A method, system and device for analyzing the state of a marine wind turbine transmission chain under wind and wave disturbance, and a storage medium

Abstract

The application discloses a kind of offshore wind turbine transmission chain state analysis method, system, device and storage medium under wind wave disturbance, belong to offshore wind generating set state analysis technical field.Method includes obtaining including the image sequence of target area and reference area on wind turbine transmission chain;In image sequence, select reference frame, any frame after reference frame in image sequence as current frame;Calculate the transformation matrix of current frame to reference frame, carry out coordinate compensation to current target feature point, obtain compensation target feature point;According to compensation target feature point, the motion parameter of target area is calculated, and the running state of wind turbine transmission chain is determined according to motion parameter;Calculate the visual measurement reliability including image quality score, compensation consistency score and parameter estimation score, and the output mode of the running state of wind turbine transmission chain is determined according to visual measurement reliability.The application realizes the accurate analysis to offshore wind turbine transmission chain state.

Description

Technical Field

[0001] This invention belongs to the field of offshore wind turbine state analysis technology, specifically relating to a method, system, device, and storage medium for analyzing the state of offshore wind turbine transmission chains under wind and wave disturbances. Background Technology

[0002] The drivetrain of a wind turbine generator typically includes the main shaft, main bearings, gearbox, high-speed shaft, coupling, generator shaft, and adjacent support structures. The motion of the drivetrain directly affects the operational stability of the wind turbine generator. Visual methods can be used for non-contact monitoring of exposed shaft sections to obtain rotational speed, radial relative displacement, or axial relative displacement. However, in the actual nacelle environment, the camera, nacelle frame, and target shaft section may be simultaneously affected by turbine vibration, tower movement, yaw action, and support micro-motions, easily resulting in superimposed field-of-view motions in the visual observation results. For offshore wind turbine generators, high humidity, salt spray, and condensation reduce the quality of the target's visual features. Wave loads causing motion in fixed foundations, towers, or floating support structures can also superimpose field-of-view motions unrelated to the local motion of the drivetrain shaft system, further reducing the reliability of visual measurement results.

[0003] Visual measurement features non-contact and multi-feature parallel observation, enabling the acquisition of rotational speed or dynamic displacement using rotating textures, artificial markers, and edge positions. Existing visual rotational speed measurement schemes can determine instantaneous rotational speed based on image time-series periods, while existing monocular dynamic displacement measurement schemes can determine multi-directional dynamic responses based on the edge or marker projection of rotating targets. These schemes provide a foundation for visual monitoring of wind turbine generator drivetrain shaft systems.

[0004] In offshore wind turbine nacelle or drivetrain test rigs, image acquisition devices and target shaft segments are typically mounted near the vibrating frame, main bearing housing, or gearbox. Wave loads causing motion in the fixed foundation, tower, nacelle, or floating support structure, along with localized changes in viewing angle and minor movements of the camera mount during yaw, result in the observed motion of the target shaft segment in the image encompassing both localized motion of the shaft segment relative to the supporting structure and a combined field-of-view drift. Directly calculating displacement using uncompensated image features easily leads to the superimposing of the combined motion of the nacelle, tower, supporting structure, or camera mount into the target shaft segment's motion result.

[0005] Furthermore, low illumination, dust, oil mist, condensation, surface reflection, local occlusion, and motion blur within the wind turbine nacelle can cause the effective quantity and fitting stability of target edge or texture features to change over time. Existing solutions typically output rotational speed or displacement parameters, but they do not jointly map image quality, reference compensation residuals, and parameter estimation residuals to a confidence level corresponding to the motion parameters, nor do they develop targeted result rejection and acquisition adjustment mechanisms based on sources of low confidence. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, system, device and storage medium for analyzing the state of offshore wind turbine transmission chains under wind and wave disturbances. It unifies the current frame to the reference frame, clearly outputs the motion parameters of the target area relative to the reference area, and evaluates the reliability of the motion parameters, thereby achieving accurate analysis of the state of offshore wind turbine transmission chains.

[0007] This invention provides the following technical solution:

[0008] In the first aspect, a method for analyzing the state of offshore wind turbine transmission chains under wind and wave disturbances is provided, including: acquiring an image sequence including a target area and a reference area on the wind turbine transmission chain;

[0009] A reference frame is selected in the image sequence. Reference reference feature points within the reference region and reference target feature points within the target region are selected in the reference frame image. Any frame after the reference frame in the image sequence is taken as the current frame. In the current frame image, the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point are determined.

