Wind power plant noise control and power optimization method and device

Abstract

The invention relates to the technical field of wind power plant optimization, and discloses a wind power plant noise control and power optimization method and device, and the method comprises the steps: obtaining the multi-source noise data of a wind power plant, constructing a multi-source noise feature database, and constructing a noise propagation theoretical model; determining a plurality of monitoring points in the wind power plant based on the noise propagation theoretical model, and collecting actual multi-source noise data and fan operation data of the plurality of monitoring points in real time; on the basis of actual multi-source noise data and fan operation data, a multi-objective optimization function fusing constraint conditions is constructed by taking minimization of dispatching instruction difference, fan operation load and noise superscale as objectives; and solving the multi-objective optimization function to generate an optimal power distribution scheme of the fan under the noise constraint. According to the invention, intelligent monitoring and intelligent control can be carried out on the noise of the wind power plant near the residential area, so that the noise influence of the wind power plant on the nearby residential area is reduced, and the contradiction between the residential area and the wind power plant is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind farm optimization, in particular to a wind farm noise control and power optimization method and device. BACKGROUND

[0002] The noise of wind turbine mainly includes mechanical noise and aerodynamic noise. The mechanical noise mainly comes from the friction of the generator and the gear box hydraulic machine and other instruments. By improving the machining process and installation accuracy, and improving the lubrication method between parts, the mechanical noise can be reduced. The aerodynamic noise is the noise generated by the interaction of airflow and blades, which is the main source of wind turbine noise. At present, the noise control of wind farm mainly includes construction stage and post-construction modification, and the noise control method is also divided into two types: one is to develop low-noise high-aerodynamic performance airfoils in the design process; the other is to control noise by external force, such as isolating noise, absorbing noise, and adding noise suppression devices such as serrated trailing edge to the blades.

[0003] At present, the wind farms that need noise optimization are mostly already built wind farms, which produce a certain degree of noise interference to the residential areas near the wind farms, causing the power of the wind farm to be artificially reduced for the purpose of controlling noise. The existing technology is to input the noise curve into the noise evaluation model and the noise propagation attenuation model to calculate the total sound power level and the noise attenuation amount, and then calculate the target optimization time according to the total sound power level and the noise attenuation amount. It is judged whether the noise of the wind farm meets the preset noise standard. If not, the conventional noise reduction is used to reduce the noise, and it is judged whether the noise after the conventional noise reduction meets the preset noise standard. Although this method has a certain effect, it lacks timeliness and cannot solve the problem of wind farm noise disturbing the public from the root, and has certain limitations. SUMMARY

[0004] The present application provides a wind farm noise control and power optimization method and device to solve the problem of wind farm noise disturbing the public from the root.

[0005] In the first aspect, the present application provides a wind farm noise control and power optimization method, which comprises: Obtaining multi-source noise data of the wind farm, and analyzing the characteristics of the multi-source noise data to construct a multi-source noise feature database; Based on the multi-source noise feature database, a noise propagation theoretical model containing terrain factors, vegetation factors, obstacle factors and meteorological factors is constructed; Based on the noise propagation theoretical model, a plurality of monitoring points in the wind farm are determined, and actual multi-source noise data and wind turbine operation data of the plurality of monitoring points are collected in real time; Based on the actual multi-source noise data and the wind turbine operation data, a multi-objective optimization function fused with constraint conditions is constructed, taking the minimization of the difference between the scheduling instructions, the wind turbine operation load and the noise exceeding amount as the target. Solving the multi-objective optimization function generates an optimal power distribution scheme of the fan under noise constraints.

[0006] The wind farm noise control and power optimization method provided by the application first accurately identifies the sources and spectral characteristics of aerodynamic, mechanical and electromagnetic noise through multi-source noise characteristic analysis and feature database construction, solving the problem of fuzzy noise source positioning in traditional technologies; then a noise propagation theoretical model is constructed in combination with topography, vegetation, obstacles and weather, greatly improving the prediction accuracy of noise propagation paths in complex environments and providing reliable basis for scientific layout of monitoring points; subsequently, a multi-objective optimization function is constructed with the minimum difference between scheduling instructions, fan operating load and noise exceeding amount as the target based on real-time collected actual noise and unit operation data, breaking the limitations of traditional technologies that only focus on noise reduction or power and ignore multi-objective balance; finally, an optimal power distribution scheme is output through efficient solution, which can not only ensure that the noise of the wind farm meets the environmental protection standards of the residential area, but also reduce the loss of power generation caused by blind noise reduction, while ensuring the safety of fan operation and the response accuracy of power grid scheduling, effectively alleviating the contradiction between the wind farm and the residential area, and constructing a full-chain technology system of "noise source analysis-multi-factor propagation prediction-dynamic monitoring-multi-objective collaborative optimization-land execution", which solves the problem of wind farm noise disturbance from the root. In an optional implementation, the multi-source noise data includes fan blade near-field transient flow field data, fan gearbox mechanical noise and electromagnetic noise data; The characteristics of the multi-source noise data are analyzed, and a multi-source noise feature database is constructed, including: The fan blade near-field transient flow field data is obtained by CFD simulation, the sound propagation is solved by combining the FW-H equation and computational aeroacoustics, the spectral characteristics of the turbulent boundary layer noise, the trailing edge noise and the tip vortex noise are analyzed, and the aerodynamic noise characteristics are obtained; The generation mechanism of the fan gearbox mechanical noise is analyzed by using acoustic emission detection technology and vibration signal analysis, and the mechanical noise characteristics are obtained; Based on electromagnetic field theory and acoustic measurement, the generation mechanism of the electromagnetic noise of the generator set is analyzed, and the electromagnetic noise characteristics are obtained; Based on the aerodynamic noise characteristics, the mechanical noise characteristics and the electromagnetic noise characteristics, the sound field contribution of each noise source is quantitatively evaluated by coherent analysis and sound source separation technology, and a multi-source noise feature database is constructed.

[0007] The application provides a wind farm noise control and power optimization method, which precisely includes three types of core noise data of fan blade near-field transient flow field, gear box machinery and generator set electromagnetism, has strong comprehensive data coverage, breaks the limitation of traditional methods focusing on single noise source, and realizes complete capture of key noises of the wind farm; for aerodynamic noise, flow field data are obtained through CFD simulation, and the spectrum characteristics of turbulent boundary layer, trailing edge and tip vortex noise are analyzed in combination with FW-H equation and calculation aerodynamic acoustics, so that the shortage of traditional surface sound pressure level measurement is broken through; for mechanical noise, noise generation mechanism is deeply excavated relying on acoustic emission detection (capturing microscopic damage) and vibration signal analysis (quantifying macroscopic response); for electromagnetic noise, electromagnetic field theory (tracing force wave root) and acoustic measurement (verifying actual noise) are combined to ensure that the analysis has theoretical support and measured basis; the sound field contribution degree of each noise source is quantified through coherence analysis and sound source separation technology, the core control object is determined, and the multi-source noise characteristic database is constructed, which has comprehensive data, accurate characteristics and clear mechanism, can provide high-quality basic support with strong targeting for subsequent noise propagation theory model construction and power optimization, and effectively solves the problem of "scattered data, fuzzy characteristics and no clear control focus" in traditional noise characteristic analysis.

[0008] In an optional implementation, based on the multi-source noise characteristic database, a noise propagation theory model containing terrain factors, vegetation factors, obstacle factors and meteorological factors is constructed, including: Based on the multi-source noise characteristic database, a sound propagation model is constructed by using ray tracing method and parabolic equation method; The influence of terrain characteristic parameters on sound wave diffraction and reflection is quantitatively analyzed, a quantitative relationship between the terrain characteristic parameters and the sound attenuation rate is constructed, and a vegetation noise reduction model is constructed based on the quantitative relationship; The shielding effect of the geometric parameters of the obstacle on sound wave propagation is analyzed, and a meteorological coupling model is constructed based on the shielding effect and meteorological factors; Based on the sound propagation model, the vegetation noise reduction model and the meteorological coupling model, a noise propagation theory model is constructed.

[0009] This invention provides a wind farm noise control and power optimization method. It relies on a multi-source noise characteristic database to obtain accurate core parameters of noise sources (such as the spectral characteristics of aerodynamic, mechanical, and electromagnetic noise), providing a highly targeted initial basis for subsequent propagation calculations and avoiding prediction biases caused by fuzzy noise source inputs in traditional models. It integrates the ray tracing method (excellent at analyzing sound wave reflection and diffraction paths) and the parabolic equation method (adaptable to long-distance propagation calculations in complex terrain) to construct a sound propagation model. Compared to single modeling methods, this method can more accurately capture the sound wave propagation patterns in complex environments. By quantitatively analyzing the relationship between terrain feature parameters (such as slope and roughness) and sound attenuation rate, it breaks through the limitations of traditional qualitative descriptions of terrain influences, and the constructed vegetation noise reduction model is quantifiable. The noise reduction effect of parameters such as coverage and type has a multi-factor analysis that is both in-depth and targeted. At the same time, it focuses on the geometric parameters of obstacles (height, width, porosity) to analyze the shielding effect, and constructs a coupled model with meteorological factors such as temperature and wind speed to achieve a refined consideration of key variables affecting noise propagation. By integrating the sound propagation model, vegetation noise reduction model and meteorological coupled model, a complete theoretical model covering the four core factors of terrain, vegetation, obstacles and meteorology is formed. It can accurately predict the propagation path and sound pressure level distribution of wind farm noise under different environmental conditions, effectively solving the problems of insufficient consideration of multiple factors and insufficient analysis depth in traditional noise propagation models. It provides highly reliable theoretical support for the subsequent scientific layout of monitoring points and the formulation of noise control strategies.

[0010] In one optional implementation, multiple monitoring points in the wind farm are determined based on a noise propagation theory model, and real-time multi-source noise data and wind turbine operation data from these monitoring points are collected, including: Candidate areas for monitoring points are determined based on the noise propagation theory model. These candidate areas include the strong radiation area of ​​the noise source, the key propagation path, the sensitive protection area, and the model validation blind zone. Within the candidate region, an improved genetic algorithm is used to iteratively optimize the number and location of monitoring points to obtain the optimal set of monitoring points. Based on the optimal set of monitoring points, distributed monitoring points are deployed within the wind farm, and real-time multi-source noise data and wind turbine operation data from multiple monitoring points are collected in accordance with a dynamic sampling strategy.

