Offshore wind power operation and maintenance method and system based on digital twinborn technology

Abstract

The invention provides an offshore wind power operation and maintenance method and system based on a digital twinning technology, and relates to the technical field of offshore wind power operation and maintenance methods, and the method comprises the steps: calculating the maximum theoretical power of a fan according to the Betz limit definition; converting a dynamic ratio of wind energy efficiency conversion into a difference equation by adopting forward Euler discretization, and constructing a state space model by adopting Kalman filtering to update the wind energy conversion efficiency of each sampling point; and constructing a dynamic threshold value of the wind energy conversion efficiency by adopting an exponential weighted movement method. The air density is dynamically corrected through a temperature compensation method, so that theoretical power calculation better fits the real atmospheric environment where the fan is located; kalman filtering is adopted to solve the discretized state space model, and prediction and actual observation values of the mechanism model can be optimally fused; a dynamic threshold value is generated by adopting an exponential weighted moving method, and the threshold value can be adaptively adjusted along with the long-term trend and short-term fluctuation of the wind energy conversion efficiency.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power operation and maintenance methods, specifically to offshore wind power operation and maintenance methods and systems based on digital twin technology. Background Technology

[0002] Traditional wind turbine performance monitoring technologies generally have the following limitations: static air density models do not consider the dynamic changes in air density caused by the nacelle temperature gradient, resulting in deviations in power curve calculations; yaw error correction relies on fixed compensation coefficients, which cannot adapt to real-time fluctuations in the angle between the wind direction and the rotor plane; and anomaly detection often uses fixed threshold methods, which are slow to respond to dynamic conditions such as sudden changes in wind speed and turbulence.

[0003] In existing technologies, wind energy conversion efficiency analysis does not separate wind speed-sensitive terms from noise terms, causing minute efficiency fluctuations to be overlooked. Threshold settings rely on historical data statistics, making it impossible to respond in real time to sudden changes in environmental parameters and difficult to capture transient anomalies. This results in a significantly increased rate of missed detection for early faults such as blade icing and bearing wear. Furthermore, the reliance on the linear mapping relationship between wind speed and power fails to consider the need for real-time dynamic adjustment of the angle between wind direction and the rotor plane, leading to a significant increase in yaw power loss compensation errors. Especially in complex marine meteorological environments, existing methods are prone to false alarms due to the coupling effect of environmental parameters and have difficulty distinguishing between short-term disturbances and real faults, resulting in high operation and maintenance costs. Therefore, there is an urgent need for an offshore wind power operation and maintenance method based on digital twin technology.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide an offshore wind power operation and maintenance method and system based on digital twin technology to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The offshore wind power operation and maintenance method based on digital twin technology includes the following specific steps: S1: Based on the temperature of the wind turbine nacelle, the air density coefficient is constructed using the temperature compensation method. The operation and maintenance monitoring window is set, and the wind speed and air density data of the environment where the wind turbine is located are obtained within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window. S2: Based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, construct the yaw error compensation term of the corrected twin power curve, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, construct a sliding exponential filter model for the power at each moment within the monitoring window, and obtain the actual power curve of the wind turbine. S3: Based on the actual power curve and the corrected power curve in the monitoring window, obtain the dynamic ratio curve of the wind turbine in the monitoring window. Differentiate the dynamic ratio curve to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, use forward Euler discretization to transform the differential equation into a difference equation and obtain the wind energy conversion efficiency at each sampling point. S4: Based on the difference equation, a state-space model is constructed using Kalman filtering to update the wind energy conversion efficiency of each sampling point. An exponential weighted moving average method is used to construct a dynamic threshold for the wind energy conversion efficiency. The updated wind energy conversion efficiency of each sampling point is compared with the dynamic threshold, and the abnormal situation of the sampling point is determined based on the result. S5: Calculate the percentage of abnormal occurrences among all sampling points, and also calculate the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number threshold, mark the fan in the monitoring window as having an abnormality.

[0007] Furthermore, the air density coefficient is constructed using a temperature compensation method based on the temperature of the wind turbine nacelle. The specific steps are as follows: the basic air density value is calculated by multiplying the standard atmospheric pressure by the air gas constant and the thermodynamic temperature obtained by adding a fixed constant to the Celsius temperature. A linear correction term is introduced, which is composed of a fixed coefficient and the difference between the current temperature and a specific reference temperature. Finally, the basic air density value is subtracted from this correction term to obtain the final temperature-compensated air density coefficient.

