operating time, the structural safety inspection of wind turbine towers has become increasingly prominent. Traditional inspection methods (such as manual visual inspection or periodic shutdown inspections) have significant drawbacks: inspection personnel must climb high-altitude towers, resulting in high operational risks and low efficiency; minor defects (such as cracks and corrosion) are easily overlooked, leading to the escalation of potential faults; the inspection process relies on subjective experience, lacking real-time performance and quantitative accuracy. Furthermore, wind turbine operation is affected by dynamic environmental factors (such as wind speed and temperature), and traditional methods cannot establish adaptive operating condition models, making early warning difficult. Summary of the Invention

[0003] In view of the technical defects and drawbacks existing in the prior art, embodiments of the present invention provide a wind turbine management method based on a modal model to overcome or at least partially solve the above problems, the specific solution of which is as follows:

[0004] As a first aspect of the present invention, a wind turbine health management method based on modal models is provided, comprising the following steps:

[0005] Step S1: Deploy a sensor array at a preset height position on the wind turbine tower to collect vibration modal data, environmental parameters, and operating status data in real time; wherein: the vibration modal data includes sway frequency and mode shape matrix; the environmental parameters include wind speed, wind direction, and temperature; and the operating status data includes power generation.

[0006] Step S2: Based on the data collected in Step S1, during the initial operating cycle after the wind turbine is newly installed or overhauled, the data is divided into three operating conditions: shutdown, low wind speed, and rated wind speed, according to the mapping relationship between wind speed and power generation. The data of each operating condition is weighted and averaged to generate the reference modal frequency and reference mode shape matrix, and a healthy baseline database of operating condition-modal mapping is established.

[0007] Step S3: During the wind turbine operation phase, based on the health baseline database of step S2, real-time data is collected through a dynamic time window, and datasets that meet the continuity of a single operating condition are selected; the deviation ratio of the real-time modal frequency relative to the reference value and the correlation coefficient between the real-time mode matrix and the reference mode matrix, i.e., the MAC value, are calculated.

[0008] Step S4: Based on the deviation ratio and MAC value calculation results from Step S3, match the preset multi-level early warning threshold table and output the alarm signal corresponding to the fault risk level.

[0009] Further, step S2 includes:

[0010] Step S2.1: Based on the power generation and wind speed data collected in step S1, perform dynamic operating condition classification:

[0011] When the power generation is zero and the wind speed is less than level 3, it is marked as a shutdown state;

[0012] When the power generation is greater than zero but less than the rated power, it is marked as a low wind speed state;

[0013] When the power generation capacity equals the rated power, it is marked as the rated wind speed state;

[0014] Specifically, when dividing the operating conditions, the jump points where the power generation power changes from 0 to non-0 and their adjacent data are removed;

[0015] Step S2.2: For each type of operating condition data, assign a weight related to wind speed, where the closer the wind speed is to the median of the typical wind speed range for the corresponding operating condition, the larger the weight.

[0016] The weights are used to calculate the weighted average of the sway frequencies to generate the reference modal frequencies for the corresponding working conditions; the weights are also used to calculate the weighted average of the column vectors of the mode shape matrix to generate the reference mode shape matrix for the corresponding working conditions.

[0017] Step S2.3: Store the reference modal frequencies and reference mode shape matrices of the three types of working conditions as a queryable mapping database according to the working condition labels, which will serve as the comparison benchmark for subsequent health status assessment.

[0018] Furthermore, S2.2 includes:

[0019] S2.21, Based on historical data or design parameters, determine the median reference wind speed v for each type of operating condition. mid ;

[0020] S2.22, For the sway frequency data of each type of operating condition, the reference modal frequency is generated according to the following rules:

[0021] Calculate the weights of each data point Where v is the real-time wind speed and σ is an adjustable standard deviation parameter; the reference modal frequency for this type of operating condition is output by weighting the swaying frequency with the aforementioned weights.

[0022] S2.23, the mode shape matrix Φ for each type of working condition n×m Generate the reference mode shape matrix according to the following rules:

[0023] The column vector of the mode shape for each mode Individually weighted, using the same weights w as S2.22, a column-vector weighted average is applied to output the reference mode shape matrix Φ. n×m ;

[0024] S2.24, store the reference modal frequencies and reference mode shape matrices into the database according to the working condition labels.