[0011] This invention provides a wind farm noise control and power optimization method. Based on a noise propagation theory model, it clearly defines the strong radiation area of ​​the noise source (capturing original noise characteristics), key propagation paths (tracking noise attenuation patterns), sensitive protection areas (assessing actual impact on residents), and model validation blind spots (feedback to model optimization). This completely solves the problem of blind point placement in traditional methods, ensuring no core omissions or redundant coverage within the monitoring range. By improving the genetic algorithm to iteratively optimize the number and location of points within the candidate area, prioritizing sound field uniformity, terrain adaptability, and cost minimization, it achieves a more even distribution compared to manual point placement. Balancing monitoring accuracy and engineering costs, this approach significantly improves the scientific rigor and feasibility of site selection. Distributed monitoring nodes are deployed based on optimal locations, employing a dynamic sampling strategy (low-frequency sampling during stable noise conditions and high-frequency capture during abrupt changes) to collect multi-source noise data in real time. Simultaneously, wind turbine operation data is acquired, ensuring both the timeliness and effectiveness of noise data and achieving deep linkage between noise and operating parameters. This provides a high-quality, strongly correlated real-time data foundation for the subsequent construction of multi-objective optimization functions, effectively avoiding the limitations of traditional data acquisition methods, such as isolation, inefficiency, and poor adaptability, and providing reliable data support for power optimization under noise constraints.

[0012] In one optional implementation, based on actual multi-source noise data and wind turbine operation data, a multi-objective optimization function with fused constraints is constructed, aiming to minimize scheduling command differences, wind turbine operating load, and noise exceedance, including: Based on actual multi-source noise data and wind turbine operation data, the deviation between the actual total power of the wind farm and the power commanded by the power grid is calculated as the dispatch command difference; the degree to which the operating load of a single wind turbine exceeds the safety limit is calculated as the wind turbine operating load; and the degree to which the noise at monitoring points around the wind farm exceeds the preset noise standard is calculated as the noise exceeding the standard. With the objectives of minimizing the deviation between the actual total power of the wind farm and the power commanded by the grid dispatch, minimizing the degree to which the operating load of a single wind turbine exceeds the safety limit, and minimizing the degree to which the noise at monitoring points around the wind farm exceeds the preset noise standard, a multi-objective optimization function with integrated constraints is constructed. The constraints include noise constraints, power constraints, and operational safety constraints.

[0013] This invention provides a wind farm noise control and power optimization method that uses real-time collected multi-source noise data and wind turbine operating data as the basis for calculation, rather than relying on theoretical assumptions. This ensures that the calculation of target parameters such as dispatch command differences, wind turbine operating load, and noise exceedances closely match the actual operating state of the wind farm, avoiding a disconnect between optimization and reality. It focuses on three core dimensions: grid response (dispatch command differences), equipment safety (wind turbine operating load), and environmental requirements (noise exceedances), covering the key needs of the grid, equipment, and residential areas in wind power operation. This breaks away from traditional optimization methods that only focus on noise reduction or power, neglecting multiple objectives. This approach overcomes the limitations of traditional optimization functions, and defines each objective through quantifiable indicators such as deviation value and degree of exceedance, making the optimization direction clear and controllable. Noise constraints (ensuring environmental compliance in residential areas) are used as the core constraint, while power constraints (ensuring the power of a single wind turbine and the entire field is within a safe range) and operational safety constraints (avoiding equipment overload damage) are simultaneously incorporated. This not only defines reasonable boundaries for optimization and avoids outputting infeasible solutions, but also ensures that the final optimization result meets environmental protection standards while taking into account the accuracy of power grid dispatch and the safety of wind turbine operation through the correspondence between constraints and objectives. This effectively solves the problems of one-sided objectives, loose constraints, and insufficient practicality of traditional optimization functions.

[0014] In one alternative implementation, the multi-objective optimization function is calculated using the following formula: ; in, This represents the actual total power of the wind farm. For power grid dispatch command power, Single fan operating load, For safety limits, For noise from monitoring points around the wind farm, This is a preset noise standard.

[0015] In one optional implementation, the noise constraint is that the noise at all monitoring points in the target area does not exceed the upper limit of the noise in the target area, the power constraint is that the power of a single wind turbine is within the safe operating range, and the operation safety constraint is that the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command. Solving the multi-objective optimization function generates the optimal power allocation scheme for wind turbines under noise constraints, including: Under the constraints that the noise at all monitoring points in the target area does not exceed the upper limit of the target area noise, the power of a single wind turbine is within the safe operating range, and the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command, an improved particle swarm optimization algorithm is used to solve the multi-objective optimization function, and the optimal power allocation scheme of wind turbines under noise constraints is generated through iterative optimization.

[0016] This invention provides a method for noise control and power optimization in wind farms. It sets a strict environmental bottom line: noise levels at all monitoring points in the target area must not exceed the upper limit; the power of a single wind turbine must be within a safe range to ensure basic equipment operation; and the load corresponding to the wind turbine power must not exceed the load demand of the dispatch command to connect with the grid. These three constraints are progressively layered, covering key dimensions of environmental protection, equipment, and the grid, completely avoiding the problem of infeasible solutions caused by fuzzy constraints in traditional optimization. An improved particle swarm optimization algorithm can accurately handle multi-objective optimization scenarios under multiple constraints. Through iterative processes, it dynamically balances the three objectives of "minimizing dispatch command differences, wind turbine operating load, and noise exceeding the scalar limit," making it easier to escape local optima and efficiently find the globally optimal solution that balances environmental protection, equipment safety, and grid response compared to traditional algorithms. The generated optimal power allocation scheme under noise constraints ensures that noise levels in residential areas meet standards to alleviate NIMBY (Not In My Backyard) conflicts, while avoiding power generation losses due to blindly reducing noise or ignoring equipment overload risks in pursuit of power. This achieves synergistic optimization of wind farm environmental compliance, equipment safety, and grid response, effectively improving the overall operational efficiency and sustainability of wind farms.

[0017] In one alternative implementation, the method further includes: Based on the optimal power allocation scheme for wind turbines, hierarchical decision control is implemented for wind farms.

[0018] In one optional implementation, based on the optimal power allocation scheme for wind turbines, hierarchical decision control of the wind farm is performed, including: Based on the optimal power allocation scheme for wind turbines, an improved particle swarm optimization algorithm is used to allocate the power of the entire wind farm, and the weights of the constraints are adjusted in real time according to the environmental noise monitoring data of the wind farm to build a collaborative mechanism for short-term scheduling and long-term optimization at the farm level. Adaptive PID control is used to track the power of a single wind turbine. By configuring a noise feedback adjustment unit and interacting in real time with a preset acoustic monitoring system, a multi-level safety protection mechanism at the unit level is constructed. A preset communication protocol is used to perform collaborative optimization control at the site level and unit level, and optimization calculations are performed periodically to form a closed-loop power control for the wind farm. This invention provides a wind farm noise control and power optimization method. Through a site-level and unit-level collaborative and closed-loop mechanism, it achieves precise, safe, and dynamic power control of the wind farm. Specifically, at the site level, an improved particle swarm optimization algorithm is used to allocate power across the entire farm. Combined with environmental noise monitoring data, constraint weights are adjusted in real time. This ensures short-term dispatch response to grid demands while also balancing noise reduction and energy efficiency in the long term, solving the problems of static rigidity and lack of long-term / short-term coordination in traditional site control. At the unit level, adaptive PID control accurately tracks the power of individual units. Combined with a noise feedback adjustment unit and real-time interaction with the acoustic system, a multi-level safety protection mechanism is constructed. This ensures that individual unit operation conforms to the power plan while avoiding noise exceeding limits and equipment risks, overcoming the limitations of insufficient precision and low safety redundancy in single control modes. Simultaneously, a preset communication protocol enables efficient collaboration between the site and units. Periodic optimization calculations form a closed-loop control, allowing the entire wind farm to dynamically adapt to noise changes, grid commands, and equipment status, ultimately achieving integrated control that is globally optimized, locally precise, and safely controllable.

[0019] In a second aspect, the present invention provides a wind farm noise control and power optimization device, the device comprising: The multi-source noise data acquisition and feature extraction module is used to acquire multi-source noise data from wind farms, analyze the characteristics of multi-source noise data, and construct a multi-source noise feature database. The noise propagation theory model building module is used to construct a noise propagation theory model that includes topographic factors, vegetation factors, obstacle factors, and meteorological factors based on a multi-source noise feature database. The monitoring point deployment and data acquisition module is used to determine multiple monitoring points in the wind farm based on the noise propagation theory model, and to collect the actual multi-source noise data and wind turbine operation data of multiple monitoring points in real time. The multi-objective optimization function construction module is used to construct a multi-objective optimization function with fused constraints based on actual multi-source noise data and wind turbine operation data, with the objectives of minimizing scheduling command differences, wind turbine operating load and noise overshoot. The function solving module is used to solve multi-objective optimization functions and generate the optimal power allocation scheme for wind turbines under noise constraints.

[0020] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the wind farm noise control and power optimization method of the first aspect or any corresponding embodiment described above.

[0021] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the wind farm noise control and power optimization method of the first aspect or any corresponding embodiment described above.

[0022] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the wind farm noise control and power optimization method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0023] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the first step of the wind farm noise control and power optimization method according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the second process of the wind farm noise control and power optimization method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the third process of the wind farm noise control and power optimization method according to an embodiment of the present invention. Figure 4 This is a schematic diagram of synchronous acquisition of mechanical noise signals in the wind farm noise control and power optimization method according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the automatic electromagnetic noise risk assessment process in the wind farm noise control and power optimization method according to an embodiment of the present invention. Figure 6 This is a schematic diagram of the multi-source noise acquisition and feature extraction process in the wind farm noise control and power optimization method according to an embodiment of the present invention. Figure 7 This is a structural block diagram of a wind farm noise control and power optimization device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0027] For existing wind turbines in wind farms, noise reduction measures have certain limitations compared to newly built wind farms. However, most current research does not address the root causes of wind farm noise; it simply involves modifying existing equipment, which fails to fundamentally reduce the noise impact on residential areas. While research on wind turbine noise and noise reduction methods is mature, there is currently no systematic approach to addressing noise pollution from the perspectives of energy efficiency management, noise sensitivity in residential areas at different times of day, and grid load demand. Furthermore, existing noise monitoring systems are mostly deployed independently, lacking deep integration with wind farm control systems, thus failing to achieve dynamic noise control.

