A method for monitoring wind turbine blade deformation

Abstract

This invention discloses a method for monitoring wind turbine blade deformation, comprising the following steps: S1, opening equalizing holes on the blade baffle and installing flow meters inside the holes to acquire airflow data during blade operation and transmitting the data to the main control system; S2, collecting SCADA data from the turbine generator set and transmitting the SCADA data to the main control system; S3, obtaining airflow data at different speeds and positions during normal blade operation based on the turbine generator set SCADA data and airflow valve data, and establishing corresponding data for different speeds and rotational positions during normal blade operation; S4, setting threshold values ​​for airflow valve flow data at different speeds and positions of the blade; S5, monitoring the airflow data and SCADA data of the airflow valve in real time, and issuing an alarm when the airflow data exceeds the set threshold. This method deeply integrates the internal and external airflow of the blade with operating parameters, utilizing the internal gas flow of the blade to reflect the deformation state, thus adapting to the complex operating environment of wind turbine blades while also considering accuracy, economy, and practicality.

Description

Technical Field

[0001] This invention relates to a method for monitoring the deformation of wind turbine blades, belonging to the technical field of wind turbine units. Background Technology

[0002] Wind turbine blades are the core components of wind turbine generators, and their operating status directly affects the generator's power generation efficiency and safe and stable operation. During long-term service, blades are susceptible to deformation due to factors such as complex alternating loads, aerodynamic impacts, and erosion. If not detected in time, this deformation can gradually worsen, potentially leading to structural damage or even breakage of the blades, causing generator shutdowns or safety accidents and resulting in severe economic losses.

[0003] Currently, the main methods for monitoring wind turbine blade deformation include visual inspection, strain sensing monitoring, and modal analysis. Visual inspection relies on manual inspection or drone photography, is greatly affected by environmental factors such as sunlight and weather, and is difficult to achieve real-time monitoring. Strain sensing monitoring requires the placement of numerous sensors on or inside the blade, which not only results in high installation and maintenance costs but may also affect the aerodynamic performance and structural integrity of the blade. Modal analysis identifies deformation by recognizing changes in blade vibration characteristics, but is easily affected by turbine operating vibration and environmental noise, and its monitoring accuracy is limited. Summary of the Invention

[0004] The purpose of this invention is to address the challenges posed by the increasing capacity and blade size of wind turbine units as wind power technology advances. Traditional monitoring methods are no longer sufficient to meet the real-time accuracy and cost-effectiveness requirements of monitoring large wind turbine blades. Therefore, this invention aims to develop a real-time, accurate, and low-cost method for monitoring wind turbine blade deformation.

[0005] The technical solution adopted in this invention is as follows: A method for monitoring the deformation of wind turbine blades, comprising the following steps:

[0006] S1. Make pressure equalization holes on the blade baffle and install a flow meter in the hole to obtain airflow data when the blade is running and transmit the data to the main control system;

[0007] S2. Collect SCADA data from the unit and transmit the SCADA data to the main control system;

[0008] S3. Based on the SCADA data of the unit and the airflow valve data, obtain the airflow data of the blades at different speeds and positions during normal operation, and establish a database of corresponding relationships of the blades at different speeds and rotation positions during normal operation.

[0009] S4. Set the threshold values ​​for airflow valve flow data under different blade speeds and positions;

[0010] S5. Monitors the airflow data of the airflow valve in real time and issues an alarm if the airflow exceeds the set threshold.

[0011] Alternatively, the pressure equalization hole may be located at the blade root baffle.

[0012] Optional SCADA data includes real-time blade rotation speed and real-time blade orientation.

[0013] SCADA data may also include wind speed and temperature data.

[0014] Optional airflow data includes flow rate and / or velocity.

[0015] Alternatively, in step S3, a predictive model of blade speed-position-airflow data is constructed based on historical SCADA data and airflow valve data, and the normal gas velocity and flow rate baseline values ​​of the blade at any speed and position are obtained by calculation and prediction.