[0025] Furthermore, the correlation coefficient (MAC) between the real-time mode shape matrix and the reference mode shape matrix is ​​calculated, including:

[0026] Extract the reference mode matrix Φ for the current operating condition from the health baseline database. n×m And obtain the real-time mode shape matrix Φ′ of the real-time dynamic window. n×m ;

[0027] Calculate Φ′ n×m With Φ n×m Projection matrix: P=Φ′ n×m ×Φ n×m T , where P is an m×m matrix and T represents the transpose calculation;

[0028] Calculate Φ′ n×m The autocorrelation diagonal vector E B =diag(Φ′ n×m ×Φ′ n×m T ) and Φ n×m The autocorrelation diagonal vector E A =diag(Φ n×m T ×Φ n×m );

[0029] via E A With E B The outer product generates the denominator matrix Thus, matrix D is an m×n matrix;

[0030] Squaring the projection matrix P element by element: P squared =P.×P, where .× indicates element-wise division and multiplication;

[0031] Dividing the numerator matrix by the denominator matrix, we get M=P squared . / D, where . / represents element-wise division;

[0032] The resulting M matrix is ​​an m×m matrix. Taking the values ​​of its diagonal elements gives the MAC value, where MAC = {M...} 11 M 22 ,...,M mm}, M ii M is the correlation calculation result between the baseline i-th mode and the real-time i-th mode.ii The closer to 1, the higher the correlation; equal to 1, they are completely identical.

[0033] Further, step S3 includes:

[0034] Based on the operating characteristics and frequency of changes in operating conditions of the wind turbine, the time window parameters are adaptively set to ensure that the window covers at least one complete operating cycle.

[0035] Modal parameter calculation is triggered if and only if both the wind speed range and the power generation range within the window are within the preset operating condition range threshold.

[0036] Real-time detection of events switching between zero and non-zero power generation values; removal of event points and their adjacent data to form a clean dataset.

[0037] Synchronous execution based on the clean dataset: Calculate the relative ratio of the weighted average real-time modal frequency to the reference frequency; calculate the MAC value between the real-time mode shape matrix and the reference mode shape matrix;

[0038] The output relative proportions and MAC values ​​are stored in a time-series database, and statistical feature vectors are generated in response to periodic analysis commands. (The preset operating condition interval thresholds are dynamically generated through historical operating condition clustering, and the periodic analysis commands are triggered by an external maintenance system or a fixed time period.)

[0039] Furthermore, the method also includes: performing temperature compensation on the real-time modal frequencies, specifically including:

[0040] A mapping model between temperature and modal frequency is established based on historical data, and the temperature response coefficient is extracted.

[0041] When the real-time temperature deviates from the reference temperature used when the health baseline was established, the compensation process is activated.

[0042] The modal frequencies acquired in real time are corrected according to the following rules:

[0043]

[0044] f raw The original frequency measurement value is represented by α, the temperature response coefficient is represented by α, and the normalized difference between the real-time temperature and the reference temperature is represented by ΔT.

[0045] Furthermore, the hierarchical early warning decision in step S4 is executed through the following conditional trigger chain:

[0046] Receive the modal frequency relative ratio β and MAC value output in step S3;

[0047] Level 1 warning trigger: When β is between the baseline preset lower limit and upper limit of fluctuation and the MAC value is not lower than the morphological correlation threshold, a level 1 warning signal is output, indicating that there is no abnormality in the structure;

[0048] Level 2 warning trigger: When β is lower than the baseline preset fluctuation lower limit but does not reach the moderate degradation threshold, and the MAC value is lower than the morphological correlation threshold, a level 2 warning signal is output to indicate the risk of connector degradation.

[0049] Level 3 warning trigger: When β reaches the moderate degradation threshold but not the severe degradation threshold, and the MAC value is lower than the morphological stability threshold, a level 3 warning signal is output to indicate the initiation of structural damage.

[0050] Level 4 warning trigger: When β exceeds the severe degradation threshold and the dynamic amplitude exceeds the design margin, a Level 4 warning signal is output to indicate the risk of structural instability.

[0051] Furthermore, the method also includes: achieving multi-height tower compatibility through a dimension-independent modal analysis architecture, specifically including:

[0052] S8.1: Based on the distribution density of tower flanges and the target height range, dynamically configure the installation position and number of sensor arrays to ensure that the sensor topology covers the key modal deformation zone of the tower.

[0053] S8.2: Encapsulate the real-time acquired vibration modal data into a dynamic dimensional mode shape matrix Φ′ n×m This matrix serves as the real-time mode shape matrix for step S3, where n is the number of effective sensors in the current tower, determined by the deployment rules in S8.1; and m is the standard mode order, which is bound to the physical properties of the tower and is independent of its height.

