A load monitoring device and prediction method for a tandem dual-rotor wind turbine generator set

Abstract

The present invention belongs to the field of wind turbine monitoring, and specifically designs a load monitoring device and prediction method for a serial twin-rotor wind turbine, comprising setting sensors at multiple key positions of the serial twin-rotor wind turbine, collecting data from the sensors, and transferring the data to an external server. Within the effective time range of the sensors, accurate monitoring of the loads at key positions of the serial twin-rotor wind turbine can be achieved, and a load prediction model is established and trained based on the data obtained from the detection; when the sensor fails, the loads at key positions of the wind turbine are predicted using the trained load prediction system.

Description

Technical Field

[0001] The present invention belongs to the field of wind turbine monitoring, and specifically provides a load monitoring device and a prediction method for a tandem double-rotor wind turbine. Background Art

[0002] Currently, wind turbines are developing towards higher power, longer blades, and taller towers. Without breakthroughs in material systems, these core technologies will be subject to numerous limitations. Tandem twin-rotor wind turbines enable cascaded utilization of wind energy, breaking the traditional single-rotor rule where blade length increases linearly with turbine capacity. This reduces turbine costs and improves site utilization, and has garnered significant attention in the wind power industry in recent years.

[0003] Compared with single-rotor wind turbines, the transmission system of tandem twin-rotor wind turbines is longer and more complex, and accurate load monitoring and prediction is more difficult than with traditional wind turbines. Summary of the Invention

[0004] The present invention provides a load monitoring device and prediction method for a tandem twin-rotor wind turbine generator set, so as to solve the problem of difficulty in load monitoring and prediction of a tandem twin-rotor wind turbine generator set in the prior art.

[0005] To achieve the above object, the present invention proposes the following technical solutions:

[0006] A tandem dual-rotor wind turbine load monitoring device includes multiple optical fiber load sensors, multiple optical fiber temperature sensors, a front rotor optical fiber sensor analyzer, a nacelle-type wind measurement laser radar, a nacelle switch, a rear rotor optical fiber sensor analyzer, a tower optical fiber sensor analyzer, and a tower base switch.

[0007] The optical fiber load sensor and the optical fiber temperature sensor are connected to the front wind wheel optical fiber sensing analyzer or the rear wind wheel optical fiber sensing analyzer or the tower optical fiber sensing analyzer;

[0008] The front wind wheel optical fiber sensor analyzer, the rear wind wheel optical fiber sensor analyzer and the tower optical fiber sensor analyzer are connected to the cabin switch arranged on the tower;

[0009] The nacelle switch and the nacelle-type wind laser radar are connected to the tower bottom switch;

[0010] The switch at the bottom of the tower is connected to the external server.

[0011] Preferably, the front wind wheel optical fiber sensor analyzer, the rear wind wheel optical fiber sensor analyzer and the tower optical fiber sensor analyzer are all provided with wireless modules.

[0012] Preferably, four optical fiber load sensors are installed at the root section and maximum chord section of each front wind rotor blade, and four optical fiber load sensors are installed at the root section and maximum chord section of each rear wind rotor blade.

[0013] Preferably, four optical fiber load sensors are provided at the upper, middle and lower ends of the tower respectively.

[0014] Preferably, an optical fiber temperature sensor is provided next to each optical fiber load sensor.

[0015] Preferably, the front wind wheel fiber optic sensor analyzer, the rear wind wheel fiber optic sensor analyzer and the tower fiber optic sensor analyzer are respectively arranged on the front wind wheel hub, the rear wind wheel hub and the tower; the tower bottom switch is arranged at the bottom of the tower; and the cabin switch is arranged on the tower.

[0016] A load prediction method for a tandem twin-rotor wind turbine generator system, comprising the following steps:

[0017] Step 1: Receive data from the tower bottom receiver, including unit operation data and load data;

[0018] Step 2: Set the kernel function of the convolutional neuron to obtain multiple convolutional neural network models. Group the unit operation data and load data at 1 m / s wind speed intervals. Each group of data is fed into a convolutional neural network model for training to obtain a load prediction model.

[0019] The load prediction model includes multiple convolutional neural network models;

[0020] Step 3: Determine the corresponding convolutional neural network based on the operating wind speed of the unit to be predicted, input the operating data of the unit to be predicted into the corresponding convolutional neural network model in the load prediction model to obtain the predicted load.