[0028] This invention provides a method for wind farm noise control and power optimization. By using intelligent acoustic sensing and wind farm group collaborative control, it overcomes the limitations of traditional noise reduction technologies, achieves precise, dynamic, and intelligent noise control, and effectively alleviates the conflict between wind farms and residential areas. It is expected to effectively reduce the cost of wind farm noise control.

[0029] According to an embodiment of the present invention, a method for wind farm noise control and power optimization is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] This embodiment provides a method for wind farm noise control and power optimization, which can be used in wind farms. Figure 1 This is a flowchart of a wind farm noise control and power optimization method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain multi-source noise data from the wind farm, analyze the characteristics of the multi-source noise data, and construct a multi-source noise feature database.

[0031] Specifically, for different core noise sources in a wind farm, multi-source noise data includes aerodynamic noise data, mechanical noise data, and electromagnetic noise data. Its characteristics revolve around source differentiation, quantifiable features, and clear correlation factors. By analyzing the characteristics of multi-source noise data, the corresponding noise source characteristics are obtained, and a structured multi-source noise feature database is constructed.

[0032] This step relies on a multi-source noise characteristic analysis system for wind turbine generators based on acoustic measurements and numerical simulations.

[0033] Step S102: Based on the multi-source noise feature database, construct a noise propagation theoretical model that includes topographic factors, vegetation factors, obstacle factors, and meteorological factors.

[0034] Specifically, based on the core parameters of noise sources (such as the spectral intensity and location of each noise type) provided by the multi-source noise feature database, a noise propagation theoretical model is constructed in modules: first, a basic sound propagation model is constructed, and then the basic sound propagation model is integrated with sub-models of terrain, vegetation, obstacles, and meteorology to form a noise propagation theoretical model covering four key influencing factors, so as to achieve accurate prediction of the noise propagation path and the sound pressure level reached in wind farms.

[0035] Step S103: Based on the noise propagation theory model, determine multiple monitoring points in the wind farm, and collect the actual multi-source noise data and wind turbine operation data of multiple monitoring points in real time.

[0036] Specifically, in order to collect actual wind turbine operation data and actual multi-source noise data of the wind farm, multiple distributed monitoring points in the wind farm are determined based on the noise propagation theory model. By deploying acoustic cameras and microphone arrays at each monitoring point, accurate collection of actual wind turbine operation data and actual multi-source noise data of the wind farm can be achieved.

[0037] Step S104: Based on actual multi-source noise data and wind turbine operation data, construct a multi-objective optimization function with fused constraints, aiming to minimize the difference in scheduling instructions, wind turbine operating load, and noise excess.

[0038] Specifically, to resolve the contradictions among the three core requirements of environmental compliance, grid response, and equipment safety in wind farm operation, and to break away from the limitations of traditional optimization schemes that prioritize single objectives, are detached from reality, and lack boundaries, it is necessary to base optimization on actual multi-source noise data and wind turbine operation data. This ensures that the optimization closely matches the real operating scenario of the wind farm, avoids the disconnect between theory and practice, and constructs an objective that covers the core requirements of the grid, equipment, and environmental protection. Specifically, the objective is to minimize the difference in dispatch instructions, wind turbine operating load, and noise exceedance. A multi-objective optimization function with integrated constraints is constructed to solve the problem of single-objective optimization that suffers from neglecting one aspect while focusing on another. The constraints are used to define the optimization boundary to avoid the infeasibility of the optimal solution and ensure the feasibility of the solution.

[0039] Step S105: Solve the multi-objective optimization function to generate the optimal power allocation scheme for the wind turbine under noise constraints.

[0040] Specifically, under the premise of satisfying the three major constraints of no noise exceeding the standard at the monitoring points in the target area, power of a single wind turbine within the safe range, and wind turbine load not exceeding the upper limit of the scheduled load, an adaptive multi-objective optimization algorithm (such as an improved particle swarm optimization algorithm) is adopted to solve the multi-objective optimization function. Through the algorithm iteration process, the target power parameters of each wind turbine are dynamically adjusted: in the initial stage, multiple sets of wind turbine power allocation schemes are generated (i.e., particles in the algorithm), the optimization function value (fitness value) corresponding to each scheme is calculated, the scheme with better fitness value is retained (i.e., individual extreme value and global extreme value), and the optimal solution is gradually approached by adjusting the search strategy (such as dynamic inertia weight and learning factor).

[0041] After multiple iterations (until the change in the optimal fitness value of consecutive iterations is less than a preset threshold), the final output is the optimal power allocation scheme for wind turbines that satisfies all constraints and minimizes the differences in scheduling instructions, wind turbine operating load, and noise overrun, thus clarifying the target output power of each wind turbine.

[0042] The wind farm noise control and power optimization method provided in this embodiment first accurately identifies the root causes and spectral characteristics of aerodynamic, mechanical, and electromagnetic noise through multi-source noise characteristic analysis and feature database construction, solving the problem of ambiguous noise source localization in traditional technologies. Then, it combines terrain, vegetation, obstacles, and meteorological factors to construct a noise propagation theoretical model, significantly improving the prediction accuracy of noise propagation paths in complex environments and providing a reliable basis for the scientific layout of monitoring points. Subsequently, based on real-time collected actual noise and turbine operation data, it constructs a system that minimizes "dispatch command differences, turbine operating load, and noise exceedance." The multi-objective optimization function with the goal of "noise source analysis - multi-factor propagation prediction - dynamic monitoring - multi-objective collaborative optimization - implementation" breaks through the limitations of traditional technologies that focus solely on noise reduction or power output while ignoring the balance of multiple objectives. Ultimately, it outputs the optimal power allocation scheme through efficient solution, which can ensure that the noise of wind farms meets the environmental protection standards of residential areas, reduce the power generation loss caused by blind noise reduction, and at the same time ensure the safe operation of wind turbines and the accuracy of grid dispatch response. It effectively alleviates the conflict between wind farms and residential areas and builds a full-chain technical system of "noise source analysis - multi-factor propagation prediction - dynamic monitoring - multi-objective collaborative optimization - implementation", which solves the problem of wind farm noise pollution from the root.

[0043] This embodiment provides a method for wind farm noise control and power optimization, which can be used in wind farms. Figure 2 This is a flowchart of a wind farm noise control and power optimization method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain multi-source noise data from the wind farm, analyze the characteristics of the multi-source noise data, and construct a multi-source noise feature database.

[0044] Specifically, the multi-source noise data includes near-field transient flow field data of the wind turbine blades, mechanical noise data of the wind turbine gearbox, and electromagnetic noise data; such as Figure 6 As shown, step S201 above includes: Step S2011: CFD simulation is used to obtain near-field transient flow field data of the wind turbine blades. The sound propagation is solved by combining the FW-H equation and computational aeroacoustics. The spectral characteristics of turbulent boundary layer noise, trailing edge noise and tip vortex noise are analyzed to obtain aerodynamic noise characteristics.

[0045] Specifically, the generation mechanism of aerodynamic noise of wind turbine blades is modeled and analyzed by using the sound intensity method and near-field acoustic holography, combined with CFD (Computational Fluid Dynamics) numerical simulation. A hybrid modeling method is used, and the specific modeling steps are as follows: Step 1: Unsteady Flow Field Simulation (Computational Fluid Dynamics, CFD). The purpose is to obtain transient pressure, velocity, and density fluctuation data (as a sound source) on the blade surface and near field. Large Eddy Simulation is used to explicitly solve the large-scale turbulent structure, which can accurately capture turbulent fluctuations, separated flows, and eddy dynamics.

[0046] Step 2: Acoustic Modeling and Propagation Calculation. Using CFD simulation results as the sound source, far-field noise is calculated. An acoustic analogy method is employed, using the FW-H equation (Fowcs Williams-Hawkings equation) to treat the wind turbine blade surface as a moving solid boundary, integrating the surface pulsating pressure (…). p The time-domain CFD simulation data and volumetric flux are calculated. Computational aeroacoustics (CAA) is used to directly solve the linearized Euler equations (LEE) or the linearized Navier-Stokes equations (LNSE). Finally, the time-domain CFD simulation data is transformed to the frequency domain using FFT, and the sound propagation is solved using the Green's function or the boundary element method (BEM).

[0047] The aerodynamic noise signal of the wind turbine is obtained through the above modeling steps. The focus is on analyzing the spectral characteristics of turbulent boundary layer noise, trailing edge noise, and tip vortex noise. Specifically, the time-domain CFD and acoustic measurement data are converted to the frequency domain through FFT, and spectral parameters such as the characteristic frequency bands, peak sound pressure level of each frequency band, and frequency band energy ratio of turbulent boundary layer noise, trailing edge noise, and tip vortex noise are extracted to obtain the aerodynamic noise characteristics.

[0048] In step S2012, the generation mechanism of mechanical noise in the wind turbine gearbox is analyzed using acoustic emission detection technology and vibration signal analysis to obtain mechanical noise characteristics.

[0049] The generation mechanism of gearbox mechanical noise was studied using acoustic emission (AE) detection technology and vibration signal analysis. A schematic diagram of synchronous acquisition of mechanical noise signals was presented, illustrating the synergistic analysis of acoustic emission, vibration, and noise modes. Figure 4 As shown, the specific steps are as follows: Step 1, Acoustic emission (AE) feature extraction: AE technology captures the origin of micro-damage (such as the initiation of initial cracks on the tooth surface), that is, it captures the AE signal parameters of micro-damage in the gearbox (such as initial cracks on the tooth surface), such as AE signal peak value, rise time, pulse width, and energy value.

[0050] Step 2, Vibration response feature extraction: Vibration analysis quantifies the macroscopic dynamic response (such as meshing impact intensity), and quantifies the vibration signal parameters of the gearbox macroscopic dynamic response, such as meshing impact intensity, vibration frequency components (such as gear meshing frequency, bearing characteristic frequency), and root mean square vibration acceleration.

[0051] Step 3, Sound radiation characteristics: The "mechanical noise sound pressure level data" (including the sound pressure level distribution of different frequency bands) collected by the microphone, and the correlation data between the vibration measurement point and the sound pressure calculated by the vibration-sound radiation transfer function, are used to confirm the final noise radiation characteristics through sound signal measurement.