[0054] S8.3: During the healthy baseline establishment phase, generate the reference mode matrix according to the same rules as in S8.2 to ensure that the dynamic dimension mode matrix is ​​aligned with the dimension of the reference mode matrix;

[0055] S8.4: When performing step S3, apply a dimension-independent MAC value calculation core to the dynamic dimension mode matrix and the reference mode matrix, so that the output index depends only on modal correlation rather than physical size.

[0056] Furthermore, the method also includes: enhancing fault location capabilities through a local-global collaborative diagnostic mechanism, specifically including:

[0057] S9.1: Based on the stress distribution model of the tower structure, identify high-damage-risk sections as local diagnostic target areas;

[0058] S9.2: In the data acquisition step S1, the sensor subset corresponding to the local diagnostic target area is activated synchronously to generate the target area vibration mode matrix Ψ. k×m, where: k is the number of effective sensors in the target area, and k < n; m is the modal order same as the global matrix;

[0059] S9.3: When performing the health assessment in step S3, synchronously calculate:

[0060] The global MAC value is based on the full tower mode shape matrix Φ n×m for a complete modal analysis;

[0061] The local diagnostic target area MAC value is based on the target area mode shape matrix Ψ k×m for the same modal calculation process analysis;

[0062] S9.4: When the deterioration rate of the local diagnostic target area MAC value relative to the full tower MAC value exceeds the preset sensitivity threshold, generate a target area damage location signal;

[0063] S9.5: Perform a logical OR operation on the target area damage location signal and the warning level in step S4 to trigger a targeted maintenance instruction.

[0064] As a second aspect of the present invention, there is provided a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a computer, the computer is made to execute the modal model-based fan tower health management method as described in any one of the above.

[0065] The present invention has the following beneficial effects:

[0066] 1. Improve the detection accuracy and real-time performance, can identify minor structural changes (such as loosening, corrosion), and reduce the missed detection rate;

[0067] 2. Reduce manual intervention and shutdown costs, and improve the operation safety of the fan through automatic warning;

[0068] 3. The system has strong robustness, is applicable to tower barrels of different heights and environments, and supports long-term structural health management;

[0069] 4. Through temperature compensation and wind speed weighting, enhance the comparability of data and reduce false alarms. Description of the Drawings

[0070] FIG. 1 is a schematic flowchart of a modal model-based fan tower health management method provided by an embodiment of the present invention. Detailed Embodiment

[0071] 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 a part of the present invention, and not all of the 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.

[0072] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0073] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0074] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0075] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0076] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0077] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information all comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example: appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely locating a specific individual.

[0078] To address at least one of the technical problems existing in the aforementioned related technologies, the present invention provides a wind turbine tower health management method based on modal models. Figure 1 is a schematic flowchart of a wind turbine tower health management method based on modal models provided in an embodiment of the present invention, including the following steps:

[0079] Step S1: Deploy a sensor array at a preset height position on the wind turbine tower to collect vibration modal data, environmental parameters, and operating status data in real time; wherein: the vibration modal data includes sway frequency and mode shape matrix; the environmental parameters include wind speed, wind direction, and temperature; and the operating status data includes power generation.

[0080] Step S2: Based on the data collected in Step S1, during the initial operating cycle after the wind turbine is newly installed or overhauled, the data is divided into three operating conditions: shutdown, low wind speed, and rated wind speed, according to the mapping relationship between wind speed and power generation. The data of each operating condition is weighted and averaged to generate the reference modal frequency and reference mode shape matrix, and a healthy baseline database of operating condition-modal mapping is established.

[0081] Step S3: During the wind turbine operation phase, based on the health baseline database of step S2, real-time data is collected through a dynamic time window, and datasets that meet the continuity of a single operating condition are selected; the deviation ratio of the real-time modal frequency relative to the reference value and the correlation coefficient between the real-time mode matrix and the reference mode matrix, i.e., the MAC value, are calculated.

[0082] Step S4: Based on the deviation ratio and MAC value calculation results from Step S3, match the preset multi-level early warning threshold table and output the alarm signal corresponding to the fault risk level.

[0083] This invention integrates vibration mode (frequency + mode shape), environmental parameters (wind speed / temperature), and operating status (power) data to overcome the limitations of traditional methods that rely on a single vibration signal. It establishes a benchmark database according to operating conditions (shutdown / low wind speed / rated wind speed) to eliminate misjudgments caused by environmental interference and improve the accuracy of health assessment. Finally, it uses the modal frequency deviation ratio (quantifying stiffness degradation) and MAC value (mode shape correlation) for collaborative judgment, which can identify minor damage earlier than traditional methods that only monitor frequency.