[0021] Preferably, the convolution neurons include one low-frequency convolution neuron and multiple high-frequency convolution neurons.

[0022] Preferably, the low-frequency convolution neuron adopts Gaussian distribution as the kernel function, which is expressed as:

[0023]

[0024] Where μ is the expected value, which is usually 0; σ is the standard deviation, which is usually 1.

[0025] Preferably, the high-frequency convolution neuron uses a sine wavelet function and a cosine wavelet function as kernel functions, wherein the expression of the sine wavelet function is:

[0026]

[0027]

[0028] The expression of the cosine wavelet function is:

[0029]

[0030]

[0031] Where f(i) is the frequency series, which is divided into several characteristic frequencies from low frequency to high frequency in multiples according to the accuracy requirements.

[0032] The present invention is beneficial in that:

[0033] A load monitoring device for a tandem twin-rotor wind turbine set is proposed, which uses sensors to accurately monitor the loads of various parts of the twin-rotor set.

[0034] A load prediction method for tandem twin-rotor wind turbines is proposed. A load learning system is established to train the model. After sensor failure, the trained load prediction system can be used to predict the loads on key parts of the wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings, which constitute part of the present invention, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0036] Figure 1 This is a schematic diagram of a load monitoring device for a tandem dual-rotor wind turbine;

[0037] Figure 2 This is a schematic diagram of a load prediction method for a tandem twin-rotor wind turbine;

[0038] Figure 3 Schematic diagram of the convolutional neural network model.

[0039] In the figure, 1 is the front wind rotor blade; 2 is the optical fiber load sensor; 3 is the optical fiber temperature sensor; 4 is the front wind rotor optical fiber sensor analyzer; 5 is the front wind rotor hub; 6 is the nacelle-type wind measurement lidar; 7 is the nacelle; 8 is the nacelle switch; 9 is the rear wind rotor blade; 10 is the rear wind rotor optical fiber sensor analyzer; 11 is the rear wind rotor hub; 12 is the tower; 13 is the tower optical fiber sensor analyzer; 14 is the tower bottom switch. DETAILED DESCRIPTION

[0040] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other.

[0041] The following detailed description is an exemplary description and is intended to provide further detailed description of the present invention. Unless otherwise indicated, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the present invention belongs. The terms used in the present invention are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention.

[0042] Example 1:

[0043] See also Figure 1 As shown, the present invention provides a tandem twin-rotor wind turbine load monitoring device, comprising an external server, a fiber optic load sensor 2, a fiber optic temperature sensor 3, a front rotor fiber optic sensor analyzer 4, a nacelle-mounted wind measurement lidar 6, a nacelle switch 8, a rear rotor fiber optic sensor analyzer 10, a tower fiber optic sensor analyzer 13, and a tower-base switch 14. The front rotor fiber optic sensor analyzer 4, the rear rotor fiber optic sensor analyzer 10, and the tower fiber optic sensor analyzer 13 are all equipped with wireless modules.

[0044] The tandem twin-rotor wind turbine includes a front rotor hub 5, a nacelle 7, rear rotor blades 9, a rear rotor hub 11 and a tower 12; wherein, the top of the tower 12 is connected to the middle section of the transversely arranged nacelle 7, and the two ends of the nacelle 7 are respectively connected to the front rotor hub 5 and the rear rotor hub 11, the front rotor hub 5 is provided with three front rotor blades 1, and the rear rotor hub 11 is provided with three rear rotor blades 9.

[0045] Four optical fiber load sensors 2 are installed on the root section and maximum chord section of each front wind rotor blade 1 by glue, and four optical fiber load sensors 2 are installed on the root section and maximum chord section of each rear wind rotor blade 9 by glue.

[0046] Four are installed at the upper, middle and lower ends of the tower 12; a fiber optic temperature sensor 3 is fixed next to each fiber optic load sensor 2 by glue. At this time, 24 fiber optic load sensors 2 and 24 fiber optic temperature sensors 3 are provided on the front wind wheel, 24 fiber optic load sensors 2 and 24 fiber optic temperature sensors 3 are provided on the rear wind wheel, and 12 fiber optic load sensors 2 and 12 fiber optic temperature sensors 3 are provided on the tower 12.