[0052] The vibration-radiative acoustic transfer function (VATF) is calculated using the following formula: (1); in, Let be the transfer function from vibration to sound pressure at the i-th monitoring point, obtained through hammer impact testing. Based on the vibration frequency response, the mechanical noise characteristics are obtained by combining the above acoustic emission (AE) characteristics, vibration response characteristics, and acoustic radiation characteristics.

[0053] Step S2013: Based on electromagnetic field theory and acoustic measurement, analyze the generation mechanism of electromagnetic noise of generator set and obtain electromagnetic noise characteristics.

[0054] Specifically, based on electromagnetic field theory and acoustic measurements, the generation mechanism of generator electromagnetic noise is analyzed, and the specific steps are as follows: Step 1, Basic operating parameter characteristics: Read data from the wind turbine SCADA system (Supervisory Control And Data Acquisition) and perform data input and preprocessing.

[0055] Step 2, Electromagnetic Force Wave Characteristics: Perform rapid calculation of electromagnetic force waves. The core formula is the spatial order formula for force waves, as follows: (2); in: The number of stator slots The number of rotor slots, , For the harmonic order of the magnetic field, These are the order correction coefficients. It is an electromagnetic force wave.

[0056] And the formula for the time frequency of electromagnetic force waves: (3); in: This is the motor's base frequency. This represents the number of pole pairs of the motor. For time harmonics, The frequency of the electromagnetic force wave is denoted by .

[0057] Step 3, Noise Risk and Proportion Characteristics: Automatic noise risk assessment, assessment method as follows... Figure 5 As shown, the electromagnetic noise risk assessment results (high risk / low risk, based on whether the force wave order is ≤4 and whether it is close to the structural mode) and the difference data between mechanical noise and operating noise after shutdown (used to quantify the proportion of electromagnetic noise in the total noise) are analyzed to determine the proportion of electromagnetic noise.

[0058] Step 4, Pole-slot combination-noise correlation characteristics: The influence of different pole-slot combinations and air gap magnetic fields on electromagnetic noise is investigated, and electromagnetic noise suppression strategies are optimized. Finally, the electromagnetic noise characteristics are obtained.

[0059] Step S2014: Based on aerodynamic noise characteristics, mechanical noise characteristics and electromagnetic noise characteristics, the sound field contribution of each noise source is quantitatively evaluated through coherent analysis and sound source separation technology, and a multi-source noise characteristic database is constructed.

[0060] Specifically, through coherence analysis and sound source separation techniques, the sound field contribution of each noise source is quantitatively evaluated, and a multi-source noise characteristic database is constructed to provide data support for noise optimization design. The formula for calculating the sound-vibration coherence function is as follows: (4); in: sound pressure and vibration The mutual power spectrum, To compose one's own music To resonate with one's own music; A value greater than 0.8 indicates a strong correlation.

[0061] For example, the proportion of the sound field contribution of aerodynamic noise, mechanical noise, and electromagnetic noise in the total noise (e.g., aerodynamic noise accounts for 70%, mechanical noise accounts for 20%, and electromagnetic noise accounts for 10% under a certain working condition), and the contribution distribution of each noise source in different frequency bands (e.g., mechanical noise contributes more in the low frequency band).

[0062] Step S202: Based on the multi-source noise feature database, construct a noise propagation theoretical model that includes topographic factors, vegetation factors, obstacle factors, and meteorological factors.

[0063] Specifically, the research on the noise propagation characteristics and prediction model of wind turbine generators focuses on solving key technical problems such as accurate prediction of noise propagation paths and characterization of noise characteristics under multiple operating conditions. This method systematically explores the influence mechanism of factors such as topography and atmospheric conditions on noise propagation, constructs a noise prediction model considering multiple influencing factors, and provides a theoretical basis for wind farm noise assessment and control. Traditional methods for studying noise propagation path characteristics have large errors in predicting sound attenuation, failing to meet the requirements for high-precision noise prediction. This embodiment proposes a comprehensive strategy for studying noise propagation path characteristics that considers the influence of multiple factors to improve the accuracy of acoustic models. Step S202 includes: Step S2021: Based on the multi-source noise feature database, a sound propagation model is constructed using the sound ray tracing method and the parabolic equation method.

[0064] Specifically, a high-precision three-dimensional sound propagation model is constructed using the ray tracing (RT) method and the parabolic equation (PE) method. The sound wave reflection, diffraction, and scattering mechanisms under complex environments are systematically studied. The formula for the sound ray trajectory is as follows: (5); in, This represents the initial vertical position of the sound emission point. The initial elevation angle of the sound ray (the angle between the ray and the horizontal plane). For the sound speed gradient, Let be the initial altitude and speed of sound. Using the above formula and combining it with the parabolic equation, a sound propagation model is constructed. This is a horizontal position variable used to describe the relationship between height and horizontal position.

[0065] Step S2022: Quantitatively analyze the influence of terrain feature parameters on sound wave diffraction and reflection, construct a quantitative relationship between terrain feature parameters and sound attenuation rate, and construct a vegetation noise reduction model based on the quantitative relationship.

[0066] Specifically, the influence of topographic relief on sound wave diffraction and reflection is quantitatively analyzed, and a quantitative relationship between topographic feature parameters (such as slope and roughness) and sound attenuation rate is constructed.

[0067] The reflection from the ground during propagation can be described by the reflection coefficient, the specific expression of which is as follows: (6); In the formula, the first term on the right represents the solution for a spherical wave in an unbounded free field; the second term represents the solution for the sound wave at the receiving point due to ground reflection or a mirror image sound source. Indicates wave number, The horizontal distance between the sound source and the receiving point. and These represent the heights of the receiving point and the sound source point from the ground, respectively. The reflection coefficient of a spherical wave. The reflection coefficient, The amplitude (or intensity) of the sound source.

[0068] Vegetation noise reduction research: This study investigates the sound absorption characteristics of different types of vegetation (trees, shrubs, grasslands) and constructs a mathematical model of the relationship between vegetation coverage, vegetation type, and noise reduction effect.

[0069] The main mechanisms of vegetation noise reduction are as follows: i. Absorption: Friction on the surfaces of plant leaves, branches, and trunks converts sound energy into heat energy (primarily effective at mid- to high frequencies). Leaf area index is an important indicator.

[0070] ii. Scattering: The complex branching structure causes sound waves to undergo multiple irregular reflections and diffractions, which destroys the coherence of sound waves and disperses sound energy.

[0071] iii. Ground effect: Vegetation cover alters surface properties (e.g., grassland replaces hard ground) and enhances ground absorption (especially at low frequencies).

[0072] iV. Barrier effect: Especially dense belts of trees and shrubs can form physical barriers, reducing noise propagation through diffraction and partial blocking (mainly effective for mid- to high-frequency frequencies).

[0073] The objective function for noise reduction (i.e., the vegetation noise reduction model) is constructed as follows: (7); in, For noise reduction, These are, respectively, vegetation cover, vegetation type, three-dimensional green volume (m² / m² or m³ / m²), structural parameters, the combination of distances (m) from the sound source to the vegetation front and from the vegetation to the receiving point, and frequency. This is the error term.

[0074] Step S2023: Analyze the shielding effect of the geometric parameters of the obstacle on the propagation of sound waves, and construct a meteorological coupling model based on the shielding effect and meteorological factors.

[0075] Specifically, obstacle shielding effect analysis: analyze the shielding effect of the geometric characteristics (height, width, porosity) of obstacles such as buildings and windbreaks on sound wave propagation, and optimize the layout of obstacles to reduce noise propagation.

[0076] The analysis of the shading effect is as follows: Phase 1: Preliminary research and parameter determination, including: 1.1 Define the research boundaries and objectives: Define the noise source type: differentiate between traffic (mid-to-low frequency, 63-1000Hz), industry (broadband, 250-4000Hz), airport (low frequency, 31.5-500Hz), etc., and determine the core noise reduction frequency band; Delineate the protected area: Identify the areas where noise reduction is needed (residential areas, schools), and mark the coordinates of the receiving point and the target sound pressure level (e.g., ≤55dB during the day and ≤45dB at night).

[0077] 1.2 Obstacle and Acoustic Parameter Acquisition: Geometric parameter measurement: Buildings: Use a total station to measure height (from eaves to ground), width (horizontal length along the direction of noise propagation), and facade porosity (the proportion of window / door opening area to total area); Windbreak forest: The average tree height and forest belt width were measured using the quadrat method (10m×10m quadrats), and the porosity (leaf and branch coverage area / quadrats area) was calculated using image analysis (drone aerial photography + AI recognition). Acoustic parameter monitoring: Use a sound level meter to set up monitoring points at the noise source (such as the road red line) and 10m / 20m / 50m before and after the obstacle, monitor continuously for 24 hours, and record the sound pressure level (A-weighted) and propagation attenuation at different frequencies. Phase 2: Masking Effect Simulation and Problem Diagnosis 2.1 Select a suitable simulation model: Choose the appropriate tool based on the type of obstacle to ensure a balance between simulation accuracy and efficiency: Rigid obstacles (buildings): The ray tracing method (software Odeon) is used to simulate the direct, reflected, and diffracted paths of sound waves and calculate the sound pressure level distribution in the sound shadow area; Porous obstacles (windbreaks): An equivalent acoustic impedance model (Sound Plan software) is used. Parameters such as porosity and leaf density are input, transmission and absorption losses are corrected, and noise attenuation curves before and after the forest belt are output.

[0078] 2.2 Comparison and analysis of simulation results and current situation: Compare simulated values ​​with field monitoring values ​​to verify the accuracy of the model (error must be ≤3dB). Diagnose existing problems such as "insufficient building height resulting in the sound shadow zone not covering residential areas", "excessive porosity of windbreak forests leading to excessive transmitted noise", and "excessive spacing of obstacles forming sound wave channels", and clarify the direction for optimization.

[0079] Meteorological factor coupling modeling: Considering the influence of meteorological factors such as temperature gradient, humidity, wind speed and wind direction on sound wave refraction, a meteorological-sound propagation coupling model is constructed to improve the prediction accuracy under different climatic conditions.