[0010] Calculate the transformation matrix from the current frame to the reference frame based on the reference feature points and the current reference feature points;

[0011] Based on the transformation matrix from the current frame to the reference frame, coordinate compensation is performed on the current target feature points to obtain the compensated target feature points;

[0012] The motion parameters of the target area are calculated based on the compensation target feature points, and the operating status of the wind turbine drive chain is determined based on the motion parameters.

[0013] Based on the transformation matrix from the current frame to the reference frame, the reference reference feature points, the current reference feature points, and the current frame image, the visual measurement confidence level, including image quality score, compensation consistency score, and parameter estimation score, is calculated. The output mode of the wind turbine drive chain's operating status is then determined based on the visual measurement confidence level.

[0014] As an optional technical solution of the present invention, the target area is located on the target shaft segment on the wind turbine drive chain, and the target shaft segment is the main shaft, high-speed shaft or exposed section of the coupling in the wind turbine drive chain.

[0015] The reference area is located on a reference structure on the wind turbine drive chain. The reference structure is a main bearing housing, gearbox housing, generator end cover, nacelle frame or other support components that are fixedly connected to the target shaft segment.

[0016] As an optional technical solution of the present invention, the transformation matrix from the current frame to the reference frame is represented as follows:

[0017] ;

[0018] in, This represents the transformation matrix from the current frame to the reference frame. Represents a two-dimensional coordinate transformation matrix. Represents a robust loss function. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents Euclidean distance. Represents the current set of reference feature points. This indicates that through optimization Obtain the minimum value.

[0019] As an optional technical solution of the present invention, the compensation target feature point is represented as follows:

[0020] ;

[0021] in, Indicates the target feature point to be compensated. This represents the current target feature point.

[0022] As an optional technical solution of the present invention, the step of calculating the motion parameters of the target area based on the compensation target feature points and determining the operating state of the wind turbine transmission chain based on the motion parameters includes:

[0023] The motion parameters include instantaneous rotational speed. Radial relative displacement and axial relative displacement At least one of them;

[0024] The instantaneous rotational speed is calculated based on the change in the rotational angle of the rotational texture on the surface of the target shaft segment.

[0025] The radial relative displacement is obtained by calculating the radial displacement of the compensation target feature point relative to the reference target feature point;

[0026] The axial relative displacement is obtained by calculating the axial displacement of the compensation target feature point relative to the reference target feature point;

[0027] The operating states of the wind turbine drive chain include normal state, radial vibration increase state, axial vibration increase state, and abnormal state;

[0028] like Consistent with the set speed, and and If all values ​​are within the preset range, the condition is considered normal.

[0029] like Consistent with the set speed Within the preset axial displacement range and If the radial displacement is not within the preset range, it is determined to be a state of increased radial vibration;

[0030] like Consistent with the set speed Within the preset radial displacement range and If the axial displacement is outside the preset range, it is determined to be a state of increased axial vibration.

[0031] If it is any other case, it is determined to be an abnormal state.

[0032] As an optional technical solution of the present invention, the reliability of the visual measurement is expressed as follows:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] in, This indicates the confidence level of the visual measurements for the current frame. This indicates the image quality score for the current frame. This represents the compensation consistency score for the current frame. This represents the parameter estimation score for the current frame. The weights representing the image quality score This indicates the weight of the compensation consistency score. This indicates the weight of the parameter estimation score. This represents the compensation residual of the current frame. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents the current set of reference feature points. This represents the transformation matrix from the current frame to the reference frame. Represents Euclidean distance. This represents the scaling parameter for the compensated residual. Indicates the total number of reference feature points. This indicates the total number of current reference feature points. Represents the residuals of the motion parameters. Represents the residual scaling parameter of the motion parameters;

[0039] The image quality score of the current frame The image is obtained by normalizing the brightness, contrast, sharpness, effective target feature ratio, and occlusion ratio of the target region in the current frame image, and then taking a weighted average or geometric average of the normalization results. The effective target feature ratio represents the proportion of target feature points in the current frame to the baseline target feature points, and the occlusion ratio represents the proportion of the occluded area in the current frame.

[0040] As an optional technical solution of the present invention, the output method for determining the operating status of the wind turbine drive train based on visual measurement reliability includes:

[0041] when At that time, the operating status of the output fan drive chain is displayed;

[0042] when At that time, the operating status of the fan drive train and the retest prompts are displayed in a reduced weight output.

[0043] when At this time, the operating status of the fan drive chain is not output;

[0044] in, Indicates the first threshold. This represents the second threshold.