[0084] In some embodiments, step S2 includes:

[0085] Step S2.1: Based on the power generation and wind speed data collected in step S1, perform dynamic operating condition classification:

[0086] When the power generation is zero and the wind speed is less than level 3, it is marked as a shutdown state;

[0087] When the power generation is greater than zero but less than the rated power, it is marked as a low wind speed state;

[0088] When the power generation capacity equals the rated power, it is marked as the rated wind speed state;

[0089] Specifically, when dividing the operating conditions, the jump points where the power generation power changes from 0 to non-0 and their adjacent data are removed;

[0090] Step S2.2: For each type of operating condition data, assign a weight related to wind speed, where the closer the wind speed is to the median of the typical wind speed range for the corresponding operating condition, the larger the weight.

[0091] The weights are used to calculate the weighted average of the sway frequencies to generate the reference modal frequencies for the corresponding working conditions; the weights are also used to calculate the weighted average of the column vectors of the mode shape matrix to generate the reference mode shape matrix for the corresponding working conditions.

[0092] Step S2.3: Store the reference modal frequencies and reference mode shape matrices of the three types of working conditions as a queryable mapping database according to the working condition labels, which will serve as the comparison benchmark for subsequent health status assessment.

[0093] In the above embodiments, power jump point filtering (such as power from 0 to non-0) is used to avoid polluting the benchmark library with transitional operating condition data and improve data purity; data close to the median of typical wind speed are given higher weights so that the benchmark modal parameters are closer to the actual operating conditions (such as tower flexibility is more significant at low wind speeds).

[0094] In some embodiments, S2.2 includes:

[0095] S2.21, Based on historical data or design parameters, determine the median reference wind speed v for each type of operating condition. mid ;

[0096] S2.22, For the sway frequency data of each type of operating condition, the reference modal frequency is generated according to the following rules:

[0097] Calculate the weights of each data point Where v is the real-time wind speed and σ is an adjustable standard deviation parameter; the reference modal frequency for this type of operating condition is output by weighting the swaying frequency with the aforementioned weights.

[0098] S2.23, the mode shape matrix Φ for each type of working conditionn×m Generate the reference mode shape matrix according to the following rules:

[0099] The column vector of the mode shape for each mode Individually weighted, using the same weights w as S2.22, a column-vector weighted average is applied to output the reference mode shape matrix Φ. n×m ;

[0100] S2.24, store the reference modal frequencies and reference mode shape matrices into the database according to the working condition labels.

[0101] In the above embodiments, an exponential weighting function is used. Strengthen the contribution of typical wind speed data, suppress the noise impact of extreme wind speeds (such as gusts), improve the stability of the reference frequency, independently weight the column vector of each mode, preserve the physical characteristics of the mode shape (such as the shape of the first bending mode), and avoid the shape distortion caused by overall averaging.

[0102] In some embodiments, calculating the correlation coefficient (MAC) between the real-time mode shape matrix and the reference mode shape matrix includes:

[0103] Extract the reference mode matrix Φ for the current operating condition from the health baseline database. n×m And obtain the real-time mode shape matrix Φ′ of the real-time dynamic window. n×m ;

[0104] Calculate Φ′ n×m With Φ n×m Projection matrix: P=Φ′ n×m ×Φ n×m T , where P is an m×m matrix and T represents the transpose calculation;

[0105] Calculate Φ′ n×m The autocorrelation diagonal vector E B =diag(Φ′ n×m ×Φ′ n×m T ) and Φ n×m The autocorrelation diagonal vector E A =diag(Φ n×m T ×Φ n×m );

[0106] via E A With E B The outer product generates the denominator matrix Thus, matrix D is an m×n matrix;

[0107] Squaring the projection matrix P element by element: P squared =P.×P, where .× indicates element-wise division and multiplication;

[0108] Dividing the numerator matrix by the denominator matrix, we get M=P squared . / D, where . / represents element-wise division;

[0109] The resulting M matrix is ​​an m×m matrix. Taking the values ​​of its diagonal elements gives the MAC value, where MAC = {M...} 11 M 22 ,...,M mm}, M ii M is the correlation calculation result between the baseline i-th mode and the real-time i-th mode. ii The closer to 1, the higher the correlation; equal to 1, they are completely identical.