[0008] Further, the corrected power curve after compensation is obtained, and the specific steps are as follows: A monitoring window for operation and maintenance is set up, and wind speed and air density data of the environment where the wind turbine is located are acquired within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window, based on the following formula: ; in, Indicates the Bates limit coefficient; This represents wind speed data; Indicates the radius of the fan rotor; Represents the twin power curve; This represents the compensation air density coefficient. Represents a time variable; The formula used to obtain the compensated corrected power curve is: ; in, This represents the corrected power curve after compensation. This indicates the angle between the wind direction and the impeller plane of the wind turbine rotor. This is the yaw error compensation term.

[0009] Furthermore, obtaining the actual power curve of the wind turbine also includes constructing a sliding exponential filter model to filter the actual power curve of the wind turbine. The specific steps are as follows: Within the monitoring window, the power output of the wind turbine is acquired. For the power data at each moment within the monitoring window, a moving average filter model is constructed to filter the data and obtain the actual power curve. The formula used is: ; in, Indicates the length of the monitoring window; Indicates the forgetting factor; express The output power of the fan at any given time; Represents the integral variable; This represents the actual power curve.

[0010] Furthermore, the differential equations containing wind speed-sensitive terms and noise terms are obtained through the following steps: The dynamic ratio curve of the wind turbine within the monitoring window is obtained by taking the ratio of the actual power curve and the corrected power curve of the wind turbine. ; in, express The dynamic ratio curve at any given time; right By taking the derivative, we obtain the differential equation containing wind speed sensitivity terms and noise terms: ; in, ; ; express Wind speed sensitivity at any given time; express Noise term at any given time; Indicates the filtered result The actual power at any given moment.

[0011] Furthermore, the wind energy conversion efficiency at each sampling point is obtained through the following steps: For differential equations, in the monitoring window Internal sampling interval is set to The differential equations are transformed into difference equations using forward Euler discretization to obtain the wind energy conversion efficiency at each sampling point. ; in, ; in, Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Sensitive parameters for wind speed changes at each sampling point; Indicates the sampling interval; Indicates the number of sampling points; Indicates the first The noise term at each sampling point.

[0012] Furthermore, a state-space model is constructed using Kalman filtering to update the wind energy conversion efficiency at each sampling point. The specific steps are as follows: Based on the difference equation, a Kalman filter is used to construct a state-space model to update the wind energy conversion efficiency at each sampling point: ; in, This represents observation noise, which follows a normal distribution. Indicates the first The updated wind energy conversion efficiency at each sampling point.

[0013] Furthermore, an exponentially weighted moving average method is used to construct a dynamic threshold for wind energy conversion efficiency. The specific steps are as follows: First, initialize the mean and variance: ; ; in, This represents the initial exponentially weighted mean; This represents the initial exponentially weighted standard deviation; This represents the wind energy conversion efficiency at the 0th sampling point; The initial standard deviation setting for wind energy conversion efficiency is set based on expert experience. Update the exponentially weighted mean and standard deviation: ; ; in, Indicates the first The exponentially weighted mean of the sampling points; Indicates the first The exponentially weighted mean of the sampling points; Indicates the smoothing factor; Indicates the first The exponentially weighted standard deviation of each sampling point; Indicates the first -1 index-weighted standard deviation of sampling points; Calculate the dynamic threshold: ; ; in, Indicates the first Dynamic upper threshold for each sampling point; Indicates the first Dynamic threshold for each sampling point; Indicates the sensitivity coefficient; The wind energy conversion efficiency is compared with a dynamic threshold, and the corresponding operation and maintenance information is output. ; in, Indicates the first Signal classification results for each sampling point Indicates the first The fans at each sampling point malfunctioned. Indicates the first The fans at all sampling points are in normal condition.

[0014] Furthermore, the fan in the monitoring window is marked as abnormal. The specific steps are as follows: Within the monitoring window, count all the times the signal classification result is 1, and divide by the total number of sampling points within the monitoring window to calculate the percentage of abnormal occurrences. At the same time, count the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number of occurrences threshold, mark the wind turbine within the monitoring window as having an abnormality.