[0045] Secondly, a system for analyzing the state of offshore wind turbine transmission chain under wind and wave disturbance is provided, including: a data acquisition module for acquiring image sequences including target areas and reference areas on the wind turbine transmission chain;

[0046] The feature point confirmation module is used to select a reference frame in the image sequence, select reference reference feature points in the reference region and reference target feature points in the target region in the reference frame image; take any frame after the reference frame in the image sequence as the current frame, and determine the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point in the current frame image.

[0047] The transformation module is used to calculate the transformation matrix from the current frame to the reference frame based on the reference feature points and the current reference feature points.

[0048] The compensation module is used to perform coordinate compensation on the current target feature points based on the transformation matrix from the current frame to the reference frame, so as to obtain the compensated target feature points.

[0049] The operation status analysis module is used to calculate the motion parameters of the target area based on the compensation target feature points, and to determine the operation status of the wind turbine drive chain based on the motion parameters;

[0050] The credibility analysis module is used to calculate the visual measurement credibility, including image quality score, compensation consistency score and parameter estimation score, based on the transformation matrix from the current frame to the reference frame, the reference reference feature point, the current reference feature point and the current frame image, and to determine the output mode of the wind turbine drive chain operation status based on the visual measurement credibility.

[0051] Thirdly, a device for analyzing the state of an offshore wind turbine drivetrain under wind and wave disturbance is provided, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method for analyzing the state of an offshore wind turbine drivetrain under wind and wave disturbance described in the first aspect.

[0052] Fourthly, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method for analyzing the state of the offshore wind turbine drive train under wind and wave disturbances as described in the first aspect.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] This invention provides a method for analyzing the state of offshore wind turbine transmission chains under wind and wave disturbances. By compensating the target feature points of the current frame with a reference frame and using the compensated target feature points to calculate motion parameters, more accurate motion parameters are obtained. At the same time, visual measurement credibility is constructed by considering image quality score, compensation consistency score and parameter estimation score. The reliability of motion parameters is evaluated by visual measurement credibility, which enables accurate analysis of the state of offshore wind turbine transmission chains. Attached Figure Description

[0055] Figure 1 This is a flowchart of the offshore wind turbine drivetrain state analysis method under wind and wave disturbance in an embodiment of the present invention.

[0056] Figure 2 This is a schematic diagram of visual monitoring of the drive train shaft system of an offshore wind turbine in an embodiment of the present invention;

[0057] Figure 3 This is a schematic diagram of the target area and reference area in an embodiment of the present invention;

[0058] Figure 4 This is a flowchart illustrating the compensation process for the current target feature point in this embodiment of the invention.

[0059] Figure 5 This is a flowchart illustrating the generation of motion parameters and visual measurement reliability in an embodiment of the present invention.

[0060] Figure 6 This is a flowchart of the state output in an embodiment of the present invention. Detailed Implementation

[0061] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0062] Example 1

[0063] This embodiment provides a method for analyzing the state of the offshore wind turbine drivetrain under wind and wave disturbances, such as... Figure 1 As shown, the specific steps include the following:

[0064] Step 1: Obtain an image sequence including the target area and reference area on the wind turbine drive train.

[0065] The target area is located on the target shaft segment of the wind turbine drive chain, and the target shaft segment is the main shaft, high-speed shaft or exposed section of the coupling in the wind turbine drive chain.

[0066] The reference area is located on a reference structure on the wind turbine drivetrain. This reference structure includes a main bearing housing, gearbox housing, generator end cover, nacelle frame, or other supporting components fixedly connected to the target shaft segment. The reference structure is used to establish a local non-rotational reference. The compensated motion parameters characterize the motion of the target shaft segment relative to the reference structure, describing the relative dynamic relationship between the target shaft segment and the main bearing housing, gearbox housing, generator end cover, or nacelle frame. These parameters are not considered as the absolute motion of the target shaft segment in the inertial coordinate system. The reference area is preferably located near the target area, and both the reference area and the target area are within the same optical channel and depth of field to reduce the impact of depth of field differences and parallax variations on coordinate compensation. During the image acquisition device installation or system calibration phase, known micro-motions are generated in the image acquisition device, wind turbine unit frame, or fixed support. The image coordinate changes of the reference area and the target area under these micro-motion conditions are calculated. Based on the residual displacement of the stationary target obtained after applying the two-dimensional coordinate transformation relationship determined by the reference area to the target area, the allowable object distance difference between the target area and the reference area is determined.