[0016] Alternatively, the prediction model may be a machine learning model.

[0017] Alternatively, the calculation and prediction process may also incorporate blade physical property parameters, including blade length, stiffness coefficient, and material elastic modulus, and correct the normal airflow data baseline value output by the prediction model through fluid dynamics simulation calculation.

[0018] Alternatively, the upper limit of statistical data or the limit calculated based on the safety margin can be used as the threshold.

[0019] Alternatively, every preset period, the corresponding relational database can be updated again using the latest collected normal operation data, and the baseline value and corresponding threshold value can be updated synchronously.

[0020] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0021] 1. The wind turbine blade deformation monitoring method provided by this invention deeply integrates the airflow inside and outside the blade with SCADA data, using the internal gas flow of the blade to reflect the deformation state. This method is suitable for the complex operating environment of wind turbine blades and also takes into account accuracy, economy, and practicality. By opening equalizing holes and installing flow meters at the blade baffle, the key physical signal reflecting deformation—internal airflow changes—can be directly captured without damaging the blade's structural integrity, reducing hardware deployment and maintenance costs. By combining the corresponding data with the operating parameters such as speed and position collected by the SCADA system, the method can accurately cover the normal airflow characteristics under any operating state of the blade, avoiding the omissions of traditional fixed threshold monitoring for specific operating conditions. The entire process from data acquisition and benchmark modeling to threshold judgment is automated, and alarms can be triggered in real time, significantly improving the monitoring response speed and reliability.

[0022] 2. The wind turbine blade deformation monitoring method provided by this invention incorporates wind speed and temperature into SCADA data, which can correct for environmental factors that interfere with airflow data, such as changes in gas density caused by high temperatures or local airflow disturbances caused by strong winds. Simultaneously, it monitors flow rate and velocity to form multi-parameter cross-validation, and uses fluid dynamics equations to infer the internal pressure distribution of the blade, improving the accuracy of anomaly diagnosis. The machine learning model based on historical data can quickly adapt to the physical characteristics of different turbine units, such as blade length and stiffness coefficient, without the need for separate modeling. After incorporating parameters such as the elastic modulus of materials, the prediction results are corrected through fluid-structure interaction simulation to ensure long-term accuracy of the benchmark value. Thresholds are set using the upper limit of normal data statistics or safety margin limits, which cover normal fluctuations and reserve safety redundancy. Updating the database every preset period can dynamically track blade performance drift and avoid static benchmark failure. SCADA data from multiple units can also be used to discover common patterns, support model migration across units, and reduce the cost of batch deployment of wind farms. Attached Figure Description

[0023] Figure 1 This is the flowchart of this method. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the accompanying drawings.

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0026] A method for monitoring wind turbine blade deformation, such as Figure 1 As shown, it includes the following steps:

[0027] S1. Make pressure equalization holes on the blade baffle and install a flow meter in the hole to obtain airflow data when the blade is running and transmit the data to the main control system;

[0028] S2. Collect SCADA data from the unit and transmit the SCADA data to the main control system;

[0029] S3. Based on the SCADA data of the unit and the airflow valve data, obtain the airflow data of the blades at different speeds and positions during normal operation, and establish a database of corresponding relationships of the blades at different speeds and rotation positions during normal operation.

[0030] S4. Set the threshold values ​​for airflow valve flow data under different blade speeds and positions;

[0031] S5: Monitors airflow data and SCADA data of the airflow valve in real time, and issues an alarm when the set threshold is exceeded.

