A noise monitoring method and system, device, medium for a wind farm

Abstract

The present application relates to the technical field of wind power generation equipment, in particular to a noise monitoring method and system for a wind farm, equipment and medium, the method provided by the present application mainly includes judging whether the current noise data and the first judgment threshold are too large; a wind turbine fault diagnosis model is constructed, and based on the basic parameters, the wind turbine fault diagnosis model is used to output the result of whether the current wind turbine has a fault; if there is a fault, a control signal is sent to stop the operation of the current wind turbine, if there is no fault, the operating parameters of the current power generation wind farm are obtained, a power generation regulation and control evaluation model is constructed, and a control signal is output to determine whether adjustment is needed. Through the above method, the factors related to the fault of the wind turbine itself are considered to analyze the current noise, determine whether to execute shutdown or execute the operation of reducing power generation power, so as to effectively reduce noise and eliminate the fault problem of the wind turbine itself.

Description

Technical Field

[0001] This invention relates to the field of wind power generation equipment technology, and more specifically, to a noise monitoring method, system, equipment, and medium for wind farms. Background Technology

[0002] Currently, the equipment employs a metrologically calibrated integrating sound level meter and an outdoor automatic monitoring instrument, coupled with meteorological sensors, to simultaneously collect wind speed, wind direction, temperature, and humidity data for noise propagation correction. Monitoring content covers equivalent continuous A-weighted sound level, noise peak value, and spectral distribution, distinguishing between fan operating noise and background environmental noise, and completing data correction and source strength calculation. Automated monitoring systems are now widely used, enabling real-time data acquisition, remote transmission, automatic storage, and batch analysis, accommodating both individual unit noise source measurement and overall site environmental noise assessment.

[0003] However, current noise detection methods overlook the problem of faults in the fan itself. The source of noise may be the normal friction sound of the blades against the air, or it may be other abnormal noises caused by the fault. Therefore, incorrect control measures may be implemented after noise is generated. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, device, and medium for noise monitoring in wind farms to solve the aforementioned problems in the prior art.

[0005] This invention is achieved through the following technical solution: In a first aspect, the present invention provides a noise monitoring method for wind farms, comprising: Obtain the current noise data and set a first judgment threshold. Determine whether the noise is too loud based on the current noise data and the first judgment threshold. If not, no action is taken; if yes, several basic parameters of the wind turbine within the collection period are obtained, a wind turbine fault diagnosis model is constructed, and the result of whether the wind turbine is faulty is output based on the basic parameters through the wind turbine fault diagnosis model. If a fault exists, a control signal to stop the current wind turbine is issued. If no fault exists, the current operating parameters of the wind farm are obtained, a power generation regulation and evaluation model is constructed, and an evaluation value is output based on the operating parameters through the power generation regulation and evaluation model. A second judgment threshold is set, and a control signal is output to determine whether adjustment is needed based on the evaluation value and the second judgment threshold.

[0006] Preferably, it also includes setting a decibel meter, including: Divide the current wind turbine into several segments along the longitudinal direction, set up a decibel meter in each segment, and obtain the decibel value collected by the current decibel meter; The target segment containing the decibel meter with the highest decibel value is retained. This target segment is then divided into several segments in the same proportion. A decibel meter is set up in each segment, and the noise data collected by the current decibel meter is obtained. The target segment containing the decibel meter with the highest decibel value is retained as the final target segment, and several decibel meters are set within the final target segment. The noise data is obtained by combining the decibel values ​​collected by several decibel meters.

[0007] Preferably, the noise data obtained by combining the decibel values ​​collected by several decibel meters includes: Set up a comprehensive calculation model, including:

[0008] In the formula, For noisy data, Let i be the weight for the i-th decibel meter. Let i be the decibel values ​​of i decibel meters.

[0009] Preferably, the construction of the wind turbine fault diagnosis model includes: The basic parameters are preprocessed, and features are extracted from the preprocessed basic parameters. Redundant and invalid features are removed and the dimensionality is reduced from the extracted features to obtain the target feature sample set. Construct a CNN convolutional neural network model, which includes convolutional layers, pooling layers, fully connected layers, and a classification output layer; Feature extraction is performed through convolutional layers, max pooling is performed through pooling layers, and the classification output layer outputs the probability of various faults.