[0067] The image sequence was acquired using a monocular image acquisition device. For example... Figure 2As shown, a single monocular image acquisition device, fixedly mounted on the wind turbine nacelle frame, drivetrain protection structure, or other non-rotating fixed structure, performs simultaneous frame imaging of the target shaft segment and the reference structure, resulting in an image sequence containing both the target and reference regions. The monocular camera is fixedly mounted on the frame inside the wind turbine nacelle or on a test bench support, with its lens facing both the target shaft segment and the reference structure. Illumination simultaneously covers both the target and reference regions. The target and reference regions are located within the same image frame and the same optical channel, and their object distance difference is within a pre-calibrated allowable range, enabling the two-dimensional coordinate transformation relationship determined based on the reference region to be used to compensate for target feature points in the target region.

[0068] Step 2: Select a reference frame in the image sequence, and select the reference reference feature point in the reference region and the reference target feature point in the target region in the reference frame image; take any frame after the reference frame in the image sequence as the current frame, and determine the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point in the current frame image.

[0069] For the reference frame image, multiple stable feature points are extracted in the reference region as reference feature points. These stable feature points are formed from the corners, centers, or natural high-contrast features of artificially marked reference points or structural surfaces. To improve the stability of the transformation estimation, the number and spatial distribution of the reference feature points meet the estimation requirements of the selected two-dimensional coordinate transformation model; when using affine transformation, there are at least three reference feature points that are not collinear; when using homography transformation, there are at least four reference feature points that do not satisfy a degenerate distribution.

[0070] In the current frame In this process, based on the local appearance or descriptive information of the baseline reference feature, the corresponding reference feature is searched, and bidirectional tracking verification or local description consistency verification is performed on the reference feature. Reference features with excessive forward and backward errors, abnormal brightness changes, or located in occluded areas are eliminated to obtain the current reference feature. The current reference feature set is represented as follows: The total number of current reference feature points is set to... Determine the reference region in the current frame image One current reference feature.

[0071] In the same continuous analysis task, the baseline frame remains unchanged. If the current number of reference features is insufficient or the compensation residual increases, reference compensation is restored first by expanding or redefining the reference area and reselecting the set of reference feature points. When the current reference features cannot be restored and the baseline frame becomes invalid, the output results are paused, and a new baseline frame is selected after recalibration. At the same time, a new monitoring data segment and a data segment start identifier are established.

[0072] like Figure 3The image shown is a schematic diagram of any image frame, where the target region and the reference region are in the same frame.

[0073] Step 3: Based on the reference feature points and the current reference feature points, calculate the transformation matrix from the current frame to the reference frame.

[0074] The transformation matrix from the current frame to the reference frame is represented as:

[0075] ;

[0076] in, This represents the transformation matrix from the current frame to the reference frame. Represents a two-dimensional coordinate transformation matrix. This represents a robust loss function used to mitigate the impact of anomalous features. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents Euclidean distance. Represents the current set of reference feature points. This indicates that through optimization Obtain the minimum value.

[0077] Two-dimensional coordinate transformation matrix A similarity transformation matrix or an affine transformation matrix is ​​used. Similarity transformations include translation, rotation, and scaling changes within the image plane; affine transformations are used when local viewpoint changes cannot be described by similarity transformations. The transformation direction is always from the current frame to the reference frame to avoid drift caused by the continuous accumulation of adjacent transformations frame by frame.

[0078] Step 4: Based on the transformation matrix from the current frame to the reference frame, perform coordinate compensation on the current target feature points to obtain the compensated target feature points.

[0079] like Figure 4 As shown, coordinate compensation is performed on the current target feature point using the transformation matrix from the current frame to the reference frame. The compensated target feature point is represented as follows:

[0080] ;

[0081] in, Indicates the target feature point for compensation. This represents the current target feature point.

[0082] When the target feature point is an edge point, coordinate transformation is performed on all edge points; when the target feature point is a marker center or a local image patch location, transformation is performed on the corresponding center coordinates or position coordinates. For relative displacement calculation, the target feature point in the reference frame is used as the relative displacement zero point. The reference frame and its relative displacement zero point are kept unchanged in the same continuous analysis task.

[0083] Step 5: Calculate the motion parameters of the target area based on the compensation target feature points, and determine the operating status of the fan drive chain based on the motion parameters.

[0084] like Figure 5 As shown, the motion parameters include instantaneous rotational speed. Radial relative displacement and axial relative displacement At least one of them. The motion parameter vector is represented as... When some motion information cannot be obtained from the current visual features, a corresponding motion parameter sub-vector is constructed based on the actual observable motion amount, such as... or .