[0015] The present invention also provides an offshore wind power operation and maintenance system based on digital twin technology, the system being used to execute the above-described operation and maintenance methods, including: The twin power curve construction module is used to construct the air density coefficient based on the temperature compensation method of the wind turbine nacelle, set the operation and maintenance monitoring window and acquire the wind speed and air density data of the environment where the wind turbine is located within the monitoring window, and use the Bates limit and the air density coefficient as a correction term to determine the twin power curve of the wind turbine within the monitoring window. The actual power curve construction module is used to construct the yaw error compensation term of the corrected twin power curve based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, and construct a sliding exponential filter model for the power at each moment within the monitoring window to filter and obtain the actual power curve of the wind turbine. The wind energy conversion efficiency acquisition module is used to obtain the dynamic ratio curve of the wind turbine in the monitoring window based on the actual power curve and the corrected power curve in the monitoring window. The dynamic ratio curve is differentiated to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, the differential equation is discretized by forward Euler to transform it into a difference equation and obtain the wind energy conversion efficiency at each sampling point. The classification results module is used to update the wind energy conversion efficiency of each sampling point by constructing a state space model based on the difference equation and Kalman filtering. It also uses the exponential weighted moving average method to construct a dynamic threshold for the wind energy conversion efficiency, compares the updated wind energy conversion efficiency of each sampling point with the dynamic threshold, and determines the abnormal situation of the sampling point based on the result. The anomaly detection module is used to count the percentage of anomalies in all sampling points, and to count the maximum number of consecutive anomalies in each sampling point. When the percentage of anomalies exceeds a preset anomaly occurrence rate threshold or the maximum number of consecutive anomalies exceeds a preset maximum number threshold, the fan in the monitoring window is marked as having an anomaly.

[0016] Compared with the prior art, the beneficial effects of the present invention are: Based on temperature compensation, the Bates limit correction and yaw error compensation power curve reconstruction reduces the air density coefficient error and improves the accuracy of yaw power loss compensation. It overcomes the temperature sensitivity defect of the static air density model, achieves dynamic compensation through temperature-density nonlinear mapping, realizes real-time capture of small fluctuations in wind energy conversion efficiency through dynamic ratio curve differential modeling, reduces the false alarm rate of anomaly detection by predicting and weighting the observation data through the Kalman gain dynamic balance model, and reduces the missed detection rate of intermittent faults by setting a threshold for the number of consecutive anomalies. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2This is a graph showing the relationship between the included angle of the impeller plane of the wind turbine rotor and the corrected power after compensation. Figure 3 This is a schematic diagram of the overall system of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example: Please see Figures 1-2 The present invention provides a technical solution: The offshore wind power operation and maintenance method based on digital twin technology includes the following specific steps: S1: Based on the temperature of the wind turbine nacelle, the air density coefficient is constructed using the temperature compensation method. The operation and maintenance monitoring window is set, and the wind speed and air density data of the environment where the wind turbine is located are obtained within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window. The air density coefficient is constructed using a temperature compensation method based on the temperature of the wind turbine nacelle. The specific steps are as follows: The basic air density value is calculated by dividing the standard atmospheric pressure by the product of the air gas constant and the thermodynamic temperature obtained by adding a fixed constant to the Celsius temperature. A linear correction term is then introduced, which is composed of a fixed coefficient and the difference between the current temperature and a specific reference temperature. Finally, the basic air density value is subtracted from this correction term to obtain the final temperature-compensated air density coefficient.

[0021] The formula upon which the above process is based is: ; in, This represents the final air density coefficient after temperature compensation; Indicates standard atmospheric pressure; Indicates the current temperature; This represents the air gas constant.

[0022] In the above process, The physical meaning is the dynamic air density value that eliminates the influence of temperature drift. It quantifies the second-order effect of temperature deviation on air density in real time through a linear compensation term, and uses the denominator term... By correcting the main terms of the ideal gas law and superimposing linear compensation terms, the accuracy of the unit power curve prediction is improved.

[0023] molecular and This constitutes the reference term of the ideal gas law, reflecting the fundamental influence of atmospheric conditions on density, with reference temperature. The calibration temperature point corresponding to the ISO standard air density, and The relationship is inversely proportional; the smaller the gas constant, the greater the density value after compensation. As the temperature increases, the denominator term... The increase in both the increase in the compensation term value and the increase in the dual effect lead to It exhibits a non-linear decrease, precisely matching the thermodynamic properties of the real environment.

[0024] S2: Based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, construct the yaw error compensation term of the corrected twin power curve, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, construct a sliding exponential filter model for the power at each moment within the monitoring window, and obtain the actual power curve of the wind turbine. The specific steps for obtaining the compensated corrected power curve are as follows: A monitoring window for operation and maintenance is set up, and wind speed and air density data of the environment where the wind turbine is located are acquired within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window, based on the following formula: ; in, Indicates the Bates limit coefficient; This represents wind speed data; Indicates the radius of the fan rotor; Represents the twin power curve; This represents the compensation air density coefficient. Represents a time variable; In the above process, by sensing the attenuation effect of temperature changes on air density in real time, the power calculation benchmark is automatically corrected, the prediction deviation caused by seasonal temperature differences is eliminated, and the principle of wind energy capture physical limit is introduced. When encountering sudden strong winds, the built-in upper limit threshold of energy conversion efficiency is automatically activated, preventing abnormal spikes in power prediction values ​​caused by sudden changes in wind speed. This ensures that the theoretical power prediction value under extreme weather conditions such as typhoons and storms remains stable within the physical carrying capacity of the equipment. A slip monitoring window is constructed to accurately capture transient characteristics such as short-term gusts and turbulent pulses, thereby improving the dynamic response speed.