[0010] Preferably, the preprocessing of the basic parameters includes: Missing value imputation:

[0011] In the formula, The data after interpolation and filling. and Let i and j be the valid data on either side of the missing position, respectively, and k be the index value of the missing position. Wavelet thresholding for noise reduction:

[0012] In the formula, The value is the result of wavelet decomposition, and h is the translation factor. The wavelet decomposition level is denoted as . For scaling function, These are low-frequency approximation coefficients. For wavelet basis functions, For high-frequency detail coefficients.

[0013] Preferably, the convolutional layer comprises:

[0014] In the formula, For the extracted features, For activation function, For convolution kernel weights, For input parameters, For bias, This represents the number of network layers.

[0015] Preferably, the construction of the power generation regulation and evaluation model includes:

[0016] In the formula, As an evaluation value, Rated power generation capacity This represents the current power generation capacity. The first threshold for safety judgment, The current wind speed, This represents the average wind speed throughout the year.

[0017] Secondly, the present invention also provides a noise monitoring system for wind farms, comprising the aforementioned noise monitoring method for wind farms, and further comprising: The data processing module is configured to acquire the current noise data and set a first judgment threshold. Based on the current noise data and the first judgment threshold, it determines whether the noise is too loud. If not, no processing is performed. If so, it acquires several basic parameters of the wind turbine within the acquisition period, constructs a wind turbine fault diagnosis model, and outputs the result of whether the wind turbine has a fault based on the basic parameters and the wind turbine fault diagnosis model. The judgment module is configured to issue a control signal to stop the current wind turbine if a fault exists, and to obtain the current operating parameters of the wind farm if no fault exists, construct a power generation regulation and evaluation model, output an evaluation value based on the operating parameters through the power generation regulation and evaluation model, set a second judgment threshold, and output a control signal to determine whether adjustment is needed based on the evaluation value and the second judgment threshold.

[0018] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for noise monitoring in a wind farm.

[0019] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for noise monitoring in a wind farm.

[0020] The technical solution of the present invention has at least the following advantages and beneficial effects: The method provided by this invention mainly includes: judging the current noise data and determining whether the noise is too high based on a first judgment threshold; constructing a wind turbine fault diagnosis model, and outputting the result of whether the current wind turbine has a fault based on the basic parameters; if a fault exists, issuing a control signal to stop the current wind turbine operation; if no fault exists, acquiring the current operating parameters of the power generation wind farm, constructing a power generation regulation and evaluation model, and outputting a control signal indicating whether adjustment is needed. By combining the above method with factors related to wind turbine malfunctions, the current noise is analyzed to determine whether to perform a shutdown or reduce power generation, thus ensuring effective noise reduction while eliminating wind turbine malfunctions. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. The independently described modules or sub-modules may or may not be physically separated; they may be implemented in software or hardware, and some modules or sub-modules may be implemented in software, with the processor calling the software to implement the function of these modules or sub-modules, while other modules or sub-modules may be implemented in hardware, such as through hardware circuits. Furthermore, some or all of the modules can be selected to achieve the purpose of this application's solution according to actual needs.

[0024] Please refer to Figures 1-2 A noise monitoring method for wind farms, comprising: S1: Obtain the current noise data and set a first judgment threshold. Based on the current noise data and the first judgment threshold, determine whether the noise is too loud. A preliminary judgment was made based on the current noise data to determine whether the current noise level has exceeded the set safety value, i.e., the first judgment threshold. If it has exceeded the threshold, the noise level is considered to be too high.

[0025] S2: If not, no action is taken; if yes, several basic parameters of the wind turbine within the collection period are obtained, a wind turbine fault diagnosis model is constructed, and the result of whether the wind turbine is faulty is output based on the basic parameters through the wind turbine fault diagnosis model. By using the above method, when excessive noise occurs, the condition of the fan itself is diagnosed, and it is considered whether the current noise is caused by a fan malfunction, so as to avoid always assuming that the noise is caused by the blades and missing the opportunity to remedy the situation.