[0085] Specifically, the instantaneous rotational speed is calculated based on the rotational angle change of the rotational texture on the surface of the target shaft segment; the radial relative displacement is obtained by calculating the radial displacement of the compensation target feature point relative to the reference target feature point; and the axial relative displacement is obtained by calculating the axial displacement of the compensation target feature point relative to the reference target feature point.

[0086] The operating states of the wind turbine drive chain include normal state, radial vibration increase state, axial vibration increase state, and abnormal state.

[0087] like Consistent with the set speed, and and If all values ​​are within the preset range, the condition is considered normal.

[0088] like Consistent with the set speed Within the preset axial displacement range and If the radial displacement is not within the preset range, it is determined to be a state of increased radial vibration.

[0089] like Consistent with the set speed Within the preset radial displacement range and If the axial displacement is outside the preset range, it is determined to be a state of increased axial vibration.

[0090] If it is any other case, it is determined to be an abnormal state.

[0091] Step 6: Based on the transformation matrix from the current frame to the reference frame, the reference reference feature points, the current reference feature points, and the current frame image, calculate the visual measurement confidence level, which includes image quality score, compensation consistency score, and parameter estimation score, and determine the output mode of the wind turbine drive chain's operating status according to the visual measurement confidence level.

[0092] like Figure 5 As shown, the image quality score is obtained by normalizing the brightness, contrast, sharpness, effective target feature ratio, and occlusion ratio of the target region in the current frame image, and then applying a weighted average or geometric average to the normalized results. The effective target feature ratio represents the proportion of target feature points in the current frame relative to the baseline target feature points, and the occlusion ratio represents the proportion of occluded area in the current frame. The image quality score decreases accordingly when the image brightness is too low, the pixel saturation ratio is too high, the edge gradient is reduced, or the number of effective target features decreases. The compensation consistency score is formed based on the reference compensation residual and the current reference feature ratio. The parameter estimation score is formed based on the motion parameter residual.

[0093] The reliability of visual measurements is expressed as follows:

[0094] ;

[0095] ;

[0096] ;

[0097] ;

[0098] ;

[0099] in, This indicates the confidence level of the visual measurements for the current frame. This indicates the image quality score for the current frame. This represents the compensation consistency score for the current frame. This represents the parameter estimation score for the current frame. The weights representing the image quality score This indicates the weight of the compensation consistency score. This indicates the weight of the parameter estimation score. This represents the compensation residual of the current frame. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents the current set of reference feature points. This represents the transformation matrix from the current frame to the reference frame. Represents Euclidean distance. This represents the scaling parameter for the compensated residual. Indicates the total number of reference feature points. This indicates the total number of current reference feature points. Represents the residuals of the motion parameters. This represents the residual scaling parameter of the motion parameters.

[0100] The compensation residual reflects whether the reference region can stably establish a fixed reference coordinate system. If the number of current reference features is insufficient, the compensation residual exceeds a set value, or the scale and rotation changes of the transformation matrix exceed the allowable range, the current frame is marked as a reference compensation aberration frame. Weights are determined by calibration data or stable operating data; the weights, residual scale parameters, and confidence level thresholds are determined through calibration data. Specifically, they can be selected through grid search, cross-validation, or minimizing the monitoring error of calibration samples. Visual measurement confidence can be calculated separately for different motion parameters. For example, rotational speed confidence uses rotational texture quality, reference compensation consistency, and rotational speed estimation residuals; displacement confidence uses corresponding edge quality, reference compensation consistency, and edge fitting residuals.

[0101] Specifically, in one implementation combining wind turbine generator operating condition information and sea state information, the wind turbine generator operation process is divided into at least one operating condition interval among wind speed, power generation, speed, pitch angle, yaw state, and nacelle vibration signal, and at least one sea state interval among significant wave height, spectral peak period, wave principal direction, floating support structure heave, pitch, and roll, based on at least one of the following: startup, grid connection and speed-up, rated power, power limiting, yaw adjustment, and shutdown. Furthermore, it is divided into at least one sea state interval among low wave disturbance, medium wave disturbance, and high wave disturbance. For different operating condition intervals and sea state intervals, fusion weights, confidence level thresholds, or state recognition thresholds for visual measurement reliability are set respectively. When the effective wave height increases, the spectral peak period approaches the main motion period of the fixed foundation or floating support structure, or the heave, pitch, or roll amplitude of the floating support structure increases, the weight of the compensation consistency score or the rejection threshold is increased. Under low-light conditions, the weight of the image quality score is increased. Under power-limited low-speed or slowly changing speed conditions, the speed analysis time window or parameter estimation score weight is adjusted. Under rapidly changing speed conditions, the weight of the parameter estimation score is increased or a shorter speed analysis time window is used. This ensures that the reliability of visual measurements reflects not only the image and parameter estimation quality but also the impact of the wind turbine's operating conditions on the reliability of visual monitoring.