[0032] In this solution, S1 directly captures the physical signals caused by blade deformation by acquiring blade airflow data. Since blade deformation changes the internal volume, leading to abnormal gas flow characteristics, this data becomes the core basis for judging deformation. S2 collects unit SCADA data including operating condition information such as blade speed and position, providing an operating condition benchmark to distinguish between normal and abnormal conditions under different operating states, avoiding misjudgments caused by a single signal due to changes in operating conditions. S3 combines the two types of data to establish a normal correspondence database. By associating speed, position, and flow data, it accurately defines the normal state under different operating conditions, eliminating the normal influence of speed difference and position change on the breathing effect, making the benchmark more consistent with the actual operating scenario. S4 sets corresponding thresholds to quantify the normal state into a comparable standard, clarifying the boundary of abnormal judgment, and ensuring that even subtle changes exceeding the normal range can be effectively identified. S5 compares real-time monitoring data with thresholds to form a closed loop from signal acquisition to abnormal warning, achieving timely detection of deformation.

[0033] Under specific rotational speeds and positions, the normal deformation of blades under centrifugal force and aerodynamic loads is regular and stable. The resulting changes in internal volume create gas flow with fixed characteristics, meaning that the flow data has a reproducible correspondence with rotational speed and position. When blades undergo abnormal deformation, the spatial morphology of their internal structure changes. At the same rotational speed and position, the deformation amplitude differs from the normal state, causing the amount and rate of change in internal volume to deviate from the norm. This directly alters the intensity and regularity of gas flow generated by the breathing effect, causing the flow data to no longer follow the normal characteristics under the corresponding operating conditions. Ultimately, this manifests as real-time monitoring data exceeding set thresholds or not matching standard ranges, thus being identified as abnormal by the system.

[0034] In particular, this solution, by establishing pressure equalization holes and monitoring airflow data, offers significant advantages over simply installing a pressure gauge inside the sealed blade. Firstly, when a pressure gauge is installed inside the sealed blade, the internal pressure changes are simultaneously affected by temperature fluctuations and the breathing effect caused by blade deformation. Temperature changes directly alter pressure through gas expansion and contraction, while blade deformation indirectly affects pressure through volume changes. The combined effect of these two factors creates a mixed pressure signal, making it difficult to accurately separate pressure changes solely caused by deformation, easily leading to misjudgments of the deformation state. With pressure equalization holes, the pressure difference caused by temperature changes is balanced, and the airflow data is primarily driven by the breathing effect generated by blade deformation, more directly reflecting the degree of deformation, reducing interference from non-deformation factors, and making the monitoring results more accurate. Furthermore, under the influence of the breathing effect, the internal pressure of the sealed blade changes with blade deformation. During blade transportation, due to regional differences, the internal pressure changes caused by gas expansion and contraction create a pressure difference with the outside environment, placing additional loads on the blade structure. Long-term effects may exacerbate blade fatigue damage and even amplify existing deformation. The equalizing orifice in this solution can balance the air pressure inside and outside the blade in real time, eliminate the pressure difference caused by breathing effect or temperature change, avoid damage to the blade by additional stress, protect the integrity of the blade structure, and make the gas flow data more purely correlated with deformation, further improving the reliability of monitoring.

[0035] In another specific implementation, the pressure equalization hole is located at the blade root baffle. The blade root is a critical part connecting the blade and the hub, with high structural strength and relatively stable stress state. Opening a pressure equalization hole at this location has a smaller impact on the overall structural integrity of the blade, and can avoid the structural damage risk that may be caused by opening holes in areas with weaker structures and higher aerodynamic loads, such as the middle section or tip of the blade.

[0036] As another specific implementation, airflow data includes flow rate and / or velocity. When the blade experiences slight local deformation, it alters the local cross-sectional dimensions of the gas flow channel, causing fluctuations in flow rate and helping the system detect initial blade anomalies earlier. Flow rate represents the total volume of gas passing through the equalizing orifice per unit time, and is more directly related to the overall change in the blade's internal volume. When the blade undergoes overall deformation, the total change in internal volume changes significantly, causing the flow rate to deviate significantly from the normal range. This avoids judgment bias caused by a single parameter due to local interference or overall fluctuations.