[0026] S3: If a fault exists, a control signal to stop the current wind turbine operation is issued. If no fault exists, the current operating parameters of the power generation wind farm are obtained, a power generation regulation and evaluation model is constructed, an evaluation value is output based on the operating parameters through the power generation regulation and evaluation model, and a second judgment threshold is set. Based on the evaluation value and the second judgment threshold, a control signal is output indicating whether adjustment is needed.

[0027] Since reducing the blade rotation speed will affect the entire wind farm, the power generation regulation and evaluation model is used to comprehensively consider the current operating parameters of the wind farm and the specific circumstances to determine whether regulation is necessary.

[0028] The method provided by this invention mainly includes: judging the current noise data and determining whether the noise is too high based on a first judgment threshold; constructing a wind turbine fault diagnosis model, and outputting the result of whether the current wind turbine has a fault based on the basic parameters; if a fault exists, issuing a control signal to stop the current wind turbine operation; if no fault exists, acquiring the current operating parameters of the power generation wind farm, constructing a power generation regulation and evaluation model, and outputting a control signal indicating whether adjustment is needed. By combining the above method with factors related to wind turbine malfunctions, the current noise is analyzed to determine whether to perform a shutdown or reduce power generation, thus ensuring effective noise reduction while eliminating wind turbine malfunctions.

[0029] In one exemplary embodiment of the present invention, the present invention also provides a method for setting a decibel meter, comprising: The current wind turbine is divided into several segments along its longitudinal direction. A decibel meter is set in each segment, and the decibel value collected by the current decibel meter is obtained. The target segment containing the decibel meter with the highest decibel value is retained. This target segment is also divided into several segments according to the same proportion. A decibel meter is set in each segment again, and the noise data collected by the current decibel meter is obtained. The target segment containing the decibel meter with the highest decibel value is retained as the final target segment. Several decibel meters are set in the final target segment. The noise data is obtained by combining the decibel values ​​collected by several decibel meters.

[0030] This solution employs a layered search deployment approach, involving progressive segmentation, iterative optimization of high-noise sections, and final fine-grained point placement. The overall monitoring logic advances layer by layer, offering significant advantages over traditional uniform, full-area deployment methods. Firstly, by monitoring the wind turbine longitudinally in segments, it can quickly and preliminarily locate core areas with high noise radiation intensity, broadly filtering high-noise sections and effectively narrowing the key monitoring scope. This avoids resource dispersion caused by indiscriminate, full-area deployment of monitoring equipment, thus improving overall monitoring efficiency.

[0031] After initial screening, the area where the noise maximum value is located is subdivided and retested in the same proportion. This enables precise locking of the noise peak area. By narrowing the range layer by layer, low-noise interference sections can be continuously eliminated, gradually approaching the location of the strongest noise source. The positioning accuracy is improved step by step, effectively avoiding the positional deviation caused by large-scale deployment, and ensuring that the selection of key monitoring areas is scientific and accurate.

[0032] After the final target segment is determined, multiple decibel meters are densely deployed within the core area to fully collect subtle differences in noise levels within the region, comprehensively capturing the noise distribution characteristics, fluctuation patterns, and sound level changes. Multi-point integrated calculation and analysis can reduce data deviations caused by accidental errors from individual devices and transient environmental interference, improving the stability and representativeness of the overall monitoring data.

[0033] Specifically, a comprehensive calculation model is set up, including:

[0034] In the formula, For noisy data, Let i be the weight for the i-th decibel meter. Let i be the decibel values ​​of i decibel meters.

[0035] The weighting settings can be configured based on the distance between each decibel meter and the wind turbine blades. The closer the distance, the higher the weight. The weights of all the calculated weights can be summed to 1.

[0036] In one exemplary embodiment of the present invention, the construction of the wind turbine fault diagnosis model includes: The basic parameters are preprocessed, and features are extracted from the preprocessed basic parameters. Redundant and invalid features are removed and the dimensionality is reduced from the extracted features to obtain the target feature sample set. Construct a CNN convolutional neural network model, which includes convolutional layers, pooling layers, fully connected layers, and a classification output layer; Feature extraction is performed through convolutional layers, max pooling is performed through pooling layers, and the classification output layer outputs the probability of various faults.