[0102] like Figure 6 As shown, the output method for determining the operating status of the wind turbine drive train based on the reliability of visual measurement includes:

[0103] when At that time, the operating status of the output fan drive chain is displayed;

[0104] when At that time, the operating status of the fan drive train and the retest prompts are displayed in a reduced weight output.

[0105] when At this time, the operating status of the fan drive chain is not output; only the cause of insufficient image quality, abnormal reference compensation, or abnormal parameter estimation is output.

[0106] in, Indicates the first threshold. This represents the second threshold.

[0107] In the data acquisition and feedback implementation, adjustments are made based on the main scores that lead to decreased credibility. The image quality score, compensation consistency score, and parameter estimation score are compared with their corresponding source determination thresholds. Information sources corresponding to scores below the corresponding source determination thresholds are identified as low-credibility sources. When multiple scores are simultaneously below their corresponding thresholds, adjustment priorities are determined based on the normalized gap of each score relative to its corresponding threshold. When insufficient image brightness, cabin dust, oil mist, or condensation cause edge quality degradation, illumination brightness is increased, exposure time is adjusted, the observation window is cleaned, or the target region of interest is adjusted. When motion blur increases, exposure time is shortened and illumination brightness is increased. When the number of effective reference features is insufficient or the compensation residual increases, the reference region of interest is expanded or redefined, or a new set of reference feature points is selected. When the parameter estimation residual increases but image quality and compensation consistency are normal, the feature tracking time window is adjusted or the motion parameter estimation strategy is switched. After adjustment, the visual measurement credibility is recalculated, forming a closed-loop monitoring process.

[0108] Example 2

[0109] This embodiment, based on Embodiment 1, provides a set of experiments to verify the feasibility of the method in Embodiment 1.

[0110] In a series of verification tests, the target shaft segment operated at a stable rotational speed of 600 r / min, with the image acquisition device's frame rate set to 120 FPS and the exposure time set to 8 ms. A controllable displacement device was used to induce periodic micro-motions of 0.30 mm amplitude and 2.0 Hz frequency in the wind turbine drivetrain test bench frame and the image acquisition device. These periodic micro-motions were considered equivalent to the wave load causing the fixed foundation, tower, nacelle, or floating support structure to move together. A 180-second image sequence was continuously acquired. Under the same image sequence, the radial relative displacement of the target shaft segment relative to the fixed reference structure was calculated using methods including no reference compensation, cumulative compensation of adjacent frames, and compensation with a fixed reference frame.

[0111] Without reference compensation, the calculated relative displacement contains a significant component of the frame's common motion, with a spurious root mean square (RMS) displacement of 0.214 mm and a peak-to-peak value of 0.618 mm. With adjacent frame cumulative compensation, the spurious RMS displacement is reduced to 0.081 mm, but due to the continuous accumulation of adjacent frame matching errors, the cumulative drift reaches 0.312 mm at the end of monitoring. Using the fixed reference frame compensation method described in this paper, the spurious RMS displacement is further reduced to 0.031 mm, and the cumulative drift at the end of monitoring is 0.044 mm. Compared with no reference compensation and adjacent frame cumulative compensation, the fixed reference frame compensation method reduces the spurious RMS displacement by 85.51% and 61.73%, respectively; compared with adjacent frame cumulative compensation, the cumulative drift at the end of monitoring is reduced by 85.90%. The results show that directly transforming the current frame to the fixed reference frame coordinate system can suppress spurious displacements caused by the common motion of the engine room frame, transmission chain test bench frame, or image acquisition device, and avoid the continuous accumulation of adjacent frame transformation errors over monitoring time.

[0112] Further, 900 low-quality images were included in the image sequence, with 300 frames each of low-light images, partially occluded images, and images with significant motion blur. Without a visual monitoring quality evaluation and grading output mechanism, 711 low-quality images were accepted as valid measurement results, resulting in a false acceptance rate of 79.00%. After adopting the visual measurement reliability grading, deweighting, and rejection mechanism of this method, 823 low-quality images were correctly identified and deweighted or rejected, leaving only 77 low-quality images that were still incorrectly accepted, reducing the false acceptance rate to 8.56%. Meanwhile, among the 2700 normal-quality images, 2639 were determined to be valid measurement results, resulting in an effective retention rate of 97.74% for normal-quality images.