[0025] The physical meaning is that the time series model of the theoretical output power of the wind turbine is corrected by real-time air density, eliminating the power prediction deviation caused by temperature fluctuations in the traditional static density model, and using the Bates limit coefficient. Constraining the maximum theoretical power, combined with the density correction term To improve the sensitivity of wind turbine performance testing. By employing a temperature-compensated dynamic density coefficient, the impact of the coupling effect of temperature and humidity on power is quantified, replacing the traditional fixed density value and avoiding overestimation or underestimation of power prediction due to environmental changes. It reflects the cubic amplification effect of wind speed on power, and monitors the dynamic characteristics of the model in response to wind speed fluctuations by real-time data input. This constant represents the theoretical maximum wind energy utilization factor, and it limits the theoretical upper limit of power. and The relationship is proportional to the cube; doubling the wind speed results in an eightfold increase in theoretical power. It is linearly positively correlated with power.

[0026] The formula used to obtain the compensated corrected power curve is: ; in, This represents the corrected power curve after compensation. This indicates the angle between the wind direction and the impeller plane of the wind turbine rotor. This is the yaw error compensation term.

[0027] In the above process, in response to the short-term yaw lag phenomenon caused by gust change, the instantaneous energy loss is captured by tracking the change in the angle between the wind direction and the impeller plane in real time. By establishing a continuous comparison channel between the corrected power and the measured value, a new dimension is provided for monitoring the health status of the yaw system. The effective windward projected area of ​​the rotor system is directly affected. When the fan is not aligned with the prevailing wind direction, the cosine cube term of this angle... The nonlinear decay characteristics of aerodynamic efficiency were accurately characterized, and its cubic relationship originated from the correlation between the third derivative of kinetic energy conversion efficiency and the angle of attack in fluid mechanics. and It exhibits strict reverse monotonicity when the yaw deviation angle Increase The value decays exponentially, and the corrected theoretical power prediction decreases non-linearly.

[0028] In the above embodiments, 20 sets of data on the angle between the wind direction and the impeller plane of the wind turbine rotor and the compensated corrected power are given to reflect the change of the compensated corrected power as the angle between the wind direction and the impeller plane of the wind turbine rotor changes, as shown in Table 1: Table 1: Relationship between the angle between the wind direction and the impeller plane of the wind turbine rotor and the compensated corrected power.

[0029] As can be seen from Table 1 above, given , and It exhibits strict inverse monotonicity, and the corrected theoretical power prediction decreases nonlinearly.

[0030] Obtaining the actual power curve of the wind turbine also includes constructing a sliding exponential filter model to filter the actual power curve of the wind turbine. The specific steps are as follows: Within the monitoring window, the power output of the wind turbine is acquired. For the power data at each moment within the monitoring window, a moving average filter model is constructed to filter the data and obtain the actual power curve. The formula used is: ; in, Indicates the length of the monitoring window; Indicates the forgetting factor; express The output power of the fan at any given time; Represents the integral variable; This represents the actual power curve.

[0031] In the above process, The equivalent power sequence after dynamic weight optimization is essentially an energy integral representation based on memory decay characteristics, and the forgetting window is also relevant. Defining the maximum scope of historical data tracing using the forgetting factor The dynamic adjustment and synergistic effect of these two components construct a time-varying filter system, which works together to achieve a balance between high-frequency noise attenuation and step response. and Showing a positive correlation, Increased power leads to accelerated memory decay, but also improves the sensitivity of the actual power curve to new operating conditions. and Showing an inverse correlation, The extension of the period and the increased participation of historical data result in a smoother power curve.