[0080] Among them, the relationship between temperature, sound speed, and meteorological factors is as follows: the sound speed is directly proportional to the square root of the absolute temperature, and the formula is: (8); in, R is the specific heat ratio, T is the universal gas constant, and M is the absolute temperature. A temperature gradient causes a change in the speed of sound in the vertical direction, thus causing sound wave refraction.

[0081] Humidity: The effect of humidity on the speed of sound can be accounted for by adjusting for air density and elastic modulus. Generally, as humidity increases, air density decreases, and the speed of sound increases slightly. Commonly used calibrated coupler models can describe the speed of sound in humid air.

[0082] Wind speed and direction: Wind speed is added to the speed of sound propagation to form the effective speed of sound. With the wind at your back, the effective speed of sound is the sum of the speed of sound and the wind speed; against the wind, the effective speed of sound is the difference between the speed of sound and the wind speed. Wind direction changes the direction of sound wave propagation.

[0083] Step S2024: Construct a theoretical model of noise propagation based on the sound propagation model, vegetation noise reduction model, and meteorological coupling model.

[0084] Experimental verification and optimization: The accuracy of the model was verified through field experiments, and the model parameters were optimized by combining measured data to form a sound propagation prediction method suitable for the target wind farm, and the noise propagation path characteristics of the target wind farm were obtained.

[0085] Step S203 involves determining multiple monitoring points in the wind farm based on a noise propagation theory model, and then collecting real-time multi-source noise data and wind turbine operation data from these monitoring points. For details, please refer to [link to relevant documentation]. Figure 1 Step S103 of the illustrated embodiment will not be described again here.

[0086] Step S204: Based on actual multi-source noise data and wind turbine operation data, a multi-objective optimization function with fused constraints is constructed, aiming to minimize scheduling command differences, wind turbine operating load, and noise exceedance. For details, please refer to [link to relevant documentation]. Figure 1 Step S104 of the illustrated embodiment will not be described again here.

[0087] Step S205 involves solving the multi-objective optimization function to generate the optimal power allocation scheme for the wind turbine under noise constraints. For details, please refer to [link to relevant documentation]. Figure 1 Step S105 of the illustrated embodiment will not be described again here.

[0088] The wind farm noise control and power optimization method provided in this embodiment overcomes the limitations of traditional noise reduction technologies through intelligent acoustic perception and farm group collaborative control, achieving precise, dynamic, and intelligent noise control.

[0089] This embodiment provides a method for wind farm noise control and power optimization, which can be used in wind farms. Figure 3 This is a flowchart of a wind farm noise control and power optimization method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Acquire multi-source noise data from the wind farm, analyze the characteristics of the multi-source noise data, and construct a multi-source noise feature database. For details, please refer to [link to relevant documentation]. Figure 2 Step S201 of the illustrated embodiment will not be described again here.

[0090] Step S302: Based on a multi-source noise characteristic database, construct a noise propagation theoretical model that includes topographic factors, vegetation factors, obstacle factors, and meteorological factors. For details, please refer to [link to relevant documentation]. Figure 2 Step S202 of the illustrated embodiment will not be described again here.

[0091] Step S303: Based on the noise propagation theory model, determine multiple monitoring points in the wind farm, and collect the actual multi-source noise data and wind turbine operation data of multiple monitoring points in real time.

[0092] Specifically, step S303 includes: Step S3031: Determine the candidate areas for monitoring points based on the noise propagation theory model. The candidate areas for monitoring points include the strong radiation area of ​​the noise source, the key propagation path, the sensitive protection area, and the model verification blind zone.

[0093] Specifically, the candidate areas for monitoring points include areas with strong noise source radiation, key propagation paths, sensitive protection areas, and blind spots for model validation.

[0094] Among them: the strong radiation zone of the noise source: the near-field strong radiation range of high-contribution noise sources such as aerodynamic noise (blade trailing edge / blade tip vortex), mechanical noise (gearbox / generator), and electromagnetic noise (motor) predicted by the model (usually the area 50-100m outside the noise source). Monitoring points need to be set up in this area to capture the original noise characteristics. Key propagation paths: The main propagation direction of noise calculated by the model (such as the path extending downhill along the terrain slope and in the downwind direction) and the energy attenuation inflection point (such as the boundary of the sound shadow zone formed by the obstruction of obstacles) need to be monitored at the path nodes to verify the propagation attenuation law. Sensitive protected areas: If the noise level of sensitive points such as residential areas and schools predicted by the model reaches the sound pressure level (it must meet the standard of ≤55dB during the day and ≤45dB at night), monitoring points need to be set up at the boundary and inside the sensitive area to assess the actual impact. Model validation blind spots: Areas in the model that may have prediction errors due to complex terrain (such as canyons and steep slopes) or dynamic weather changes (such as gusts) need to have additional monitoring points added to provide feedback for model optimization.

[0095] Step S3032: Within the candidate region, an improved genetic algorithm is used to iteratively optimize the number of monitoring points and their locations to obtain the optimal set of monitoring points.

[0096] Specifically, an improved genetic algorithm is employed, with the optimization objectives of sound field distribution uniformity, terrain feature adaptability, and monitoring cost minimization, to achieve efficient layout of monitoring points and ensure the spatial representativeness and economy of noise data. The core optimization objectives are as follows: Sound field distribution uniformity: Ensure that the monitoring points are spatially distributed reasonably within the target area, effectively capture sound pressure level changes at different locations, maximize the spatial coverage and representativeness of the sound field data, and avoid monitoring blind spots or data redundancy.

[0097] Terrain adaptability: The impact of complex terrain features (such as undulations, obstacles, water bodies, and buildings) on sound wave propagation in the actual monitoring area should be fully considered. The optimization process should proactively adapt to these features so that the location of monitoring points can both avoid sound field distortion caused by terrain and reflect the objective effect of terrain on noise distribution, thereby improving the authenticity and accuracy of the monitoring data.

[0098] Minimize monitoring costs: While meeting the above technical performance requirements, strictly control the overall monitoring costs. This includes, but is not limited to: the purchase and installation costs of monitoring equipment, long-term operation and maintenance costs (such as electricity, communication, and maintenance), and manpower and transportation costs related to the accessibility of monitoring points, to obtain the optimal set of monitoring points.

[0099] Step S3033: Based on the optimal set of monitoring points, distributed monitoring points are deployed in the wind farm, and real-time multi-source noise data and wind turbine operation data of multiple monitoring points are collected in accordance with the dynamic sampling strategy.

[0100] Specifically, low-power, high-precision noise sensor nodes, i.e., intelligent sensor node modules, are developed, integrating noise signal acquisition, preprocessing, and wireless transmission functions to support real-time data acquisition and synchronous monitoring of environmental parameters. For example, a distributed measurement system based on acoustic cameras and microphone arrays is designed, as follows: A1. Microphone Array Design: Array type: It adopts a hybrid array of "circular + linear". The circular array (30cm in diameter, 16 microphones) achieves 360° horizontal positioning, while the linear array (50cm in length, 8 microphones) improves vertical resolution. The total number of channels is 24.

[0101] Microphone parameters: Select a high-precision MEMS microphone (such as ADI ADMP401), with a frequency response of 20Hz-20kHz (covering the main frequency bands of industrial and environmental noise), sensitivity of -38dBV / Pa, and signal-to-noise ratio ≥65dB to ensure accurate acquisition of weak noise signals.

[0102] Array calibration: Built-in calibration sound source (1kHz sine wave, 94dB sound pressure level) automatically completes gain and phase calibration of each channel every time it is started, with a calibration error of ≤0.5dB, avoiding measurement deviations caused by differences in microphone performance.

[0103] A2. Miniature acoustic camera module: Core components: Employs a 1 / 2.3-inch CMOS image sensor (1920×1080 resolution) + a 128-channel acoustic imaging chip (such as Sound Scape SS128) to achieve synchronous image and sound acquisition, with an audio-visual fusion frame rate of ≥25fps.

[0104] Visualization function: Supports real-time sound pressure level color mapping (red - high noise, blue - low noise), which can be superimposed on optical images to generate acoustic images, intuitively displaying the spatial distribution of noise sources. The smallest identifiable noise source size is ≤5cm (at a distance of 10m).

[0105] Through a distributed measurement system, 12 core features are extracted from three dimensions—time domain, frequency domain, and time-frequency domain—to provide a basis for noise source type identification. Temporal characteristics: peak sound pressure level, root mean square sound pressure level, pulse width, rise time (reflecting noise intensity and dynamic characteristics).

[0106] Frequency domain characteristics: center frequency, frequency band energy ratio, peak frequency, and spectral flatness (to distinguish between mechanical noise, airflow noise, and other types).

[0107] Time-frequency domain features: wavelet packet energy entropy, short-time fourier transform (STFT) peak, and Hilbert-Huang transform (HHT) marginal spectrum (capturing the instantaneous features of non-stationary noise).

[0108] Through the above processing, spatial localization and feature extraction of noise sources are achieved, improving the accuracy of noise source identification. After the distributed measurement system is deployed, comprehensive wind turbine operating noise characteristics are obtained by changing the wind turbine's operating conditions, mainly including the following conditions: A1. Rated Power Operating Condition Test: Under rated power operating conditions, study the spectral characteristics of aerodynamic noise (such as blade turbulence noise) and mechanical noise (such as gearbox and generator noise), and analyze their interaction mechanism.

[0109] A2. Low wind speed start-up test: Analyze the influence of blade dynamic stall phenomenon on aerodynamic noise during low wind speed start-up, and construct a correlation model between stall state and noise spectrum.

[0110] A3. High wind speed pitch control test: This study investigates the modulation effect of pitch angle adjustment on the noise spectrum under high wind speed and quantifies the relationship between pitch control and noise characteristics.

[0111] Finally, the actual noise characteristic database was constructed: multi-condition test data were integrated to build an actual noise characteristic database covering aerodynamic noise, mechanical noise and modulation effect, providing data support for wind turbine noise prediction and noise reduction design.

[0112] When predicting wind turbine noise, a noise prediction model is constructed, and high-precision noise prediction is achieved through feature engineering and multi-algorithm fusion. Specific technical solutions include: 1. Feature engineering processing: The feature importance analysis method is used to extract key feature parameters related to noise intensity from the wind turbine operating parameters, including but not limited to operating status parameters such as wind speed, wind direction, power, speed, and pitch angle.