[0113] The above results show that the proposed method can reduce false displacement and long-term cumulative drift caused by the joint movement of the cabin frame, transmission test bench frame, or image acquisition device through fixed reference compensation. By combining the image quality of the target area, the reference compensation residual, and the motion parameter estimation residual to generate visual measurement reliability, it can reduce the situation where low-quality measurement results enter deterministic judgment under imaging degradation conditions while maintaining the effective utilization rate of normal quality images.

[0114] Example 3

[0115] This embodiment provides a system for analyzing the state of offshore wind turbine transmission chains under wind and wave disturbances, including:

[0116] pass Figure 2 The data processing apparatus in the embodiment performs the method in Example 1, which includes:

[0117] The data acquisition module is used to acquire image sequences including the target area and reference area on the wind turbine drive chain;

[0118] The feature point confirmation module is used to select a reference frame in the image sequence, select reference reference feature points in the reference region and reference target feature points in the target region in the reference frame image; take any frame after the reference frame in the image sequence as the current frame, and determine the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point in the current frame image.

[0119] The transformation module is used to calculate the transformation matrix from the current frame to the reference frame based on the reference feature points and the current reference feature points.

[0120] The compensation module is used to perform coordinate compensation on the current target feature points based on the transformation matrix from the current frame to the reference frame, so as to obtain the compensated target feature points.

[0121] The operation status analysis module is used to calculate the motion parameters of the target area based on the compensation target feature points, and to determine the operation status of the wind turbine drive chain based on the motion parameters;

[0122] The credibility analysis module is used to calculate the visual measurement credibility, including image quality score, compensation consistency score and parameter estimation score, based on the transformation matrix from the current frame to the reference frame, the reference reference feature point, the current reference feature point and the current frame image, and to determine the output mode of the wind turbine drive chain operation status based on the visual measurement credibility.

[0123] Example 4

[0124] This embodiment provides a device for analyzing the state of an offshore wind turbine drive train under wind and wave disturbances, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the offshore wind turbine drive train state analysis method under wind and wave disturbances described in Embodiment 1.

[0125] Example 5

[0126] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the offshore wind turbine drivetrain state analysis method under wind and wave disturbances described in Embodiment 1.

[0127] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will 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 program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0129] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0131] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for analyzing the state of a offshore wind turbine drive train under wind and wave disturbance, characterized in that, include: Acquire an image sequence including the target area and reference area on the wind turbine drive train; A reference frame is selected in the image sequence, and reference reference feature points within the reference region and reference target feature points within the target region are selected in the reference frame image. Take any frame after the reference frame in the image sequence as the current frame, and determine the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point in the current frame image; Calculate the transformation matrix from the current frame to the reference frame based on the reference feature points and the current reference feature points; Based on the transformation matrix from the current frame to the reference frame, coordinate compensation is performed on the current target feature points to obtain the compensated target feature points; The motion parameters of the target area are calculated based on the compensation target feature points, and the operating status of the wind turbine drive chain is determined based on the motion parameters. Based on the transformation matrix from the current frame to the reference frame, the reference reference feature points, the current reference feature points, and the current frame image, the visual measurement confidence level, including image quality score, compensation consistency score, and parameter estimation score, is calculated. The output mode of the wind turbine drive chain's operating status is then determined based on the visual measurement confidence level.

2. The method for analyzing the state of the offshore wind turbine transmission chain under wind and wave disturbance as described in claim 1, characterized in that, The target area is located on the target shaft segment of the wind turbine drive chain, and the target shaft segment is the main shaft, high-speed shaft or exposed section of the coupling in the wind turbine drive chain. The reference area is located on a reference structure on the wind turbine drive chain. The reference structure is a main bearing housing, gearbox housing, generator end cover, nacelle frame or other support components that are fixedly connected to the target shaft segment.

3. The method for analyzing the state of the offshore wind turbine drive train under wind and wave disturbance as described in claim 1, characterized in that, The transformation matrix from the current frame to the reference frame is represented as follows: ； in, This represents the transformation matrix from the current frame to the reference frame. Represents a two-dimensional coordinate transformation matrix. Represents a robust loss function. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents Euclidean distance. Represents the current set of reference feature points. This indicates that through optimization Obtain the minimum value.

4. The method for analyzing the state of the offshore wind turbine transmission chain under wind and wave disturbance as described in claim 3, characterized in that, The compensation target feature points are represented as follows: ； in, Indicates the target feature point for compensation. This represents the current target feature point.