[0032] S3: Based on the actual power curve and the corrected power curve in the monitoring window, obtain the dynamic ratio curve of the wind turbine in the monitoring window. Differentiate the dynamic ratio curve to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, use forward Euler discretization to transform the differential equation into a difference equation and obtain the wind energy conversion efficiency at each sampling point. The specific steps for obtaining the differential equation with wind speed sensitive terms and noise terms are as follows: The dynamic ratio curve of the wind turbine within the monitoring window is obtained by taking the ratio of the actual power curve and the corrected power curve of the wind turbine. ; in, express The dynamic ratio curve at any given time; right By taking the derivative, we obtain the differential equation containing wind speed sensitivity terms and noise terms: ; in, ; ; express Wind speed sensitivity at any given time; express Noise term at any given time; Indicates the filtered result The actual power at any given moment.

[0033] In the above process, It directly characterizes the sensitivity of wind speed changes to the dynamic ratio, and its physical meaning is the dynamic ratio attenuation rate caused by a unit change in wind speed. In the denominator, the noise amplitude is converted into a ratio relative to the theoretical power, and the exponential decay factor is used. Quantify the contribution weight of historical noise to the current moment.

[0034] The specific steps for obtaining the wind energy conversion efficiency at each sampling point are as follows: For differential equations, in the monitoring window Internal sampling interval is set to The differential equations are transformed into difference equations using forward Euler discretization to obtain the wind energy conversion efficiency at each sampling point. ; in, ; in, Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Sensitive parameters for wind speed changes at each sampling point; Indicates the sampling interval; Indicates the number of sampling points; Indicates the first The noise term at each sampling point.

[0035] In the above process, The dynamic energy conversion efficiency factor, characterized in the discretized time domain, is essentially a composite state variable encompassing inertial decay, wind speed sensitivity, and noise coupling. The item inherits the valid state information from the previous time step, and the weight coefficients are... Map the wind speed gradient of 3 m / s to fluctuation, The item quantifies the residual impact of historical power fluctuations; Its coefficient is used as an inertial reference term. This constitutes a first-order inertial element. The rate of change of the axial inducible factor is directly correlated, converting wind speed increments into contributions to aerodynamic torque fluctuations. Characterizing the discretization accumulation effect of unattenuated noise within the time window, and Showing an inverse correlation, As the value increases, the discrete truncation error rises, and the efficiency factor tracking lag increases. and Showing a negative correlation, As the number of states increases, the retention rate of historical states decreases, and the transient response speed increases. and Showing a positive correlation, As the absolute value increases, the modulating effect of wind speed changes on the efficiency factor strengthens.

[0036] S4: Based on the difference equation, a state-space model is constructed using Kalman filtering to update the wind energy conversion efficiency of each sampling point. An exponential weighted moving average method is used to construct a dynamic threshold for the wind energy conversion efficiency. The updated wind energy conversion efficiency of each sampling point is compared with the dynamic threshold, and the abnormal situation of the sampling point is determined based on the result. The specific steps for using Kalman filtering to construct a state-space model and update the wind energy conversion efficiency at each sampling point are as follows: Based on the difference equation, a Kalman filter is used to construct a state-space model to update the wind energy conversion efficiency at each sampling point: ; in, This represents observation noise, which follows a normal distribution. Indicates the first The updated wind energy conversion efficiency at each sampling point.

[0037] The specific steps for constructing the dynamic threshold for wind energy conversion efficiency using the exponentially weighted moving average method are as follows: First, initialize the mean and variance: ; ; in, This represents the initial exponentially weighted mean; This represents the initial exponentially weighted standard deviation; This represents the wind energy conversion efficiency at the 0th sampling point; The initial standard deviation setting for wind energy conversion efficiency is set based on expert experience. Update the exponentially weighted mean and standard deviation: ; ; in, Indicates the first The exponentially weighted mean of the sampling points; Indicates the first The exponentially weighted mean of the sampling points; Indicates the smoothing factor; Indicates the first The exponentially weighted standard deviation of each sampling point; Indicates the first -1 index-weighted standard deviation of sampling points; During the above process, based on the exponentially weighted average formula, the system generates a dynamic trend center line for energy conversion efficiency in real time. During the phase of gradual wind speed change, the smoothing factor It significantly suppresses transient noise interference, reduces the standard deviation of trend line fluctuations, and in scenarios with sudden wind speed changes, the current observation value is... It rapidly injects new information, solving the lag problem of traditional static thresholds under varying operating conditions. The threshold bandwidth is dynamically adjusted through recursive calculation when the observed value When there is a sudden increase, the variance response amplitude increases within 3 periods; The dynamic trend centerline characterizing the efficacy factor is obtained through... Smooth the weighting of historical data and current observations. The dynamic tolerance range for quantified performance fluctuations is adaptively adjusted through recursive calculation of the threshold bandwidth. As a historical trend benchmark, its weight To control sensitivity to sudden anomalies, current observations are used... Injecting new information into the coefficients gives the mean update the characteristics of inertial filtering. The fluctuation memory effect is formed through the transmission of historical variance. During a sudden increase, The response amplitude increases within the cycle. and Showing a positive correlation, Increased weighting of historical data enhances the smoothness of trend lines. and Showing a positive correlation, As the threshold bandwidth increases, the immediate impact of new fluctuations on the threshold bandwidth is amplified.

