A wind power tower control method and system based on fatigue damage driving

Abstract

The application discloses a wind power generation tower control method and system based on fatigue damage driving, relates to the technical field of wind power generation control, and comprises the following steps: step S1, collecting tower drum structure response data, three-dimensional wind field parameters and power grid scheduling instructions of each wind power generation tower in a wind power plant through a distributed sensing network in real time; step S2, constructing a tower drum structure digital twin model, and dynamically predicting the residual fatigue life and damage evolution trend of each tower drum key position based on real-time structure response data; step S3, performing double-layer collaborative optimization control; and step S4, issuing the optimization instruction to an executing mechanism, and updating the digital twin model through real-time feedback data. The wind power generation tower control method and system based on fatigue damage driving have the advantages that the safety operation guarantee, fatigue damage inhibition and power generation efficiency collaborative optimization of the wind power generation tower are realized, and the unit life and power generation performance are balanced.

Description

Technical Field

[0001] This invention relates to the field of wind power generation control technology, and in particular to a wind power tower control method and system based on fatigue damage. Background Technology

[0002] Traditional control strategies aim to maximize the power of a single unit, neglecting the cumulative fatigue effect of the tower, leading to an increased risk of fracture during service life. Monitoring with only finite strain gauges, with a sampling rate of less than 1Hz, cannot capture transient resonances.

[0003] The wake effect between units in the wind farm causes power generation loss. The existing yaw control only passively avoids the problem and does not consider the downstream tower structure. The structural damage diagnosis and power control are decoupled, and the control command generation delay exceeds 60 seconds, which is difficult to adapt to the grid frequency regulation requirements. Summary of the Invention

[0004] The purpose of this invention is to provide a wind power tower control method and system based on fatigue damage. It addresses the problems in traditional wind power tower control, such as high risk of tower fatigue accumulation, insufficient monitoring of transient resonance, inadequate avoidance of wake effects, and control response delay. Through multi-source sensing and dynamic optimization technology, it achieves safe operation assurance, fatigue damage suppression, and synergistic optimization of power generation efficiency of wind power towers, balancing unit life and power generation performance.

[0005] This invention provides a wind power tower control method and system based on fatigue damage, including the following steps: Step S1, real-time acquisition of tower structure response data, three-dimensional wind field parameters and power grid dispatch instructions of each wind power tower in the wind farm through a distributed sensor network;

[0006] Step S2: Construct a digital twin model of the tower structure and dynamically predict the remaining fatigue life and damage evolution trend of each key part of the tower based on real-time structural response data;

[0007] Step S3: Execute two-layer collaborative optimization control;

[0008] Local control layer: Based on the vibration mode prediction results output by the digital twin model of the tower, an adaptive model predictive control algorithm is used to generate pitch and damper adjustment commands to suppress tower resonance;

[0009] Central Coordination Layer: Using the remaining fatigue life of each tower as a constraint, dynamically adjust the power distribution and yaw angle of the unit to make the wake centerline offset by ≥15° relative to the incoming wind direction in order to avoid high fatigue risk areas.

[0010] Step S4: Issue optimization instructions to the execution mechanism and update the digital twin model through real-time feedback data.

[0011] Preferably, in step S1, the distributed sensor network includes a high-density fiber optic grating sensor array, a pulsed lidar, and temperature, humidity, vibration, and lightning strike monitoring units. The high-density fiber optic grating sensor array is installed on the inner wall of the tower, with one set of high-density fiber optic grating sensor arrays arranged every 90 degrees along the circumference of the tower. The pulsed lidar is deployed at the leading edge and wake region of the wind farm. The temperature, humidity, vibration, and lightning strike monitoring units are located at the top of the tower.

[0012] Preferably, in step S1, the data sampling frequency of the fiber Bragg grating sensor array is dynamically adjusted.

[0013] Preferably, in step S3, the adaptive model predictive control algorithm includes predicting the tower top displacement response spectrum within the next 30 seconds based on a digital twin model, and solving the objective function to generate the optimal pitch sequence, as shown in the following equation:

[0014]

[0015] Where a(t) is the real-time vibration acceleration; Δθ(k) is the change in pitch angle at time k; ω1 is the vibration suppression weight coefficient; ω2 is the pitch action penalty weight; and N is the prediction time domain.