[0110] In the above embodiments, the projection matrix P and the autocorrelation vector E are used... A / E B The element-wise operation decouples the MAC value calculation from the number of sensors n, supports dynamic sensor deployment (such as when a sensor is damaged), and only requires that the modal order m be consistent (such as the first 3 modes inherent in the tower), thus solving the problem of data incomparability caused by the difference in the number of sensors for towers of different heights.

[0111] In some embodiments, step S3 includes:

[0112] Based on the operating characteristics and frequency of changes in operating conditions of the wind turbine, the time window parameters are adaptively set to ensure that the window covers at least one complete operating cycle.

[0113] Modal parameter calculation is triggered if and only if both the wind speed range and the power generation range within the window are within the preset operating condition range threshold.

[0114] Real-time detection of events switching between zero and non-zero power generation values; removal of event points and their adjacent data to form a clean dataset.

[0115] Synchronous execution based on the clean dataset: Calculate the relative ratio of the weighted average real-time modal frequency to the reference frequency, and calculate the MAC value of the real-time mode shape matrix and the reference mode shape matrix;

[0116] The output relative proportions and MAC values ​​are stored in a time series database, and statistical feature vectors are generated in response to periodic analysis commands.

[0117] The preset operating condition range threshold is dynamically generated through historical operating condition clustering, and the periodic analysis command is triggered by an external maintenance system or a fixed time period.

[0118] In the above embodiments, by using an adaptive window to cover the complete operating cycle (such as a 10-minute low wind speed steady state), misjudgments caused by mixed operating data are avoided, power switching event points (such as shutdown → startup) are removed in real time, transient vibration interference is eliminated, and the dataset is purified to reduce MAC calculation errors.

[0119] In some embodiments, the method further includes: performing temperature compensation on the real-time modal frequencies, specifically including:

[0120] A mapping model between temperature and modal frequency is established based on historical data, and the temperature response coefficient is extracted.

[0121] When the real-time temperature deviates from the reference temperature used when the health baseline was established, the compensation process is activated.

[0122] The modal frequencies acquired in real time are corrected according to the following rules:

[0123]

[0124] f raw The original frequency measurement value is represented by α, the temperature response coefficient is represented by ΔT, and the normalized difference between the real-time temperature and the reference temperature is represented by ΔT.

[0125] The corrected f adj Enter step S3 and verify whether the frequency deviation converges after compensation.

[0126] In the above embodiments, the modal frequency is corrected by the temperature response coefficient α (e.g., α≈-0.03% / ℃ for steel), eliminating measurement deviations caused by temperature differences and reducing seasonal false alarm rates. Triggering is only performed when the temperature difference ΔT exceeds the limit (e.g., >5℃), avoiding unnecessary computational overhead.

[0127] In some embodiments, the hierarchical early warning decision in step S4 is executed through the following conditional trigger chain:

[0128] Receive the modal frequency relative ratio β and MAC value output in step S3;

[0129] Level 1 warning trigger: When β is between the baseline preset lower limit and upper limit of fluctuation and the MAC value is not lower than the morphological correlation threshold, a level 1 warning signal is output, indicating that there is no abnormality in the structure;

[0130] Level 2 warning trigger: When β is lower than the baseline preset fluctuation lower limit but does not reach the moderate degradation threshold, and the MAC value is lower than the morphological correlation threshold, a level 2 warning signal is output to indicate the risk of connector degradation.

[0131] Level 3 warning trigger: When β reaches the moderate degradation threshold but not the severe degradation threshold, and the MAC value is lower than the morphological stability threshold, a level 3 warning signal is output to indicate the initiation of structural damage.

[0132] Level 4 warning trigger: When β exceeds the severe degradation threshold and the dynamic amplitude exceeds the design margin, a Level 4 warning signal is output to indicate the risk of structural instability.

[0133] The table below shows a specific form of early warning at various levels provided in the embodiments of the present invention.

[0134]

[0135] In the above embodiments, the second-level warning indicates the risk of connector deterioration, such as loose bolts (improving sensitivity); the third-level warning indicates the initiation of structural damage, such as crack initiation (improving early detection rate); and the fourth-level warning indicates the risk of structural instability. By combining frequency deviation (stiffness deterioration) and MAC value (morphological distortion), the damage to connectors and the damage to the structural body are distinguished, guiding targeted maintenance.