[0038] Calculate the dynamic threshold: ; ; in, Indicates the first Dynamic upper threshold for each sampling point; Indicates the first Dynamic threshold for each sampling point; Indicates the sensitivity coefficient; In the above process, and The dynamic upper and lower limit thresholds characterizing the performance factor are constructed based on historical mean and standard deviation to create an adaptive volatility tolerance range. Coefficient adjustment threshold bandwidth As a threshold center benchmark, its historical trend transmission gives the threshold band inertial tracking characteristics. The volatility memory effect is transmitted through historical standard deviation. As an externally adjustable parameter, it directly controls the sensitivity of the threshold to fluctuations. and and Positive correlation The increase in efficiency shifts the threshold band upwards overall, adapting to seasonal trends in energy conversion efficiency. and and Positive correlation Increase, threshold bandwidth expansion, and and Positive correlation As the threshold band increases, it widens symmetrically.

[0039] The wind energy conversion efficiency is compared with a dynamic threshold, and the corresponding operation and maintenance information is output. ; in, Indicates the first Signal classification results for each sampling point Indicates the first The fans at each sampling point malfunctioned. Indicates the first The fans at all sampling points are in normal condition.

[0040] During the above process, when the wind energy conversion efficiency is greater than the dynamic upper threshold or less than the lower threshold, the operation and maintenance information is marked as 1 to indicate an abnormal situation; otherwise, it is marked as 0 to indicate a normal state.

[0041] S5: Calculate the percentage of abnormal occurrences among all sampling points, and also calculate the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number threshold, mark the fan in the monitoring window as having an abnormality.

[0042] The specific steps for marking the fan in the monitoring window as abnormal are as follows: Within the monitoring window, count all the times the signal classification result is 1, and divide by the total number of sampling points within the monitoring window to calculate the percentage of abnormal occurrences. At the same time, count the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number of occurrences threshold, mark the wind turbine within the monitoring window as having an abnormality.

[0043] In the above process, the long-term monitoring data of the equipment under normal conditions is collected, covering at least 12 months, including different seasons and wind speed ranges. The distribution of the proportion of abnormal occurrences in each monitoring window and the number of consecutive abnormal occurrences in each detection window are calculated. The 99th percentile of the proportion of abnormal occurrences in normal data is taken as the upper limit of the threshold, which are respectively used as the preset abnormal occurrence rate threshold and the preset maximum number of occurrences threshold.

[0044] Please see Figure 3 The present invention also provides an offshore wind power operation and maintenance system based on digital twin technology, the system being used to execute the above-described operation and maintenance methods, including: The twin power curve construction module is used to construct the air density coefficient based on the temperature compensation method of the wind turbine nacelle, set the operation and maintenance monitoring window and acquire the wind speed and air density data of the environment where the wind turbine is located within the monitoring window, and use the Bates limit and the air density coefficient as a correction term to determine the twin power curve of the wind turbine within the monitoring window. The actual power curve construction module is used to construct the yaw error compensation term of the corrected twin power curve based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, and construct a sliding exponential filter model for the power at each moment within the monitoring window to filter and obtain the actual power curve of the wind turbine. The wind energy conversion efficiency acquisition module is used to obtain the dynamic ratio curve of the wind turbine in the monitoring window based on the actual power curve and the corrected power curve in the monitoring window. The dynamic ratio curve is differentiated to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, the differential equation is discretized by forward Euler to transform it into a difference equation and obtain the wind energy conversion efficiency at each sampling point. The classification results module is used to update the wind energy conversion efficiency of each sampling point by constructing a state space model based on the difference equation and Kalman filtering. It also uses the exponential weighted moving average method to construct a dynamic threshold for the wind energy conversion efficiency, compares the updated wind energy conversion efficiency of each sampling point with the dynamic threshold, and determines the abnormal situation of the sampling point based on the result. The anomaly detection module is used to count the percentage of anomalies in all sampling points, and to count the maximum number of consecutive anomalies in each sampling point. When the percentage of anomalies exceeds a preset anomaly occurrence rate threshold or the maximum number of consecutive anomalies exceeds a preset maximum number threshold, the fan in the monitoring window is marked as having an anomaly.