[0016] Preferably, in step S3, the optimization function of the central coordination layer is defined as follows:

[0017]

[0018] Among them, P i D represents the real-time power generation of the i-th wind turbine generator; j Let λ be the fatigue damage increment of high-risk unit j; λ be the damage penalty coefficient; and H be the set of high-risk units.

[0019] The set of high-risk units H is shown in the following formula:

[0020]

[0021] k is the number of the wind turbine generator set; This represents the percentage of remaining life of unit k. This is the arithmetic average of the remaining lifespan of all units in the field.

[0022] Preferably, in step S3, the wake path offset is achieved by calculating the upstream unit yaw compensation angle Δyaω, as shown in the following formula:

[0023]

[0024] Where, d safe L represents the safe offset distance calculated based on the location of the high-risk tower; L is the unit spacing.

[0025] Preferably, in step S3, the stress change rate at key measuring points is calculated in real time as dσ / dt.

[0026] Preferably, in step S2, the digital twin model includes a parameterized structural mechanics model based on finite element analysis, a fatigue life predictor trained by an LSTM-Transformer neural network, and a parameter self-calibration module that receives control verification data.

[0027] Preferably, in step S3, the central coordination layer implements a tower life balancing strategy: when the remaining life of the tower... At that time, its power setpoint P set The adjustment is as follows:

[0028]

[0029] Among them, P rated The rated power of the unit; L rem The percentage of the unit's remaining service life;

[0030] At the same time, load reduction will be transferred to The generator set.

[0031] Preferably, it includes a sensing layer, an edge computing layer, a central decision-making layer, and an execution layer; a real-time data bus with end-to-end latency connects each layer; the sensors include a hardware network of fiber optic sensor arrays, pulsed lidar groups, and environmental monitoring units; the edge computing layer includes local controllers deployed on each tower, configured with dynamically adaptable processors to execute vibration suppression algorithms; the central decision-making layer includes a server for a spatiotemporal database engine and an optimization solver; the execution layer includes a pitch system, an active damper, and a yaw actuator.

[0032] Therefore, the present invention adopts the above-mentioned wind power tower control method and system based on fatigue damage, which addresses the problems of high tower fatigue accumulation risk, insufficient transient resonance monitoring, insufficient avoidance of wake effects, and control response delay in traditional wind power tower control. Through multi-source sensing and dynamic optimization technology, it realizes the safe operation guarantee, fatigue damage suppression and power generation efficiency synergistic optimization of wind power tower, and balances the unit life and power generation performance.

[0033] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the overall process of a wind power tower control method based on fatigue damage driven by the present invention.

[0035] Figure 2 This is a schematic diagram of the specific process in step S2 of the wind power tower control method based on fatigue damage driven by the present invention.

[0036] Figure 3 This is a system block diagram of a wind power tower control system based on fatigue damage driven by the present invention. Detailed Implementation

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0038] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0039] The terms "first," "second," and similar words 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 after 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.

[0040] Example 1

[0041] like Figures 1-3 As shown, the present invention discloses a wind power tower control method and system based on fatigue damage, comprising the following steps:

[0042] Step S1: Collect in real time the tower structure response data, three-dimensional wind field parameters and power grid dispatch instructions of each wind power tower in the wind farm through a distributed sensor network.

[0043] In step S1, the distributed sensor network includes a high-density fiber optic grating sensor array, a pulsed lidar, and temperature, humidity, vibration, and lightning strike monitoring units. The high-density fiber optic grating sensor array is deployed on the inner wall of the tower, with one array arranged every 90° along the circumference of the tower and an axial spacing of ≤0.5m. The distance between the sensor and the tower weld is ≤10mm. The pulsed lidar is deployed at the leading edge and wake of the wind farm. The temperature, humidity, vibration, and lightning strike monitoring units are located at the top of the tower.

[0044] In step S1, the data sampling frequency of the fiber optic grating sensor array is dynamically adjusted; when the vibration amplitude is detected to exceed the dynamic fatigue damage threshold of the tower, the sampling frequency is increased from the basic 1Hz to 10-50Hz.