[0136] In some embodiments, the method further includes: achieving multi-height tower compatibility through a dimension-independent modal analysis architecture, specifically including:

[0137] S8.1: Based on the distribution density of tower flanges and the target height range, dynamically configure the installation position and number of sensor arrays to ensure that the sensor topology covers the key modal deformation zone of the tower.

[0138] S8.2: Encapsulate the real-time acquired vibration modal data into a dynamic dimensional mode shape matrix Φ n×m This matrix serves as the real-time mode shape matrix for step S3, where n is the number of effective sensors in the current tower, determined by the deployment rules in S8.1; and m is the standard mode order, which is bound to the physical properties of the tower and is independent of its height.

[0139] S8.3: During the healthy baseline establishment phase, generate the reference mode matrix according to the same rules as in S8.2 to ensure that the dynamic dimension mode matrix is ​​aligned with the dimension of the reference mode matrix;

[0140] S8.4: When performing step S3, adjust the dynamic dimension mode matrix Φ n×m The application of dimension-independent MAC value calculation core with the reference mode matrix makes the output index depend only on modal correlation rather than physical dimensions.

[0141] The number of sensors (n) varies depending on the tower height, resulting in different mode shape matrices (Φ). n×m The row dimension changes, which traditional matrix operations cannot directly compare. The core of this invention uses mathematical reconstruction to make the MAC value calculation depend only on the shape similarity of the mode shape, and is independent of the number of sensors n (i.e., the tower height).

[0142] MAC is calculated as follows:

[0143] ;

[0144] The scalar MAC value is calculated independently for each mode i, getting rid of the constraint of the matrix dimension n. Here, MACi represents the criterion value of the i-th mode, that is, the calculated output, which reflects the similarity between the real-time vibration mode and the reference vibration mode. is the reference vibration mode vector, representing the standardized vibration form of the i-th mode in the healthy baseline database, that is, the i-th column of Φ n×m ; is the real-time vibration mode vector, representing the standardized vibration form of the i-th mode in the current dynamic window, that is, the i-th column of Φ'. n×m ;

[0145] In the above embodiment, the number of rows n of the vibration mode matrix adapts to the number of sensors (for example, n = 8 for a 70m tower barrel and n = 12 for a 140m tower barrel), while m is fixed (for example, the first 3 modes are analyzed), enabling the same analysis kernel to be shared by tower barrels of different heights. The core formula gets rid of the dependence on n, making the MAC results of 70m / 140m tower barrels directly comparable.

[0146] In some embodiments, the method further includes: enhancing the fault location ability through a local-global collaborative diagnosis mechanism, specifically including:

[0147] S9.1: Based on the tower barrel structure stress distribution model, identify the high-damage-risk section as the local diagnosis target area;

[0148] S9.2: In the data acquisition of step S1, synchronously activate the sensor subset corresponding to the target area to generate the target area vibration mode matrix Ψ k×m , where: k is the number of effective sensors in the target area, k < n; m is the same mode order as the global matrix;

[0149] S9.3: When performing the health assessment in step S3, synchronously calculate:

[0150] The global MAC value is based on the complete modal analysis of the full-tower vibration mode matrix Φ n×m ;

[0151] The local MAC value is based on the same modal calculation process analysis of the target area vibration mode matrix Ψ k×m ;

[0152] S9.4: When the deterioration rate of the target area MAC value relative to the full-tower MAC value exceeds the preset sensitivity threshold, generate a target area damage location signal;

[0153] S9.5: Perform a logical OR operation on the target area damage location signal and the warning level in step S4 to trigger a targeted maintenance instruction.

[0154] The above embodiments are generally applicable, based on the target area MAC degradation rate (e.g., flange area Ψ). k×m The MAC decreases faster than the global rate, improving fault location accuracy from "tower level" to "flange level". A subset of local sensors focuses on high-stress areas (such as the bottom of the tower), amplifying the MAC value change of microcracks by many times.

[0155] This invention also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps of any of the modal model-based wind turbine tower health management methods described in the above embodiments. The computer-readable storage medium can be volatile or non-volatile.

[0156] This invention also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described wind turbine tower health management method based on modal model.