[0045] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0046] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0047] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0048] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for offshore wind power operation and maintenance based on digital twin technology, characterized by the following steps: include: S1: Based on the temperature of the wind turbine nacelle, the air density coefficient is constructed using the temperature compensation method. The operation and maintenance monitoring window is set, and the wind speed and air density data of the environment where the wind turbine is located are obtained within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window. S2: Based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, construct the yaw error compensation term of the corrected twin power curve, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, construct a sliding exponential filter model for the power at each moment within the monitoring window, and obtain the actual power curve of the wind turbine. S3: Based on the actual power curve and the corrected power curve in the monitoring window, obtain the dynamic ratio curve of the wind turbine in the monitoring window. Differentiate the dynamic ratio curve to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, use forward Euler discretization to transform the differential equation into a difference equation and obtain the wind energy conversion efficiency at each sampling point. S4: Based on the difference equation, a state-space model is constructed using Kalman filtering to update the wind energy conversion efficiency of each sampling point. An exponential weighted moving average method is used to construct a dynamic threshold for the wind energy conversion efficiency. The updated wind energy conversion efficiency of each sampling point is compared with the dynamic threshold, and the abnormal situation of the sampling point is determined based on the result. S5: Calculate the percentage of abnormal occurrences among all sampling points, and also calculate the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number threshold, mark the fan in the monitoring window as having an abnormality.

2. The offshore wind power operation and maintenance method based on digital twin technology according to claim 1, characterized in that, The air density coefficient is constructed using a temperature compensation method based on the temperature of the wind turbine nacelle. The specific steps are as follows: The basic air density value is calculated by dividing the standard atmospheric pressure by the product of the air gas constant and the thermodynamic temperature obtained by adding a fixed constant to the Celsius temperature. A linear correction term is then introduced, which is composed of a fixed coefficient and the difference between the current temperature and a specific reference temperature. Finally, the basic air density value is subtracted from this correction term to obtain the final temperature-compensated air density coefficient.

3. The offshore wind power operation and maintenance method based on digital twin technology according to claim 2, characterized in that, The specific steps for obtaining the compensated corrected power curve are as follows: A monitoring window for operation and maintenance is set up, and wind speed and air density data of the environment where the wind turbine is located are acquired within the monitoring window. The Bates limit is used, and the air density coefficient is used as a correction term to determine the twin power curve of the wind turbine within the monitoring window, based on the following formula: ； in, Indicates the Bates limit coefficient; This represents wind speed data; Indicates the radius of the fan rotor; Represents the twin power curve; This represents the compensation air density coefficient. Represents a time variable; The formula used to obtain the compensated corrected power curve is: ； in, This represents the corrected power curve after compensation. This indicates the angle between the wind direction and the impeller plane of the wind turbine rotor. This is the yaw error compensation term.

4. The offshore wind power operation and maintenance method based on digital twin technology as described in claim 3, characterized in that, Obtaining the actual power curve of the wind turbine also includes constructing a sliding exponential filter model to filter the actual power curve of the wind turbine. The specific steps are as follows: Within the monitoring window, the power output of the wind turbine is acquired. For the power data at each moment within the monitoring window, a moving average filter model is constructed to filter the data and obtain the actual power curve. The formula used is: ； in, Indicates the length of the monitoring window; Indicates the forgetting factor; express The output power of the fan at any given time; Represents the integral variable; This represents the actual power curve.

5. The offshore wind power operation and maintenance method based on digital twin technology according to claim 4, characterized in that, The specific steps for obtaining the differential equation with wind speed sensitive terms and noise terms are as follows: The dynamic ratio curve of the wind turbine within the monitoring window is obtained by taking the ratio of the actual power curve and the corrected power curve of the wind turbine. ； in, express The dynamic ratio curve at any given time; This represents the corrected power curve after compensation. right By taking the derivative, we obtain the differential equation containing wind speed sensitivity terms and noise terms: ； in, ； ； express Wind speed sensitivity at any given time; express Noise term at any given time; Indicates the filtered result Actual power at any given time; This represents wind speed data; This represents the twin power curve.

6. The offshore wind power operation and maintenance method based on digital twin technology according to claim 5, characterized in that, The specific steps for obtaining the wind energy conversion efficiency at each sampling point are as follows: For differential equations, in the monitoring window Internal sampling interval is set to The differential equations are transformed into difference equations using forward Euler discretization to obtain the wind energy conversion efficiency at each sampling point. ； in, ； in, Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind energy conversion efficiency at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Wind speed measurements at each sampling point; Indicates the first Sensitive parameters for wind speed changes at each sampling point; Indicates the sampling interval; Indicates the number of sampling points; Indicates the first The noise term at each sampling point.