[0113] 2. Random Forest Model Construction: Based on the random forest algorithm, a quantitative relationship model between noise intensity and key parameters such as wind speed and wind direction is constructed, and the weight of each parameter on the impact of noise is determined by ranking the importance of features.

[0114] 3. Deep Neural Network Model: Construct a deep neural network model to handle the nonlinear noise prediction problem under the coupling of multiple parameters such as power and speed, and use dropout and regularization techniques to improve the model's generalization ability.

[0115] 4. Online learning algorithm: Develop an online model update mechanism to continuously optimize model parameters through real-time noise monitoring data, enabling the prediction model to adapt to environmental changes.

[0116] 5. Visualization Prediction Platform: Integrate the above models to develop a visualization noise prediction platform, realize an intuitive display of the spatiotemporal distribution of noise, and provide decision support for wind farm noise control.

[0117] By leveraging a distributed data processing architecture based on edge computing technology, a distributed data processing architecture can be constructed to achieve localized feature extraction, anomaly detection, and preliminary analysis of noisy data, thereby reducing data transmission load and improving system response speed.

[0118] Step S304: Based on actual multi-source noise data and wind turbine operation data, construct a multi-objective optimization function with fused constraints, aiming to minimize the difference in scheduling instructions, wind turbine operating load, and noise excess.

[0119] Specifically, step S304 includes: Step S3041: Based on actual multi-source noise data and wind turbine operation data, calculate the deviation between the actual total power of the wind farm and the power commanded by the power grid dispatch as the dispatch command difference, calculate the degree to which the operating load of a single wind turbine exceeds the safety limit as the wind turbine operating load, and calculate the degree to which the noise at the monitoring points around the wind farm exceeds the preset noise standard as the noise exceeding the standard amount.

[0120] Step S3042: With the objectives of minimizing the deviation between the actual total power of the wind farm and the power commanded by the grid dispatch, minimizing the degree to which the operating load of a single wind turbine exceeds the safety limit, and minimizing the degree to which the noise at the monitoring points around the wind farm exceeds the preset noise standard, a multi-objective optimization function with integrated constraints is constructed. The constraints include noise constraints, power constraints, and operational safety constraints.

[0121] Specifically, the multi-objective optimization function uses power grid dispatch instructions, wind turbine operating status parameters, and noise monitoring data as input variables, and is calculated using the following formula: (9); in, This represents the actual total power of the wind farm. Power for power grid dispatch commands. Single fan operating load, For safety limits, For noise from monitoring points around the wind farm, To preset noise standards, To minimize scheduling instruction differences, To minimize the operating load on the wind turbine, To minimize noise overscalar.

[0122] Step S305: Solve the multi-objective optimization function to generate the optimal power allocation scheme for the wind turbine under noise constraints.

[0123] Specifically, step S305 includes: Step a: Under the constraints that the noise at all monitoring points in the target area does not exceed the upper limit of the target area noise, the power of a single wind turbine is within the safe operating range, and the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command, the improved particle swarm optimization algorithm is used to solve the multi-objective optimization function, and the optimal power allocation scheme of wind turbines under noise constraints is generated through iterative optimization.

[0124] It should be noted that before solving the multi-objective optimization function, the noise constraints are processed by receiving environmental noise data collected by the acoustic monitoring device in real time and converting it into constraints that the algorithm can handle.

[0125] When solving multi-objective optimization functions using an improved particle swarm optimization algorithm, an adaptive inertia weight adjustment strategy is introduced to dynamically adjust the search range based on the population convergence state; a dynamic learning factor mechanism is employed to balance the algorithm's global exploration and local exploitation capabilities. The solution process includes: I. Algorithm Initialization: Constructing the Particle Swarm Optimization and Fitness Function: 1. Particle swarm initialization: Using the target power of a single wind turbine X = [x1, x2, ..., x...] n [n] represents the number of particles (where n is the total number of wind turbines), and each particle corresponds to one power allocation scheme. Initialization rules: ①x i Randomly generated in [x min ,x max ];② Satisfy the total power of the entire field ∑x i ≈P ref（ ③ Preliminary filtering of particles with excessive noise (predicted using the power-noise mapping model, removing particles whose predicted LAeq>Lstd). Population size set at 50-100 (balancing solution efficiency and accuracy).

[0126] 2. Fitness Function Definition: The multi-objective optimization function is used as the particle fitness value; the smaller the fitness value, the better the solution. For particles that violate constraints (such as x...), ... i Exceeding [x min ,x max [LAeq>Lstd], with an added penalty term (such as fitness value × 10), to force it to be eliminated by the algorithm.

[0127] II. Iterative Solution: Core Execution of the Improved Particle Swarm Optimization Algorithm: Based on the logic of global exploration, local development, and dynamic adjustment, the particle positions (i.e., power allocation scheme) are iteratively updated. Specific steps include: 1. Initialization of individual and global extreme values: Calculate the fitness value of each particle in the initial population, set the optimal fitness of each particle as the individual extreme value pbest, and set the particle with the minimum fitness in the population as the global extreme value gbest.

[0128] 2. Particle velocity and position update: Update particle velocity v according to the following formula i and position x i (t is the current iteration number): Speed ​​update: v i (t+1)=w v i (t)+c1 r1 (pbest i -x i (t))+c2 r2 (gbest-x i (t)).

[0129] Where w is the adaptive inertia weight: w=0.9 when convergence is slow (expanding the search), w=0.4 when convergence is fast (focusing on the local); c1 / c2 are dynamic learning factors: c1=2.5 / c2=1.5 in the early stage (prioritizing exploration), c1=1.5 / c2=2.5 in the later stage (prioritizing development); r1 / r2 are random numbers in [0,1].

[0130] Location update: x i (t+1)=x i (t)+v i (t+1), and force x i (t+1) is within [xmin, xmax] (if it exceeds, it will be truncated).

[0131] 3. Fitness Value Recalculation and Extreme Value Update: Calculate the updated fitness value for each particle. If the current particle's fitness is less than pbest... i Then update pbest i If the minimum fitness of the population is less than gbest, then update gbest.

[0132] 4. Iteration termination judgment: If the condition "the number of iterations is ≥100 and gbest remains unchanged for 20 consecutive iterations" is met, then the iteration is terminated; otherwise, return to the "particle velocity and position update" step and continue iterating.

[0133] III. Constraint Verification: Feasible solution selection with priority given to noise constraints: After the iteration terminates, the "preliminary optimal power allocation scheme" corresponding to gbest is subjected to multi-dimensional constraint verification to ensure that the scheme conforms to engineering practice: Noise constraint verification: Based on the power-noise mapping model (input target power x of each wind turbine) i ), calculate the theoretical LAeq for each monitoring point: ① If LAeq ≤ Lstd for all monitoring points, proceed to the next step; ② If there is an excess (e.g., LAeq = 57dB(A) for a certain monitoring point), then the power of the corresponding wind turbine (e.g., No. 1, No. 3) is fine-tuned (reduced by 5% each time until LAeq ≤ Lstd). After fine-tuning, the deviation between the total power of the entire field and the dispatch command needs to be re-verified (ΔP ≤ 5%).

[0134] Operational safety constraint verification: Calculate the target power x of each wind turbine i The corresponding theoretical load Fload_target (derived from a model fitting historical power-load data): ① If all Fload_target ≤ Fmax (e.g., 0.85), the scheme is feasible; ② If there is a load exceeding the standard (e.g., Fload_target = 0.9), then reduce the power of the wind turbine (to Fload_target = 0.85), and simultaneously adjust the power of other wind turbines to make up for the total power (ensuring that ΔP remains unchanged).

[0135] Power grid constraint verification: Confirm the total target power ∑x of the entire field. i Deviation from Pref ΔP ≤ 3%: If the deviation exceeds 3%, adjust the power of all fans proportionally (e.g., ∑x). i =Pref×0.95, then all x i (×1.05), after scaling, noise and load constraints need to be checked again.

[0136] IV. Solution Selection: Priority Matching Across Multiple Scenarios If multiple non-dominated solutions (different solutions with similar fitness values) exist after iteration, the final solution is selected based on the priority of the real-time running scenario: Scenario 1: In noise-sensitive scenarios (such as nighttime 22:00-6:00), the solution with "noise overshoot ΔL=0 and minimum LAeq" should be selected first, even if ΔP is slightly higher (≤5%) and the load is slightly higher (≤Fmax).

[0137] Scenario 2: Grid dispatch priority scenario (e.g., peak electricity consumption 18:00-22:00): Prioritize the scheme with "minimum ΔP (≤2%)", noise must meet LAeq≤Lstd, and load≤Fmax.

[0138] Scenario 3: Equipment maintenance priority scenario (such as the high incidence of wind turbine failure): Prioritize the scheme with "minimum average load ΔF (≤0.05)" as long as the noise and scheduling deviation meet the basic constraints.

[0139] V. Solution Outputs: Standardized Results and Implementation Instructions: Output optimal power allocation scheme: Generate structured results: ① Target power table for a single fan (including fan number, x) i ① Theoretical Fload_target and corresponding noise contribution); ② Summary of all indicators (∑x i ① ΔP, average ΔF, and LAeq of each monitoring point; ② Explanation of constraint satisfaction (e.g., LAeq of all monitoring points ≤ 45dB, load ≤ 0.85, ΔP = 2%).

[0140] Execution instruction format conversion: Convert the target power of each wind turbine to x i Convert the command to a format recognizable by the wind turbine controller (such as Modbus protocol message), indicate the execution time (such as immediate execution or execution after 5 minutes, with data transmission time reserved), and set an abnormal rollback mechanism (if the noise exceeds the standard after execution, automatically roll back to the previous feasible solution).

[0141] The dynamic update mechanism is defined as follows: A scheme update cycle is set (e.g., re-solving every 15 minutes), or update conditions are triggered (e.g., noise exceeding the limit by ≥2dB, scheduling command change ≥10%, wind speed sudden change ≥3m / s) to ensure the scheme adapts to the dynamic changes of the wind farm in real time. This involves designing an adaptive sampling control module: employing a dynamic sampling algorithm to automatically adjust the sampling frequency based on the time-varying characteristics of the noise signal (e.g., amplitude change rate, spectral characteristics), optimizing system energy efficiency and extending equipment runtime while ensuring data validity.