[0157] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0158] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0159] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0160] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0161] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0162] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0163] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0164] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0165] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0166] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

[0167] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wind turbine health management method based on modal models, characterized in that, Includes the following steps: Step S1: Deploy a sensor array at a preset height on the wind turbine tower to collect vibration modal data, environmental parameters, and operational status data in real time; wherein: the vibration modal data includes swaying frequency and mode shape matrix; the environmental parameters include wind speed, wind direction, and temperature; and the operational status data includes power generation. Step S2: Based on the data collected in Step S1, during the initial operating cycle after a new installation or major overhaul of the wind turbine, divide the data into three operating conditions according to the mapping relationship between wind speed and power generation: shutdown, low wind speed, and rated wind speed; perform weighted averaging on the data for each operating condition to generate a reference modal frequency and reference mode shape matrix, establishing a healthy baseline database for the operating condition-modal mapping. Step S3: During the wind turbine operation phase, based on the healthy baseline database from Step S2, through dynamic... The method involves collecting real-time data within a time window and filtering datasets that meet the requirements of a single operating condition. It calculates the deviation ratio of the real-time modal frequency relative to the reference value, as well as the correlation coefficient (MAC) between the real-time mode shape matrix and the reference mode shape matrix. Step S4: Based on the deviation ratio and MAC value calculation results from step S3, it matches a preset multi-level early warning threshold table and outputs an alarm signal corresponding to the fault risk level. The method further includes: achieving multi-height tower compatibility through a dimension-independent modal analysis architecture, specifically including: S8.1: Dynamically configuring the installation position and number of sensor arrays based on the tower flange distribution density and target height range, ensuring the sensor topology covers the key modal deformation zone of the tower; S8.2: Encapsulating the real-time collected vibration modal data into a dynamic dimension mode shape matrix Φ′. n×m This matrix serves as the real-time mode shape matrix for step S3, where n is the number of effective sensors in the current tower and m is the standard mode order; S8.3: During the healthy baseline establishment phase, a reference mode shape matrix is ​​generated according to the same rules as in S8.2 to ensure that the dynamic dimension mode shape matrix and the reference mode shape matrix are dimensionally aligned; S8.4: When executing step S3, a dimension-independent MAC value calculation core is applied to the dynamic dimension mode shape matrix and the reference mode shape matrix, so that the output index depends only on modal correlation rather than physical dimensions.

2. The wind turbine tower health management method based on modal model according to claim 1, characterized in that, Step S2 includes: Step S2.1: Based on the power generation and wind speed data collected in Step S1, perform dynamic operating condition classification: when the power generation is zero and the wind speed is less than level 3, mark it as a shutdown state; when the power generation is greater than zero and less than the rated power, mark it as a low wind speed state; when the power generation is equal to the rated power, mark it as a rated wind speed state; wherein, during the operating condition classification, the jump point of power generation from 0 to non-zero and its adjacent data are removed; Step S2.2: For each type of operating condition data, assign a weight related to wind speed, wherein the closer the wind speed is to the median of the typical wind speed range of the corresponding operating condition, the larger the weight; use the weight to perform a weighted average of the sway frequency to generate the reference modal frequency of the corresponding operating condition; use the weight to perform a weighted average of the column vectors of the mode shape matrix to generate the reference mode shape matrix of the corresponding operating condition; Step S2.3: Store the reference modal frequencies and reference mode shape matrices of the three types of operating conditions in a queryable mapping database according to the operating condition labels, as a comparison benchmark for subsequent health status assessment.

3. The wind turbine tower health management method based on modal model according to claim 2, characterized in that, S2.2 includes: S2.21, determining the median reference wind speed v for each type of operating condition based on historical data or design parameters. mid S2.22, For the sway frequency data of each type of working condition, generate the reference modal frequency according to the following rules: Calculate the weight of each data point. Where v is the real-time wind speed and σ is an adjustable standard deviation parameter; the swaying frequency is weighted and averaged using the aforementioned weights to output the reference modal frequency S2.23 for this type of working condition, and the mode shape matrix Φ for each type of working condition is... n×m Generate the reference mode shape matrix according to the following rules: for each mode, the mode shape column vector... Individually weighted, using the same weights w as S2.22, a column-vector weighted average is applied to output the reference mode shape matrix Φ. n×m S2.24, store the reference modal frequencies and reference mode shape matrices into the database according to the working condition labels.