7. The offshore wind power operation and maintenance method based on digital twin technology according to claim 6, characterized in that, The specific steps for using Kalman filtering to construct a state-space model and update the wind energy conversion efficiency at each sampling point are as follows: Based on the difference equation, a Kalman filter is used to construct a state-space model to update the wind energy conversion efficiency at each sampling point: ； in, This represents observation noise, which follows a normal distribution. Indicates the first The updated wind energy conversion efficiency at each sampling point.

8. The offshore wind power operation and maintenance method based on digital twin technology according to claim 7, characterized in that, The specific steps for constructing the dynamic threshold for wind energy conversion efficiency using the exponentially weighted moving average method are as follows: First, initialize the mean and variance: ； ； in, This represents the initial exponentially weighted mean; This represents the initial exponentially weighted standard deviation; This represents the wind energy conversion efficiency at the 0th sampling point; The initial standard deviation setting for wind energy conversion efficiency is set based on expert experience. Update the exponentially weighted mean and standard deviation: ； ； in, Indicates the first The exponentially weighted mean of the sample points; Indicates the first The exponentially weighted mean of the sample points; Indicates the smoothing factor; Indicates the first The exponentially weighted standard deviation of each sampling point; Indicates the first -1 index-weighted standard deviation of sampling points; Calculate the dynamic threshold: ； ； in, Indicates the first Dynamic upper threshold for each sampling point; Indicates the first Dynamic threshold for each sampling point; Indicates the sensitivity coefficient; The wind energy conversion efficiency is compared with a dynamic threshold, and the corresponding operation and maintenance information is output. ； in, Indicates the first Signal classification results for each sampling point Indicates the first The fans at each sampling point malfunctioned. Indicates the first The fans at all sampling points are in normal condition.

9. The offshore wind power operation and maintenance method based on digital twin technology according to claim 1, characterized in that, The specific steps for marking the fan in the monitoring window as abnormal are as follows: Within the monitoring window, count all the times the signal classification result is 1, and divide by the total number of sampling points within the monitoring window to calculate the percentage of abnormal occurrences. At the same time, count the maximum number of consecutive abnormal occurrences for each sampling point. When the percentage of abnormal occurrences exceeds the preset abnormal occurrence rate threshold or the maximum number of consecutive abnormal occurrences exceeds the preset maximum number of occurrences threshold, mark the wind turbine within the monitoring window as having an abnormality.

10. An offshore wind power operation and maintenance system based on digital twin technology, characterized in that: The operation and maintenance system is used to execute the operation and maintenance method according to any one of claims 1-9, including: The twin power curve construction module is used to construct the air density coefficient based on the temperature compensation method of the wind turbine nacelle, set the operation and maintenance monitoring window and acquire the wind speed and air density data of the environment where the wind turbine is located within the monitoring window, and use the Bates limit and the air density coefficient as a correction term to determine the twin power curve of the wind turbine within the monitoring window. The actual power curve construction module is used to construct the yaw error compensation term of the corrected twin power curve based on the angle between the wind direction and the impeller plane of the wind turbine rotor within the monitoring window, obtain the compensated corrected power curve, synchronously collect the power output of the wind turbine within the monitoring window, and construct a sliding exponential filter model for the power at each moment within the monitoring window to filter and obtain the actual power curve of the wind turbine. The wind energy conversion efficiency acquisition module is used to obtain the dynamic ratio curve of the wind turbine in the monitoring window based on the actual power curve and the corrected power curve in the monitoring window. The dynamic ratio curve is differentiated to obtain a differential equation with wind speed sensitive terms and noise terms. Based on the set sampling interval, the differential equation is discretized by forward Euler to transform it into a difference equation and obtain the wind energy conversion efficiency at each sampling point. The classification results module is used to update the wind energy conversion efficiency of each sampling point by constructing a state space model based on the difference equation and Kalman filtering. It also uses the exponential weighted moving average method to construct a dynamic threshold for the wind energy conversion efficiency, compares the updated wind energy conversion efficiency of each sampling point with the dynamic threshold, and determines the abnormal situation of the sampling point based on the result. The anomaly detection module is used to count the percentage of anomalies in all sampling points, and to count the maximum number of consecutive anomalies in each sampling point. When the percentage of anomalies exceeds a preset anomaly occurrence rate threshold or the maximum number of consecutive anomalies exceeds a preset maximum number threshold, the fan in the monitoring window is marked as having an anomaly.

Images