[0037] Preprocessing the basic parameters includes: Missing value imputation:

[0038] In the formula, The data after interpolation and filling. and Let i and j be the valid data on either side of the missing position, respectively, and k be the index value of the missing position. Wavelet thresholding for noise reduction:

[0039] In the formula, The value is the result of wavelet decomposition, and h is the translation factor. The wavelet decomposition level is denoted as . For scaling function, These are low-frequency approximation coefficients. For wavelet basis functions, For high-frequency detail coefficients.

[0040] Convolutional layers include:

[0041] In the formula, For the extracted features, For activation function, For convolution kernel weights, For input parameters, For bias, This represents the number of network layers.

[0042] An exemplary embodiment of the present invention includes constructing a power generation regulation and evaluation model comprising:

[0043] In the formula, As an evaluation value, Rated power generation capacity This represents the current power generation capacity. The first threshold for safety judgment, The current wind speed, This represents the average wind speed throughout the year.

[0044] The higher the evaluation value, the more necessary it is to perform noise reduction operations, such as reducing power generation or blade rotation speed. When the evaluation value exceeds the set second judgment threshold, a control signal that needs to be adjusted is output. Any operation aimed at reducing noise is acceptable. The setting of the second judgment and threshold can be based on historical data.

[0045] In this embodiment, power generation is first introduced as a calculation parameter to balance the operating efficiency of the wind turbine and the requirements for noise control. The power generation of the wind turbine directly reflects the unit's operating load and working status. The sound source radiation characteristics differ under different output conditions. Combining power calculation can avoid excessive restriction of unit power generation due to simple noise reduction, balance the economic benefits of power generation and the goals of sound environment management, and achieve a synergistic balance between noise reduction optimization and power output. When the current power generation is higher and closer to the rated power, it indicates that the power consumption of the grid is higher at this time, making it inconvenient to perform noise reduction operations. Therefore, negative feedback is set.

[0046] Meanwhile, wind speed, in conjunction with other parameters, not only affects the intensity of turbine blade rotational noise and aerodynamic noise, but also alters the noise's airborne attenuation and diffusion patterns, making it a significant factor influencing noise fluctuations. Introducing wind speed into the correction calculations eliminates data interference from natural wind conditions, distinguishes between ambient wind speed and the unit's own noise source intensity, and improves the accuracy of the evaluation value calculation. Higher current wind speeds are more suitable for effective wind power generation, but also less conducive to noise reduction operations; therefore, in this embodiment, negative feedback is used.

[0047] The evaluation value obtained by comprehensively coupling these three factors can fully match the actual operating conditions of the wind turbine, resulting in a more objective evaluation and avoiding control deviations caused by one-sided judgments based on a single indicator. Based on this, noise control and regulation can be carried out with more targeted control strategies, accurately suppressing excessive noise while reasonably ensuring the unit's power generation efficiency. This adapts to the complex and ever-changing operating environment of field wind farms, providing reliable data support for the refined noise reduction and control of wind turbines.

[0048] A noise monitoring system for wind farms, comprising the aforementioned noise monitoring method for wind farms, further comprising: The data processing module is configured to acquire the current noise data and set a first judgment threshold. Based on the current noise data and the first judgment threshold, it determines whether the noise is too loud. If not, no processing is performed. If so, it acquires several basic parameters of the wind turbine within the acquisition period, constructs a wind turbine fault diagnosis model, and outputs the result of whether the wind turbine has a fault based on the basic parameters and the wind turbine fault diagnosis model. The judgment module is configured to issue a control signal to stop the current wind turbine if a fault exists, and to obtain the current operating parameters of the wind farm if no fault exists, construct a power generation regulation and evaluation model, output an evaluation value based on the operating parameters through the power generation regulation and evaluation model, set a second judgment threshold, and output a control signal to determine whether adjustment is needed based on the evaluation value and the second judgment threshold.