[0142] Step S306: Based on the optimal power allocation scheme for wind turbines, perform hierarchical decision control for the wind farm.

[0143] Specifically, step S306 includes: Step S3061: Based on the optimal power allocation scheme for wind turbines, an improved particle swarm optimization algorithm is used to allocate the power of the entire wind farm, and the constraint weights are adjusted in real time according to the environmental noise monitoring data of the wind farm to build a collaborative mechanism for short-term scheduling and long-term optimization at the farm level.

[0144] Specifically, a station-level global optimization control module is constructed, and an improved particle swarm optimization algorithm is used to optimize the power distribution across the entire field through the station-level global optimization control module; a dynamic adjustment strategy for acoustic weight factors is designed, that is, the noise constraint weights are adjusted in real time according to environmental noise monitoring data, and a multi-time-scale optimization mechanism is constructed to maximize operational efficiency through the collaboration of short-term scheduling layer and long-term optimization layer.

[0145] Step S3062: Adaptive PID (Proportional-Integral-Derivative Control) is used to track the power of a single wind turbine. By configuring a noise feedback adjustment unit and interacting in real time with the preset acoustic monitoring system, a multi-level safety protection mechanism at the unit level is constructed.

[0146] Specifically, a unit-level distributed execution module is constructed, and an adaptive PID control algorithm is used through the unit-level distributed execution module to achieve accurate tracking of the power of a single unit; a noise feedback adjustment unit is designed in the unit-level distributed execution module to build a real-time data interaction channel with the upper-level acoustic monitoring system; and a multi-level safety protection mechanism is configured, including over-limit alarm, power limiting and emergency shutdown protection.

[0147] Step S3063: Use a preset communication protocol to perform collaborative optimization control at the site level and unit level, periodically perform optimization calculations and form a closed-loop power control for the wind farm.

[0148] Specifically, a collaborative optimization mechanism is constructed at the station and unit levels: optimization instructions are issued and executed via the OPC (OLE for Process Control) communication protocol; data verification algorithms are used to ensure the reliability of control instruction transmission; and an exception handling process is designed to deal with communication interruptions and other fault situations.

[0149] Construct a closed-loop optimization control system: collect wind turbine operating status parameters and environmental noise data in real time; perform optimization calculations and update control commands periodically; and achieve closed-loop power control of the entire field through the SCADA system.

[0150] It should be noted that this embodiment also provides a wind farm noise and power dynamic control system based on intelligent acoustic monitoring. Specifically, based on acoustic monitoring and wind farm group collaborative control technology, the constructed wind farm power dynamic control system achieves optimized power allocation and energy efficiency improvement within the wind farm group under noise constraints, constructing a closed-loop management system of "monitoring-control-evaluation" to promote intelligent and low-noise operation of the wind farm. This system includes the following modular components: The system comprises a data acquisition and communication module, an energy efficiency management module, an operation optimization control module, and a noise reduction effect evaluation module. It adopts a modular design architecture and utilizes a multi-source real-time data transmission mechanism combining the OPC communication protocol and a 5G communication network to achieve system integration and efficient processing of noise monitoring data. Specifically, the data acquisition and communication module is responsible for acquiring and preprocessing multi-source environmental noise data; the energy efficiency management module dynamically monitors system energy consumption and optimizes energy allocation; the operation optimization control module adaptively adjusts the operating parameters of the field equipment based on real-time noise data; and the noise reduction effect evaluation module quantitatively evaluates noise reduction measures using data analysis algorithms. All modules interact via the OPC communication protocol and rely on the 5G communication network to achieve high-bandwidth, low-latency data transmission, thereby ensuring the overall operating efficiency and noise reduction performance of the system.

[0151] The energy efficiency management module includes: 1. Multi-source heterogeneous data fusion interface: Develop data interface modules that support multiple industrial protocols such as Modbus and IEC61850 to realize the acquisition and standardized fusion of multi-source heterogeneous data from field equipment; 2. Energy efficiency evaluation system and algorithm: Construct an energy efficiency evaluation system for power plant equipment, design energy efficiency assessment algorithms and equipment health status monitoring models, and optimize the overall power generation efficiency; 3. Intelligent optimization control algorithm: Based on deep reinforcement learning (DRL), a multi-objective optimization control algorithm under noise constraints is developed to achieve a dynamic balance between power generation efficiency and noise suppression; 4. Hierarchical decision control: A hierarchical decision architecture is adopted, which combines global optimization at the upper level with individual unit adjustment at the lower level. The upper level formulates optimization strategies based on global energy efficiency targets, while the lower level executes dynamic parameter adjustments for individual units to ensure maximum power generation efficiency under noise constraints.

[0152] The energy efficiency management module achieves efficient, low-noise, and intelligent management of the station operation through the synergistic effect of multi-source data fusion, energy efficiency assessment, and intelligent optimization control.

[0153] This embodiment also provides an effect evaluation system for intelligent optimization of wind farms based on noise constraints, including: 1. Operation Optimization Control Module: Based on the spatiotemporal distribution characteristics of noise and grid load demand of the above research results, a dynamic weight optimization algorithm is designed. According to different operating needs, an intelligent switching system supporting noise reduction priority mode, energy efficiency priority mode and hybrid mode is designed to improve the dynamic adaptability of field group control. 2. Quantitative evaluation system for noise reduction effect: Construct a multi-dimensional noise reduction effect evaluation model that includes sound pressure level attenuation, spectral characteristic improvement and economic indicators to achieve real-time quantitative evaluation of noise reduction measures; 3. Visual monitoring platform: Develop data visualization and interactive tools to dynamically display the analysis results of noise distribution, power generation efficiency and system operating status.

[0154] The wind farm noise control and power optimization method provided in this embodiment, verified through actual wind farm deployment, achieves technical indicators such as noise reduction ≥4dB, power generation efficiency improvement ≥3%, control response time ≤500ms, and system availability ≥99.9%, providing a standardized solution for intelligent wind farm operation. By combining dynamic optimization control with multi-dimensional effect evaluation, it balances environmental protection and noise reduction with power generation benefits, promoting the intelligent upgrading and large-scale application of the wind power industry. This embodiment also provides a wind farm noise control and power optimization device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0155] This embodiment provides a wind farm noise control and power optimization device, such as Figure 7 As shown, it includes: The multi-source noise data acquisition and feature extraction module 701 is used to acquire multi-source noise data from wind farms, analyze the characteristics of multi-source noise data, and construct a multi-source noise feature database.

[0156] The noise propagation theory model construction module 702 is used to construct a noise propagation theory model that includes topographic factors, vegetation factors, obstacle factors and meteorological factors based on a multi-source noise feature database.

[0157] The monitoring point deployment and data acquisition module 703 is used to determine multiple monitoring points in the wind farm based on the noise propagation theory model, and to collect the actual multi-source noise data and wind turbine operation data of multiple monitoring points in real time.

[0158] The multi-objective optimization function construction module 704 is used to construct a multi-objective optimization function with fused constraints based on actual multi-source noise data and wind turbine operation data, with the objectives of minimizing scheduling command differences, wind turbine operating load and noise overshoot.

[0159] The function solving module 705 is used to solve the multi-objective optimization function and generate the optimal power allocation scheme for wind turbines under noise constraints.

[0160] In some optional implementations, the multi-source noise data includes near-field transient flow field data of the wind turbine blades, mechanical noise of the wind turbine gearbox, and electromagnetic noise data; the multi-source noise data acquisition and feature extraction module 701 includes: The aerodynamic noise feature extraction unit is used to obtain near-field transient flow field data of wind turbine blades using CFD simulation, and to solve sound propagation by combining the FW-H equation and computational aeroacoustics, analyze the spectral characteristics of turbulent boundary layer noise, trailing edge noise and tip vortex noise, and obtain aerodynamic noise features.

[0161] The mechanical noise feature extraction unit is used to analyze the generation mechanism of mechanical noise in the wind turbine gearbox by using acoustic emission detection technology and vibration signal analysis, and to obtain mechanical noise features.

[0162] The electromagnetic noise characteristic unit is used to analyze the generation mechanism of electromagnetic noise of generator sets based on electromagnetic field theory and acoustic measurement, and to obtain electromagnetic noise characteristics.

[0163] The multi-source noise feature database construction unit is used to quantitatively evaluate the sound field contribution of each noise source based on aerodynamic noise features, mechanical noise features, and electromagnetic noise features, through coherent analysis and sound source separation technology, and to construct a multi-source noise feature database.

[0164] In some optional implementations, the noise propagation theory model building module 702 includes: The sound propagation model construction unit is used to construct a sound propagation model based on a multi-source noise feature database, using the ray tracing method and the parabolic equation method.

[0165] The vegetation noise reduction model building unit is used to quantitatively analyze the influence of terrain feature parameters on sound wave diffraction and reflection, build a quantitative relationship between terrain feature parameters and sound attenuation rate, and build a vegetation noise reduction model based on the quantitative relationship.

[0166] The meteorological coupling model building unit is used to analyze the shielding effect of the geometric parameters of obstacles on sound wave propagation, and to build a meteorological coupling model based on the shielding effect and meteorological factors.

[0167] The noise propagation theory model building unit is used to construct a noise propagation theory model based on the sound propagation model, vegetation noise reduction model and meteorological coupling model.

[0168] In some optional implementations, the monitoring point deployment and data acquisition module 703 includes: The monitoring point candidate region determination unit is used to determine the monitoring point candidate region based on the noise propagation theory model. The monitoring point candidate region includes the noise source strong radiation area, the key propagation path, the sensitive protection area, and the model verification blind zone.

[0169] The optimal monitoring point set acquisition unit is used to iteratively optimize the number and location of monitoring points within the candidate region using an improved genetic algorithm to obtain the optimal monitoring point set.

[0170] The distributed monitoring point deployment and data acquisition unit is used to deploy distributed monitoring points in the wind farm based on the optimal set of monitoring points, and to collect real-time multi-source noise data and wind turbine operation data from multiple monitoring points according to a dynamic sampling strategy.

[0171] In some optional implementations, the multi-objective optimization function construction module 704 includes: The multi-objective calculation unit is used to calculate the deviation between the actual total power of the wind farm and the power commanded by the grid dispatch as the dispatch command difference, the degree to which the operating load of a single wind turbine exceeds the safety limit as the wind turbine operating load, and the degree to which the noise at the monitoring points around the wind farm exceeds the preset noise standard as the noise exceedance amount, based on actual multi-source noise data and wind turbine operation data.