[0045] Step S2: Construct a digital twin model of the tower structure and dynamically predict the remaining fatigue life and damage evolution trend of each key part of the tower based on real-time structural response data.

[0046] In step S2, the digital twin model includes a parameterized structural mechanics model based on finite element analysis, a fatigue life predictor trained by an LSTM-Transformer neural network, and a parameter self-calibration module that receives control verification data.

[0047] Step S3: Execute two-layer collaborative optimization control;

[0048] Local control layer: Based on the vibration mode prediction results output by the digital twin model of the tower, an adaptive model predictive control algorithm is used to generate pitch and damper adjustment commands to suppress tower resonance.

[0049] In step S3, the adaptive model predictive control algorithm includes predicting the tower top displacement response spectrum within the next 30 seconds based on a digital twin model, solving the objective function to generate the optimal pitch sequence, as shown in the following equation:

[0050]

[0051] Where a(t) is the real-time vibration acceleration; Δθ(k) is the change in pitch angle at time k; ω1 is the vibration suppression weight coefficient; ω2 is the pitch action penalty weight; N is the prediction time domain;

[0052] Central Coordination Layer: Using the remaining fatigue life of each tower as a constraint, dynamically adjust the power distribution and yaw angle of the unit to make the wake centerline offset by ≥15° relative to the incoming wind direction in order to avoid high fatigue risk areas.

[0053] In step S3, the optimization function of the central coordination layer is defined as follows:

[0054]

[0055] Among them, P i D represents the real-time power generation of the i-th wind turbine generator; j Let λ be the fatigue damage increment of high-risk unit j; λ be the damage penalty coefficient; and H be the set of high-risk units.

[0056] The set of high-risk units H is shown in the following formula:

[0057]

[0058] k is the number of the wind turbine generator set; This represents the percentage of remaining life of unit k. This is the arithmetic average of the remaining lifespan of all units in the field.

[0059] In step S3, the wake path offset is achieved by calculating the upstream unit yaw compensation angle Δyaω, as shown in the following equation:

[0060]

[0061] Where, d safe L represents the safe offset distance calculated based on the location of the high-risk tower; L is the unit spacing.

[0062] In step S3, the stress change rate at key measuring points is calculated in real time as dσ / dt. When dσ / dt > α, the stress change rate is calculated according to r. lim =r0·(1-β·dσ / dt) Limiting power ramp rate, where α and β are material damage sensitivity coefficients.

[0063] In step S3, the central coordination layer implements a tower life balancing strategy: when the tower's remaining lifespan... At that time, its power setpoint P set The adjustment is as follows:

[0064]

[0065] Among them, P rated The rated power of the unit; L rem The percentage of the unit's remaining service life;

[0066] At the same time, load reduction will be transferred to The generator set.

[0067] Step S4: Issue optimization instructions to the execution mechanism and update the digital twin model through real-time feedback data.

[0068] A fatigue-damage-driven wind turbine tower control system includes a sensing layer, an edge computing layer, a central decision-making layer, and an execution layer. A real-time data bus with end-to-end delay connects each layer. Sensors include a hardware network of fiber optic sensor arrays, pulsed lidar arrays, and environmental monitoring units. The edge computing layer includes local controllers deployed on each tower section, configured with dynamically adaptable processors to execute vibration suppression algorithms. The central decision-making layer includes a server for a spatiotemporal database engine and an optimization solver. The execution layer includes a pitch system, an active damper, and a yaw actuator.

[0069] Therefore, the present invention adopts the above-mentioned wind power tower control method and system based on fatigue damage, which addresses the problems of high tower fatigue accumulation risk, insufficient transient resonance monitoring, insufficient avoidance of wake effects, and control response delay in traditional wind power tower control. Through multi-source sensing and dynamic optimization technology, it realizes the safe operation guarantee, fatigue damage suppression and power generation efficiency synergistic optimization of wind power tower, and balances the unit life and power generation performance.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention.

[0071] These modifications or equivalent substitutions cannot cause the modified technical solution to deviate from the technical specifications of this invention.