[0049] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0050] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer software product, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A noise monitoring method for wind farms, characterized in that, include: Obtain the current noise data and set a first judgment threshold. Determine whether the noise is too loud based on the current noise data and the first judgment threshold. If not, no action is taken; if yes, several basic parameters of the wind turbine within the collection period are obtained, a wind turbine fault diagnosis model is constructed, and the result of whether the wind turbine is faulty is output based on the basic parameters through the wind turbine fault diagnosis model. If a fault exists, a control signal to stop the current wind turbine is issued. If no fault exists, the current operating parameters of the wind farm are obtained, a power generation regulation and evaluation model is constructed, and an evaluation value is output based on the operating parameters through the power generation regulation and evaluation model. A second judgment threshold is set, and a control signal is output to determine whether adjustment is needed based on the evaluation value and the second judgment threshold.

2. The noise monitoring method for wind farms according to claim 1, characterized in that, It also includes setting up a decibel meter, including: Divide the current wind turbine into several segments along the longitudinal direction, set up a decibel meter in each segment, and obtain the decibel value collected by the current decibel meter; The target segment containing the decibel meter with the highest decibel value is retained. This target segment is then divided into several segments in the same proportion. A decibel meter is set up in each segment, and the noise data collected by the current decibel meter is obtained. The target segment containing the decibel meter with the highest decibel value is retained as the final target segment, and several decibel meters are set within the final target segment. The noise data is obtained by combining the decibel values ​​collected by several decibel meters.

3. The noise monitoring method for wind farms according to claim 2, characterized in that, The noise data obtained by combining the decibel values ​​collected by several decibel meters includes: Set up a comprehensive calculation model, including: In the formula, For noisy data, Let i be the weight for the i-th decibel meter. Let i be the decibel values ​​of i decibel meters.

4. The noise monitoring method for wind farms according to claim 3, characterized in that, The construction of the wind turbine fault diagnosis model includes: The basic parameters are preprocessed, and features are extracted from the preprocessed basic parameters. Redundant and invalid features are removed and the dimensionality is reduced from the extracted features to obtain the target feature sample set. Construct a CNN convolutional neural network model, which includes convolutional layers, pooling layers, fully connected layers, and a classification output layer; Feature extraction is performed through convolutional layers, max pooling is performed through pooling layers, and the classification output layer outputs the probability of various faults.

5. A noise monitoring method for wind farms according to claim 4, characterized in that, The preprocessing of the basic parameters includes: Missing value imputation: In the formula, The data after interpolation and filling. and Let i and j be the valid data on either side of the missing position, respectively, and k be the index value of the missing position. Wavelet thresholding for noise reduction: In the formula, The value is the result of wavelet decomposition, and h is the translation factor. The wavelet decomposition level is denoted as . For scaling function, These are low-frequency approximation coefficients. For wavelet basis functions, For high-frequency detail coefficients.

6. A noise monitoring method for wind farms according to claim 5, characterized in that, The convolutional layer includes: In the formula, For the extracted features, For activation function, For convolution kernel weights, For input parameters, For bias, This represents the number of network layers.

7. A noise monitoring method for wind farms according to claim 6, characterized in that, The constructed power generation regulation and evaluation model includes: In the formula, As an evaluation value, Rated power generation capacity This represents the current power generation capacity. The first threshold for safety judgment, The current wind speed, This represents the average wind speed throughout the year.

8. A noise monitoring system for wind farms, characterized in that, The method for noise monitoring in a wind farm, comprising any one of claims 1-7, further comprising: The data processing module is configured to acquire the current noise data and set a first judgment threshold. Based on the current noise data and the first judgment threshold, it determines whether the noise is too loud. If not, no processing is performed. If so, it acquires several basic parameters of the wind turbine within the acquisition period, constructs a wind turbine fault diagnosis model, and outputs the result of whether the wind turbine has a fault based on the basic parameters and the wind turbine fault diagnosis model. The judgment module is configured to issue a control signal to stop the current wind turbine if a fault exists, and to obtain the current operating parameters of the wind farm if no fault exists, construct a power generation regulation and evaluation model, output an evaluation value based on the operating parameters through the power generation regulation and evaluation model, set a second judgment threshold, and output a control signal to determine whether adjustment is needed based on the evaluation value and the second judgment threshold.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the noise monitoring method for wind farms as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements a noise monitoring method for a wind farm as described in any one of claims 1-7.

Images