4. The wind turbine tower health management method based on modal model according to claim 2, characterized in that, Calculating the correlation coefficient (MAC) between the real-time mode shape matrix and the reference mode shape matrix includes: extracting the reference mode shape matrix Φ for the current operating condition from the health baseline database. n×m And obtain the real-time mode shape matrix Φ′ of the real-time dynamic window. n×m ; Calculate Φ′ n×m With Φ n×m Projection matrix: P=Φ′ n×m ×Φ n×m T Where P is an m×m projection matrix, and T denotes transpose calculation; calculate Φ′ n×m The autocorrelation diagonal vector E B =diag(Φ′ n×m ×Φ′ n×m T ) and Φ n×m The autocorrelation diagonal vector E A =diag(Φ n×m T ×Φ n×m ); via E A With E B The outer product generates the denominator matrix The resulting matrix D is an m×n matrix; the projection matrix P is squared element-wise: P squared =P.×P, where .× indicates element-wise division and multiplication; dividing the numerator matrix by the denominator matrix yields M=P. squared . / D, where . / represents element-wise division; the resulting M matrix is ​​an m×m matrix, and the values ​​of its diagonal elements are the MAC values, MAC={M 11 M 22 ,...,M mm }, M ii M is the correlation calculation result between the baseline i-th mode and the real-time i-th mode. ii The closer to 1, the higher the correlation; equal to 1, they are completely identical.

5. The wind turbine tower health management method based on modal model according to claim 1, characterized in that, Step S3 includes: adaptively setting time window parameters based on the wind turbine's operating characteristics and the frequency of operating condition changes to ensure that the window covers at least one complete operating condition cycle; triggering modal parameter calculation if and only if both the wind speed range and the power generation range within the window are within the preset operating condition range threshold; real-time detection of zero-value and non-zero-value switching events in power generation, removing event points and their adjacent data to form a clean dataset; synchronously executing the following based on the clean dataset: calculating the relative ratio of the weighted average real-time modal frequency to the reference frequency; calculating the MAC value of the real-time mode shape matrix and the reference mode shape matrix; storing the output relative ratio and MAC value in the time series database, and generating statistical feature vectors in response to periodic analysis commands.

6. The wind turbine tower health management method based on modal model according to claim 1, characterized in that, The method further includes: temperature compensation for real-time modal frequencies, specifically including: establishing a mapping model between temperature and modal frequencies based on historical data, and extracting temperature response coefficients; activating the compensation process when the deviation between the real-time temperature and the reference temperature at which the healthy baseline was established exceeds a preset threshold; and correcting the real-time acquired modal frequencies according to the following rules: ;f raw The original frequency measurement value is represented by α, the temperature response coefficient is represented by α, and the normalized difference between the real-time temperature and the reference temperature is represented by ΔT.

7. The wind turbine tower health management method based on modal model according to claim 1, characterized in that, The graded early warning decision in step S4 is executed through the following condition trigger chain: receiving the modal frequency relative ratio β and MAC value output in step S3; Level 1 early warning trigger: when β is between the baseline preset lower limit and upper limit of fluctuation and the MAC value is not lower than the morphological correlation threshold, a level 1 early warning signal is output, indicating that there is no abnormality in the structure; Level 2 warning trigger: When β is lower than the baseline preset fluctuation lower limit but not the moderate degradation threshold, and the MAC value is lower than the morphological correlation threshold, a Level 2 warning signal is output, indicating the risk of connector degradation; Level 3 warning trigger: When β reaches the moderate degradation threshold but not the severe degradation threshold, and the MAC value is lower than the morphological stability threshold, a Level 3 warning signal is output, indicating the initiation of structural damage; Level 4 warning trigger: When β exceeds the severe degradation threshold and the dynamic amplitude exceeds the design margin, a Level 4 warning signal is output, indicating the risk of structural instability.

8. The wind turbine tower health management method based on modal model according to claim 1, characterized in that, The method further includes: enhancing the fault location ability through a local-global collaborative diagnosis mechanism, specifically including: S9.1: Based on the tower structure stress distribution model, identify the high-damage-risk section as the local diagnosis target area; S9.2: In the data acquisition of step S1, synchronously activate the corresponding sensor subset in the target area to generate the target area mode shape matrix Ψ k×m , where: k is the number of effective sensors in the target area, k < n; m is the same modal order as the global matrix; S9.3: When performing the health assessment in step S3, synchronously calculate: the global MAC value is based on the full-tower mode shape matrix Φ n×m for the complete modal analysis; the local MAC value is based on the same modal calculation process analysis of the target area mode shape matrix Ψ k×m ; S9.4: When the deterioration rate of the target area MAC value relative to the full-tower MAC value exceeds the preset sensitivity threshold, generate a target area damage location signal; S9.5: Perform a logical OR operation on the target area damage location signal and the warning level in step S4 to trigger a targeted maintenance instruction.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a computer, causes the computer to perform the wind turbine tower health management method based on a modal model as described in any one of claims 1 to 8.

Images