[0072] The spirit and scope of the technical plan.

Claims

1. A wind turbine tower control method based on fatigue damage, characterized in that, Includes the following steps: Step S1: Collect in real time the tower structure response data, three-dimensional wind field parameters and grid dispatch instructions of each wind power tower in the wind farm through a distributed sensor network; Step S2: Construct a digital twin model of the tower structure and dynamically predict the remaining fatigue life and damage evolution trend of each key part of the tower based on real-time structural response data; Step S3: Execute two-layer collaborative optimization control; Local control layer: Based on the vibration mode prediction results output by the digital twin model of the tower, an adaptive model predictive control algorithm is used to generate pitch and damper adjustment commands to suppress tower resonance; Central Coordination Layer: Using the remaining fatigue life of each tower as a constraint, dynamically adjust the power distribution and yaw angle of the unit to make the wake centerline offset by ≥15° relative to the incoming wind direction in order to avoid high fatigue risk areas. In step S3, the adaptive model predictive control algorithm includes predicting the tower top displacement response spectrum within the next 30 seconds based on a digital twin model, solving the objective function to generate the optimal pitch sequence, as shown in the following equation: ； in, Real-time vibration acceleration; For the first k The amount of change in the pitch angle at any given moment; This is the vibration suppression weighting coefficient; Penalty weights for pitching maneuvers; N For prediction in the time domain; In step S3, the optimization function of the central coordination layer is defined as follows: ； in, P i For the first i Real-time power generation of typhoon turbine generators; D j For the fatigue damage increment of high-risk unit j; This is the damage penalty coefficient; H A collection of high-risk units; High-risk unit collection H As shown in the following formula: ； Number the wind turbine generator set; This represents the percentage of remaining life of unit k. This is the arithmetic average of the remaining lifespan of all units in the field. In step S3, the wake path offset is calculated by solving the yaw compensation angle of the upstream unit. The implementation is shown in the following formula: ； in, L represents the safe offset distance calculated based on the location of the high-risk tower section; L is the unit spacing. In step S3, the central coordination layer implements a tower life balancing strategy: when the tower's remaining lifespan... At that time, its power setpoint The adjustment is as follows: ； in, This refers to the rated power of the unit. The percentage of the unit's remaining service life; At the same time, load reduction will be transferred to >1.2 The units; Step S4: Issue optimization instructions to the execution mechanism and update the digital twin model through real-time feedback data.

2. The wind turbine tower control method based on fatigue damage as described in claim 1, characterized in that, In step S1, the distributed sensor network includes a high-density fiber optic grating sensor array, a pulsed lidar, and temperature, humidity, vibration, and lightning strike monitoring units. The high-density fiber optic grating sensor array is installed on the inner wall of the tower, with one array arranged every 90 degrees along the circumference of the tower. The pulsed lidar is deployed at the leading edge and wake of the wind farm. The temperature, humidity, vibration, and lightning strike monitoring units are located at the top of the tower.

3. The wind turbine tower control method based on fatigue damage as described in claim 2, characterized in that, In step S1, the data sampling frequency of the fiber Bragg grating sensor array is dynamically adjusted.

4. The wind turbine tower control method based on fatigue damage as described in claim 1, characterized in that, In step S3, the stress change rate at key measuring points is calculated in real time. .

5. The wind turbine tower control method based on fatigue damage as described in claim 1, characterized in that, In step S2, the digital twin model includes a parameterized structural mechanics model based on finite element analysis, a fatigue life predictor trained by an LSTM-Transformer neural network, and a parameter self-calibration module that receives control verification data.

6. A wind power tower control system based on fatigue damage drive according to claim 1, characterized in that, It includes a sensing layer, an edge computing layer, a central decision-making layer, and an execution layer; a real-time data bus with end-to-end latency connects each layer; the sensors include a hardware network of fiber optic sensor arrays, pulsed lidar groups, and environmental monitoring units; the edge computing layer includes local controllers deployed on each tower, configured with dynamically adaptable processors to execute vibration suppression algorithms; the central decision-making layer includes a spatiotemporal database engine and an optimization solver server; the execution layer includes a pitch system, active dampers, and yaw actuators.

Images