[0172] The multi-objective optimization function construction unit is used to construct a multi-objective optimization function with the objectives of minimizing the deviation between the actual total power of the wind farm and the power commanded by the grid dispatch, minimizing the degree to which the operating load of a single wind turbine exceeds the safety limit, and minimizing the degree to which the noise of the monitoring points around the wind farm exceeds the preset noise standard. The constraints include noise constraints, power constraints, and operation safety constraints.

[0173] In some alternative implementations, the multi-objective optimization function is calculated using the following formula: ; in, This represents the actual total power of the wind farm. For power grid dispatch command power, Single fan operating load, For safety limits, For noise from monitoring points around the wind farm, This is a preset noise standard.

[0174] In some optional implementations, the noise constraint is that the noise at all monitoring points in the target area does not exceed the upper limit of the target area noise; the power constraint is that the power of a single wind turbine is within the safe operating range; and the operational safety constraint is that the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command. The function solving module 705 includes: The solution unit is used to solve the multi-objective optimization function using an improved particle swarm optimization algorithm under the constraints that the noise at all monitoring points in the target area does not exceed the upper limit of the target area noise, the power of a single wind turbine is within the safe operating range, and the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command. It generates the optimal power allocation scheme for wind turbines under noise constraints through iterative optimization.

[0175] In some alternative implementations, the wind farm noise control and power optimization device further includes: The hierarchical decision control module is used to perform hierarchical decision control of wind farms based on the optimal power allocation scheme of wind turbines.

[0176] In some optional implementations, the hierarchical decision control module includes: The wind farm control unit is used to allocate power across the entire wind farm based on the optimal power allocation scheme for wind turbines, employing an improved particle swarm optimization algorithm. It also adjusts the constraint weights in real time based on environmental noise monitoring data of the wind farm, thus constructing a collaborative mechanism for short-term scheduling and long-term optimization at the wind farm level.

[0177] The unit control unit is used to track the power of a single wind turbine using adaptive PID control. It interacts in real time with the preset acoustic monitoring system through a noise feedback adjustment unit to build a multi-level safety protection mechanism at the unit level.

[0178] The collaborative optimization unit is used to perform collaborative optimization control at the site level and unit level using a preset communication protocol, periodically perform optimization calculations, and form a closed-loop power control for the wind farm.

[0179] The wind farm noise control and power optimization device provided in this embodiment of the invention can execute the wind farm noise control and power optimization method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0180] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0181] The following is a detailed reference. Figure 8 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0182] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0183] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the wind farm noise control and power optimization method of the embodiments of the present invention.

[0184] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0185] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the wind farm noise control and power optimization method shown in the above embodiments is implemented.

[0186] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0187] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for noise control and power optimization in wind farms, characterized in that, The method includes: Acquire multi-source noise data from wind farms, analyze the characteristics of the multi-source noise data, and construct a multi-source noise feature database; Based on the multi-source noise feature database, a noise propagation theoretical model including terrain factors, vegetation factors, obstacle factors and meteorological factors is constructed. Based on the aforementioned noise propagation theoretical model, multiple monitoring points in the wind farm were determined, and real-time multi-source noise data and wind turbine operation data from these monitoring points were collected. Based on the actual multi-source noise data and wind turbine operation data, a multi-objective optimization function with fused constraints is constructed with the objectives of minimizing scheduling command differences, wind turbine operating load and noise exceedance. The multi-objective optimization function is solved to generate the optimal power allocation scheme for the wind turbine under noise constraints.

2. The method according to claim 1, characterized in that, The multi-source noise data includes near-field transient flow field data of wind turbine blades, mechanical noise data of wind turbine gearbox, and electromagnetic noise data. The characteristics of the multi-source noise data are analyzed to construct a multi-source noise feature database, including: CFD simulation was used to obtain near-field transient flow field data of wind turbine blades. The sound propagation was solved by combining the FW-H equation and computational aeroacoustics. The spectral characteristics of turbulent boundary layer noise, trailing edge noise and tip vortex noise were analyzed to obtain aerodynamic noise characteristics. The generation mechanism of mechanical noise in the wind turbine gearbox was analyzed using acoustic emission detection technology and vibration signal analysis, and the characteristics of mechanical noise were obtained. Based on electromagnetic field theory and acoustic measurements, the generation mechanism of electromagnetic noise of generator sets is analyzed, and the characteristics of electromagnetic noise are obtained. Based on the aforementioned aerodynamic noise characteristics, mechanical noise characteristics, and electromagnetic noise characteristics, the sound field contribution of each noise source is quantitatively evaluated through coherent analysis and sound source separation technology, and a multi-source noise characteristic database is constructed.

3. The method according to claim 1, characterized in that, Based on the aforementioned multi-source noise feature database, a noise propagation theoretical model is constructed, incorporating topographic, vegetation, obstacle, and meteorological factors, including: Based on the aforementioned multi-source noise feature database, a sound propagation model is constructed using the ray tracing method and the parabolic equation method. The influence of terrain feature parameters on sound wave diffraction and reflection is quantitatively analyzed, a quantitative relationship between terrain feature parameters and sound attenuation rate is constructed, and a vegetation noise reduction model is constructed based on the quantitative relationship. The shielding effect of the geometric parameters of obstacles on sound wave propagation is analyzed, and a meteorological coupling model is constructed based on the shielding effect and meteorological factors. A theoretical model for noise propagation is constructed based on the aforementioned sound propagation model, vegetation noise reduction model, and meteorological coupling model.

4. The method according to claim 1, characterized in that, Based on the aforementioned noise propagation theoretical model, multiple monitoring points were determined in the wind farm, and real-time multi-source noise data and wind turbine operation data from these monitoring points were collected, including: Based on the aforementioned noise propagation theoretical model, candidate regions for monitoring points are determined, including strong radiation areas of noise sources, critical propagation paths, sensitive protection areas, and model validation blind zones. Within the candidate region, an improved genetic algorithm is used to iteratively optimize the number and location of monitoring points to obtain the optimal set of monitoring points. Based on the optimal set of monitoring points, distributed monitoring points are deployed within the wind farm, and real-time multi-source noise data and wind turbine operation data from multiple monitoring points are collected in accordance with a dynamic sampling strategy.

5. The method according to claim 1, characterized in that, Based on the actual multi-source noise data and wind turbine operation data, a multi-objective optimization function with fused constraints is constructed, aiming to minimize scheduling command differences, wind turbine operating load, and noise exceedance, including: Based on the actual multi-source noise data and wind turbine operation data, the deviation between the actual total power of the wind farm and the power of the grid dispatch command is calculated as the dispatch command difference; the degree to which the operating load of a single wind turbine exceeds the safety limit is calculated as the wind turbine operating load; and the degree to which the noise at the monitoring points around the wind farm exceeds the preset noise standard is calculated as the noise exceeding the standard amount. With the objectives of minimizing the deviation between the actual total power of the wind farm and the power commanded by the grid dispatch, minimizing the degree to which the operating load of a single wind turbine exceeds the safety limit, and minimizing the degree to which the noise at monitoring points around the wind farm exceeds the preset noise standard, a multi-objective optimization function with integrated constraints is constructed. The constraints include noise constraints, power constraints, and operational safety constraints.

6. The method according to claim 5, characterized in that, The multi-objective optimization function is calculated using the following formula: ； in, This represents the actual total power of the wind farm. Power for power grid dispatch commands. Single fan operating load, For safety limits, For noise from monitoring points around the wind farm, This is a preset noise standard.

7. The method according to claim 5, characterized in that, The noise constraint condition is that the noise at all monitoring points in the target area does not exceed the upper limit of the noise in the target area; the power constraint is that the power of a single wind turbine is within the safe operating range; and the operation safety constraint is that the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the power grid dispatch instruction. Solving the multi-objective optimization function to generate the optimal power allocation scheme for wind turbines under noise constraints includes: Under the constraints that the noise at all monitoring points in the target area does not exceed the upper limit of the target area noise, the power of a single wind turbine is within the safe operating range, and the load corresponding to the power of a single wind turbine does not exceed the load corresponding to the grid dispatch command, the improved particle swarm optimization algorithm is used to solve the multi-objective optimization function, and the optimal power allocation scheme of wind turbines under noise constraints is generated through iterative optimization.

8. The method according to claim 1, characterized in that, The method further includes: Based on the optimal power allocation scheme for the wind turbines, hierarchical decision control is implemented for the wind farm.

9. The method according to claim 8, characterized in that, Based on the optimal power allocation scheme for the wind turbines, hierarchical decision control of the wind farm is implemented, including: Based on the optimal power allocation scheme of the wind turbines, an improved particle swarm optimization algorithm is used to allocate the power of the entire wind farm, and the weight of the constraint conditions is adjusted in real time according to the environmental noise monitoring data of the wind farm to build a collaborative mechanism for short-term scheduling and long-term optimization at the farm level. Adaptive PID control is used to track the power of a single wind turbine. By configuring a noise feedback adjustment unit and interacting in real time with a preset acoustic monitoring system, a multi-level safety protection mechanism at the unit level is constructed. A preset communication protocol is used for collaborative optimization control at the site and unit levels, and optimization calculations are performed periodically to form a closed-loop power control for the wind farm.

10. A wind farm noise control and power optimization device, characterized in that, The device includes: The multi-source noise data acquisition and feature extraction module is used to acquire multi-source noise data from the wind farm, analyze the characteristics of the multi-source noise data, and construct a multi-source noise feature database. The noise propagation theory model construction module is used to construct a noise propagation theory model that includes topographic factors, vegetation factors, obstacle factors and meteorological factors based on the multi-source noise feature database. The monitoring point deployment and data acquisition module is used to determine multiple monitoring points in the wind farm based on the noise propagation theoretical model, and to collect the actual multi-source noise data and wind turbine operation data of multiple monitoring points in real time. The multi-objective optimization function construction module is used to construct a multi-objective optimization function with fused constraints based on the actual multi-source noise data and wind turbine operation data, with the objectives of minimizing scheduling command differences, wind turbine operating load and noise overshoot. The function solving module is used to solve the multi-objective optimization function and generate the optimal power allocation scheme for the wind turbine under noise constraints.

Images