[0047] The 24 optical fiber load sensors 2 and 24 optical fiber temperature sensors 3 on the front wind wheel are connected to the front wind wheel optical fiber sensing analyzer 4, and the 24 optical fiber load sensors 2 and 24 optical fiber temperature sensors 3 on the rear wind wheel are connected to the rear wind wheel optical fiber sensing analyzer 10.

[0048] The front wind wheel optical fiber sensor analyzer 4 and the rear wind wheel optical fiber sensor analyzer 10 are respectively arranged on the front wind wheel hub 5 and the rear wind wheel hub 11.

[0049] The front wind wheel optical fiber sensor analyzer 4 and the rear wind wheel optical fiber sensor analyzer 10 transmit the data from the optical fiber load sensor 2 and the optical fiber temperature sensor 3 to the cabin switch 8 through the wireless module.

[0050] The 12 optical fiber load sensors 2 and 12 optical fiber temperature sensors 3 on the tower 12 are connected to a tower optical fiber sensor analyzer 13 , and the tower optical fiber sensor analyzer 13 transmits data to a tower bottom switch 14 provided at the bottom of the tower 12 via optical cables.

[0051] The cabin switch 8 is set on the tower 12. The cabin switch 8 and the tower bottom switch 14 are connected through optical fiber. The data of the cabin switch 8 and the data of the tower bottom switch 14 are sent to the external server after the tower bottom switch 14 completes the convergence.

[0052] A nacelle-type wind laser radar 6 is provided on the nacelle 7 for measuring wind speed and sending the data to an external server. The data may or may not pass through the tower bottom switch 14.

[0053] Example 2:

[0054] See also Figure 2 As shown, the present invention provides a load prediction method for a tandem twin-rotor wind turbine, which receives turbine operation data and load data. In this embodiment, an external server is used for receiving data, and the received data comes from a nacelle-type wind laser radar 6 and a tower-bottom switch 14.

[0055] Unit operating data: wind speed, yaw error, front wind rotor pitch angle, front wind rotor generator torque, front wind rotor generator speed, front wind rotor azimuth, rear wind rotor pitch angle, rear wind rotor generator torque, rear wind rotor generator speed, rear wind rotor azimuth, nacelle front and rear acceleration, nacelle left and right acceleration.

[0056] The load data includes: the swing and flapping bending moments of the blade root section and the maximum chord section, the in-plane bending moment, out-of-plane bending moment and aerodynamic thrust of the hub center plane, and the pitching bending moment and overturning bending moment of the upper, middle and lower sections of the tower.

[0057] Use the model training module to process the data. The processing process is as follows Figure 2 As shown, specifically:

[0058] A convolutional neuron kernel function is set to obtain multiple convolutional neural network models. The data is grouped at a wind speed interval of 1 m / s. Each group of data is brought into a convolutional neural network model for training to obtain a load prediction model, which includes multiple convolutional neural network models.

[0059] In the load prediction module, the corresponding convolutional neural network is determined according to the unit operating wind speed, and the unit operating data is input into the corresponding convolutional neural network model to obtain the load of key parts at the corresponding moment.

[0060] Convolutional neural network model participation Figure 3 , which consists of three layers: input layer, hidden layer and output layer; the input layer corresponds to the wind turbine data, the hidden layer corresponds to the convolutional neurons, including: 1 low-frequency convolutional neuron and multiple high-frequency convolutional neurons, and the output layer corresponds to the load of the key parts of the wind turbine, that is, the load of the key parts measured by the dual wind wheel load monitoring system.

[0061] The convolutional neuron algorithm can be expressed as:

[0062]

[0063] Where h is the feature map; g is the kernel function; and f is the input function.

[0064] The low-frequency convolution neuron uses Gaussian distribution as the kernel function, which is expressed as:

[0065]

[0066] Where μ is the expected value, usually 0; σ is the standard deviation, usually 1. The low-frequency convolution neuron is equivalent to a low-frequency filter, which performs low-frequency filtering on the input signal.

[0067] High-frequency convolution neurons use sine wavelet function and cosine wavelet function as kernel functions to extract the features of the input signal.

[0068] Among them, the expression of the cosine wavelet function is:

[0069]

[0070]

[0071] The expression of the sine wavelet function is:

[0072]

[0073]

[0074] Where f(i) is the frequency series, which is divided into several characteristic frequencies from low frequency to high frequency in multiples according to the accuracy requirements.