[0037] As another specific implementation, SCADA data includes the real-time rotational speed of the blades, the real-time azimuth position of the blades, and the operating status parameters of the unit. The real-time rotational speed is directly related to the magnitude of the centrifugal force and aerodynamic load it experiences during rotation. The degree of normal blade deformation varies at different rotational speeds, resulting in differences in the intensity of the breathing effect. The real-time azimuth position reflects its spatial angle during the rotation cycle. Different positions experience different wind loads, leading to differences in the normal deformation state and changes in the gas flow characteristics generated by the breathing effect.

[0038] As another specific implementation method, SCADA data also includes wind speed and temperature data. Wind speed directly affects the aerodynamic load on the blades. Including wind speed data allows predictive models to accurately distinguish normal flow differences under different wind conditions, avoiding misjudging flow fluctuations caused by wind changes as deformation anomalies. Temperature data reflects the thermal state of the gas inside the blades in real time, causing changes in gas density and pressure. Combining temperature data can correct for interference caused by thermal effects. This further refines the control over the correlation between the blade operating environment and airflow data, providing a more comprehensive basis for establishing normal baselines and judging anomalies.

[0039] As another specific implementation method, in step S3, a predictive model of blade speed-position-airflow data is constructed based on historical SCADA data and airflow valve data. This model calculates and predicts the baseline values ​​of normal gas velocity and flow rate for the blade at any speed and position. During blade operation, the combinations of speed and position are diverse and influenced by environmental factors. Simple historical data statistics are insufficient to cover all operating conditions. However, the predictive model, by learning the implicit nonlinear relationships in historical data, can understand the changing patterns of airflow data at different speeds and positions, and thus output corresponding normal baseline values ​​for any combination of speed and position, filling the baseline gaps for unobserved operating conditions. Furthermore, real-time collected wind speed and temperature data can be input into the predictive model, and the initial predicted baseline values ​​of normal airflow data can be dynamically adjusted using preset environmental influence coefficients or trained correlation sub-models. This enhances the dynamism of the baseline values; with the continuous input of new operating data, the model can iteratively optimize, adapting to normal performance drift caused by slight aging during long-term blade operation, providing a more reliable reference for subsequent threshold setting and anomaly judgment.

[0040] As another specific implementation, the prediction model is a machine learning model. Using a machine learning model as the prediction model allows for in-depth analysis of the complex nonlinear relationship between blade rotation speed and position and airflow data. During blade operation, the impact of changes in rotation speed and position on the breathing effect is not a simple linear correlation; it is also indirectly affected by environmental factors such as wind speed and temperature, forming a complex mapping relationship with multiple coupled variables. The machine learning model, through a multi-layered network structure or feature mapping mechanism, can automatically identify these complex correlations hidden in historical data, accurately capturing the changing patterns of airflow data under different operating conditions without the need for manually pre-setting mathematical formulas.

[0041] As another specific implementation, the calculation and prediction process also incorporates blade physical property parameters, including blade length, stiffness coefficient, and material elastic modulus. Through fluid dynamics simulation calculations, the baseline values ​​of normal airflow data output by the prediction model are corrected. Blade length directly determines the distribution of centrifugal force and the range of internal volume changes during rotation; stiffness coefficient reflects the blade's ability to resist deformation; and material elastic modulus is related to the reversibility and magnitude of blade deformation. By inputting the blade's physical property parameters, the simulation can model the actual deformation process of the blade at different speeds and positions, and then calculate the internal gas flow pattern under this deformation state.

[0042] As another specific implementation method, the statistical upper limit of normal data and the limit calculated based on the safety margin are used as thresholds. The statistical upper limit ensures that the threshold is inclusive of daily operational fluctuations, allowing the system to maintain stable and normal judgment under complex operating conditions and reducing the impact of false alarms on operation and maintenance efficiency; the safety margin limit builds a solid defense for blade safety, and can accurately identify and warn even the slight deformations that do not appear in historical data, thus preventing the risk from escalating.