5. The method for analyzing the state of the offshore wind turbine transmission chain under wind and wave disturbance as described in claim 2, characterized in that, The step of calculating the motion parameters of the target area based on the compensated target feature points, and determining the operating state of the wind turbine drive chain based on the motion parameters, includes: The motion parameters include instantaneous rotational speed. Radial relative displacement and axial relative displacement At least one of them; The instantaneous rotational speed is calculated based on the change in the rotational angle of the rotational texture on the surface of the target shaft segment. The radial relative displacement is obtained by calculating the radial displacement of the compensation target feature point relative to the reference target feature point; The axial relative displacement is obtained by calculating the axial displacement of the compensation target feature point relative to the reference target feature point; The operating states of the wind turbine drive chain include normal state, radial vibration increase state, axial vibration increase state, and abnormal state; like Consistent with the set speed, and and If all values ​​are within the preset range, the condition is considered normal. like Consistent with the set speed Within the preset axial displacement range and If the radial displacement is not within the preset range, it is determined to be a state of increased radial vibration; like Consistent with the set speed Within the preset radial displacement range and If the axial displacement is outside the preset range, it is determined to be a state of increased axial vibration. If it is any other case, it is determined to be an abnormal state.

6. The method for analyzing the state of the offshore wind turbine drive train under wind and wave disturbance as described in claim 1, characterized in that, The reliability of the visual measurement is expressed as follows: ； ； ； ； ； in, This indicates the confidence level of the visual measurements for the current frame. This indicates the image quality score for the current frame. This represents the compensation consistency score for the current frame. This represents the parameter estimation score for the current frame. The weights representing the image quality score This indicates the weight of the compensation consistency score. This indicates the weight of the parameter estimation score. This represents the compensation residual of the current frame. Indicates the relationship with the first The reference feature point corresponding to the current reference feature point, Indicates the first One current reference feature point, Represents the current set of reference feature points. This represents the transformation matrix from the current frame to the reference frame. Represents Euclidean distance. This represents the scaling parameter for the compensated residual. Indicates the total number of reference feature points. This indicates the total number of current reference feature points. Represents the residuals of the motion parameters. Represents the residual scaling parameter of the motion parameters; The image quality score of the current frame The image is obtained by normalizing the brightness, contrast, sharpness, effective target feature ratio, and occlusion ratio of the target region in the current frame image, and then taking a weighted average or geometric average of the normalization results. The effective target feature ratio represents the proportion of target feature points in the current frame to the baseline target feature points, and the occlusion ratio represents the proportion of the occluded area in the current frame.

7. The method for analyzing the state of the offshore wind turbine drive train under wind and wave disturbance as described in claim 6, characterized in that, The output method for determining the operating status of the wind turbine drive train based on visual measurement reliability includes: when At that time, the operating status of the output fan drive chain is displayed; when At that time, the operating status of the fan drive train and the retest prompts are displayed in a reduced weight output. when At this time, the operating status of the fan drive chain is not output; in, Indicates the first threshold. This represents the second threshold.

8. A system for analyzing the state of a marine wind turbine drive train under wind and wave disturbance, characterized in that, include: The data acquisition module is used to acquire image sequences including the target area and reference area on the wind turbine drive chain; The feature point confirmation module is used to select a reference frame in the image sequence, and to select reference reference feature points in the reference region and reference target feature points in the target region in the reference frame image. Take any frame after the reference frame in the image sequence as the current frame, and determine the current reference feature point corresponding to the reference reference feature point and the current target feature point corresponding to the reference target feature point in the current frame image; The transformation module is used to calculate the transformation matrix from the current frame to the reference frame based on the reference feature points and the current reference feature points. The compensation module is used to perform coordinate compensation on the current target feature points based on the transformation matrix from the current frame to the reference frame, so as to obtain the compensated target feature points. The operation status analysis module is used to calculate the motion parameters of the target area based on the compensation target feature points, and to determine the operation status of the wind turbine drive chain based on the motion parameters; The credibility analysis module is used to calculate the visual measurement credibility, including image quality score, compensation consistency score and parameter estimation score, based on the transformation matrix from the current frame to the reference frame, the reference reference feature point, the current reference feature point and the current frame image, and to determine the output mode of the wind turbine drive chain operation status based on the visual measurement credibility.

9. A device for analyzing the state of a marine wind turbine transmission chain under wind and wave disturbance, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the offshore wind turbine drivetrain state analysis method under wind and wave disturbance as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the offshore wind turbine drivetrain state analysis method under wind and wave disturbance as described in any one of claims 1 to 7.

Images