[0075] By setting up a model training module according to the above method and placing the model training module in an external server, the load monitoring and prediction of the tandem dual-rotor wind turbine can be realized. When the sensor fails, the trained load prediction model can predict the load of the key parts of the wind turbine.

[0076] It is understood from common technical knowledge that the present invention may be implemented by other embodiments that do not depart from its spirit or essential features. Therefore, the embodiments disclosed above are, in all respects, merely illustrative and not exclusive. All modifications within the scope of the present invention or equivalent to the scope of the present invention are intended to be encompassed by the present invention.

[0077] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0078] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0079] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0080] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A load prediction method for a tandem twin-rotor wind turbine generator set, based on a load monitoring device for a tandem twin-rotor wind turbine generator set, the device comprising a plurality of optical fiber load sensors (2), a plurality of optical fiber temperature sensors (3), a front wind rotor optical fiber sensor analyzer (4), a nacelle-type wind measurement laser radar (6), a nacelle switch (8), a rear wind rotor optical fiber sensor analyzer (10), a tower optical fiber sensor analyzer (13) and a tower bottom switch (14); the optical fiber load sensors (2) and the optical fiber temperature sensors (3) are connected to the front wind rotor optical fiber sensor analyzer (4) or the rear wind rotor optical fiber sensor analyzer (10) or the tower fiber sensor analyzer (13); the front wind rotor optical fiber sensor analyzer (4), the rear wind rotor optical fiber sensor analyzer (10) and the tower optical fiber sensor analyzer (13) are connected to the nacelle switch (8) arranged on the tower (12); the nacelle switch (8) and the nacelle-type wind measurement laser radar (6) are connected to the tower bottom switch (14); the tower bottom switch (14) is connected to an external server; and the method is characterized in that: The specific steps of the method include: Step 1: receiving data from the tower bottom switch (14), including unit operation data and load data; Step 2: Set the kernel function of the convolutional neuron to obtain multiple convolutional neural network models. Group the unit operation data and load data at 1 m / s wind speed intervals. Each group of data is fed into a convolutional neural network model for training to obtain a load prediction model. The load prediction model includes multiple convolutional neural network models; The convolution neurons include one low-frequency convolution neuron and multiple high-frequency convolution neurons; The low-frequency convolution neuron uses Gaussian distribution as the kernel function, which is expressed as: Where μ is the expected value, which is usually 0; σ is the standard deviation, which is usually 1; The high-frequency convolution neuron uses the sine wavelet function and the cosine wavelet function as kernel functions, wherein the expression of the sine wavelet function is: The expression of the cosine wavelet function is: Where f(i) is the frequency series, which is divided into several characteristic frequencies from low frequency to high frequency according to the accuracy requirements; Step 3: Determine the corresponding convolutional neural network based on the operating wind speed of the unit to be predicted, input the operating data of the unit to be predicted into the corresponding convolutional neural network model in the load prediction model to obtain the predicted load.

2. The load prediction method for a tandem twin-rotor wind turbine generator system according to claim 1, wherein: The front wind wheel optical fiber sensor analyzer (4), the rear wind wheel optical fiber sensor analyzer (10) and the tower optical fiber sensor analyzer (13) are all provided with wireless modules.

3. The load prediction method for a tandem twin-rotor wind turbine generator system according to claim 1, wherein: Four optical fiber load sensors (2) are installed at the blade root section and the maximum chord length section of each front wind rotor blade (1), and four optical fiber load sensors (2) are installed at the blade root section and the maximum chord length section of each rear wind rotor blade (9).

4. The load prediction method for a tandem twin-rotor wind turbine generator system according to claim 3, wherein: Four optical fiber load sensors (2) are respectively provided at the upper, middle and lower ends of the tower (12).

5. The load prediction method for a tandem twin-rotor wind turbine generator system according to claim 4, characterized in that: An optical fiber temperature sensor (3) is provided next to each optical fiber load sensor (2).

6. The method for predicting load of a tandem twin-rotor wind turbine according to claim 4, wherein: The front wind wheel optical fiber sensor analyzer (4), the rear wind wheel optical fiber sensor analyzer (10), and the tower optical fiber sensor analyzer (13) are respectively arranged on the front wind wheel hub (5), the rear wind wheel hub (11), and the tower (12); the tower bottom switch (14) is arranged at the bottom of the tower (12); and the cabin switch (8) is arranged on the tower (12).

Images