[0043] As another specific implementation, the corresponding relational database is updated every preset period using the latest collected normal operation data, synchronously updating the baseline value and the corresponding threshold. Blade stiffness may slightly decrease with increasing operating time, causing a slow change in the intensity of the normal breathing effect at the same rotational speed and position. The updated baseline value can accurately match the normal operation characteristics of the current stage, and the corresponding threshold can be adjusted accordingly. In this scheme, completely new data from each period can be used for recalculation, or a rolling overlay method can be adopted.

[0044] SCADA data can be generated from multiple wind turbines. The operational data of a single wind turbine is often limited by its specific installation environment, turbine specifications, and operating cycle, resulting in a relatively limited sample type. This may mean that the predictive model can only be adapted to the specific operating conditions of that turbine, making it difficult to generalize to other turbines. SCADA data generated from multiple wind turbines can cover operating conditions in different wind farm environments, with different turbine specifications and at different operating stages. This not only supplements the extreme operating conditions not covered by a single turbine but also provides common operating patterns for similar turbines. Furthermore, multi-turbine SCADA data can provide a comparative basis for anomaly detection. When a turbine detects airflow data deviating from the baseline, normal data from other turbines under the same operating conditions can be used for verification, effectively reducing false alarms caused by anomalies in data from a single turbine and improving the accuracy of anomaly detection.

[0045] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. The invention extends to any new features or combinations disclosed in this specification, and any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention. It is obvious to those skilled in the art that the invention is not limited to the details of the above exemplary embodiments, and that any detailed technical features not disclosed in these embodiments are prior art. Those skilled in the art can understand the specific manner in which the above terms are used in the embodiments of the invention according to the specific circumstances, and this disclosure does not specifically limit the embodiments in this regard.

Claims

1. A method for monitoring wind turbine blade deformation, characterized in that: Includes the following steps: S1. Make pressure equalization holes at the blade root baffle and install a flow meter in the holes to obtain airflow data during blade operation and transmit the data to the main control system. S2. Collect SCADA data from the unit and transmit the SCADA data to the main control system; S3. Based on the SCADA data and flow meter data of the unit, obtain the airflow data at different speeds and positions of the blades during normal operation, and establish the corresponding relationship data of the blades at different speeds and rotation positions during normal operation. S4. Set the flow meter flow data threshold for different blade rotation speeds and positions; S5: Monitors the airflow data and SCADA data of the flow meter in real time, and issues an alarm when the airflow data exceeds the set threshold.

2. The wind turbine blade deformation monitoring method as described in claim 1, characterized in that: The SCADA data includes the real-time rotational speed of the blades and the real-time orientation of the blades.

3. The wind turbine blade deformation monitoring method as described in claim 2, characterized in that: The SCADA data also includes wind speed and temperature data.

4. The wind turbine blade deformation monitoring method as described in claim 1, characterized in that: The airflow data includes flow rate and / or velocity.

5. The wind turbine blade deformation monitoring method as described in claim 1, characterized in that: In step S3, a predictive model of blade speed-position-airflow data is constructed based on historical SCADA data and flow meter data. The normal gas velocity and flow rate baseline values ​​of the blade at any speed and position are obtained by calculation and prediction.

6. The wind turbine blade deformation monitoring method as described in claim 5, characterized in that: The prediction model is a machine learning model.

7. The wind turbine blade deformation monitoring method as described in claim 5, characterized in that: The calculation and prediction process also incorporates blade physical property parameters, including blade length, stiffness coefficient, and material elastic modulus. Through fluid dynamics simulation calculations, the baseline values ​​of normal airflow data output by the prediction model are corrected.

8. The wind turbine blade deformation monitoring method as described in claim 1, characterized in that: The threshold is calculated using the upper limit of statistical data based on the safety margin.

9. The wind turbine blade deformation monitoring method as described in claim 1, characterized in that: Every preset period, the corresponding relationship data is updated again using the latest collected normal operation data, and the baseline value and corresponding threshold are updated synchronously.

Images