A method and system for noise assessment simulation of a centrifugal blower

Abstract

This invention discloses a noise assessment simulation method and system for centrifugal blowers. The method specifically includes: calculating mechanical vibration noise using a multibody dynamics model and coupling it with electromagnetic excitation, while simultaneously performing unsteady CFD simulation to extract aerodynamic sound sources; based on the vibration and aerodynamic sound source data, generating an acoustic mesh for the volute structure using BEM in an acoustic solver and applying AML technology to form an acoustic simulation environment; within the acoustic simulation environment, mapping the vibration sound source data to equivalent body forces and performing phase correction on the aerodynamic sound source data, outputting the sound pressure level spectrum; based on the sound pressure level spectrum, embedding a silencer structure into the flow field model, correlating transmission loss with the overall machine sound power level, and forming a visual evaluation model for noise reduction effectiveness. This invention achieves accurate simulation of the entire process of centrifugal blower noise generation, assessment, noise reduction, and dynamic characteristic analysis, providing a comprehensive and reliable basis for noise control and optimization design.

Description

Technical Field

[0001] This invention relates to the field of centrifugal blower technology, and in particular to a noise assessment simulation method and system for centrifugal blowers. Background Technology

[0002] Noise assessment and control are crucial aspects of the design and operation of centrifugal blowers. With increasingly stringent requirements for equipment operating environments in industry, accurately and comprehensively assessing the noise characteristics of centrifugal blowers and implementing effective noise reduction measures has become a critical issue that urgently needs to be addressed. However, existing noise assessment simulation methods have many shortcomings, severely hindering the development of centrifugal blower noise control technology.

[0003] First, traditional noise assessment methods typically treat vibration noise and aerodynamic noise as independent factors and predict them separately. However, in actual operation, there is a complex and strong coupling effect between mechanical vibration and airflow excitation. For example, mechanical vibrations generated by gear meshing and bearing oil film oscillations are transmitted through the unit structure and excite surrounding air vibrations, generating vibration noise. Simultaneously, airflow pulsations caused by impeller rotation and turbulent motion within the volute generate aerodynamic noise, which in turn further affects the unit's vibration characteristics. This mutual influence leads to significant deviations between the predicted results of a single noise source and the actual situation, failing to accurately reflect the true noise generation mechanism of the centrifugal blower and thus affecting the accuracy of noise assessment.

[0004] Secondly, the casing structure of centrifugal blowers, such as the volute, is typically irregular in shape, resulting in an extremely complex internal sound field distribution. When sound waves propagate within this irregular casing, various physical phenomena occur, including reflection, scattering, and diffraction. These effects significantly influence the sound field distribution and propagation characteristics. However, existing simulation methods often struggle to fully quantify these acoustic effects when dealing with complex casing sound fields, leading to significant discrepancies between simulation results and actual sound field distributions. For example, when simulating the sound field inside the volute, traditional methods cannot accurately predict the reflection and scattering of sound waves at different curved surfaces, thus failing to precisely calculate the sound pressure level at each location and affecting the assessment of centrifugal blower radiated noise.

[0005] Finally, silencers are commonly used noise reduction components in the noise reduction design of centrifugal blowers. However, existing silencer performance evaluation methods are often independent of the overall machine noise evaluation system, resulting in a lack of effective correlation between silencer design and the overall acoustic performance of the machine. Under this fragmented design model, the noise reduction effect of the silencer may not match the acoustic characteristics of the entire machine, failing to achieve optimal noise reduction. For example, the silencer design may only optimize for noise in a specific frequency band, ignoring the contribution of that frequency band to the overall machine noise and its interaction with other frequency bands, thus resulting in an unsatisfactory actual noise reduction effect of the silencer.

[0006] Furthermore, most existing simulation models for centrifugal blower noise assessment are based on steady-state operating conditions and cannot accurately reflect the impact of dynamic processes such as variable speed and load on noise characteristics. In actual operation, the speed and load of centrifugal blowers often adjust continuously with changes in operating conditions. This dynamic change leads to significant alterations in the characteristics of the noise source and the sound field distribution. For example, when the speed changes, the impeller's rotational frequency and aerodynamic excitation frequency also change accordingly, thus affecting the generation of aerodynamic noise. Simultaneously, changes in load alter the unit's vibration characteristics, thereby affecting the intensity and frequency distribution of vibration noise. Because traditional simulation models cannot adapt to these dynamic operating conditions, they cannot accurately predict the noise characteristics of centrifugal blowers under different operating conditions, limiting the practicality and accuracy of noise assessment. Summary of the Invention

[0007] The purpose of this invention is to provide a noise assessment simulation method and system for centrifugal blowers, which realizes accurate simulation of the entire process of centrifugal blower noise generation, assessment, noise reduction and dynamic characteristic analysis, providing a comprehensive and reliable basis for noise control and optimization design, and solving at least one of the above-mentioned problems in the prior art.

[0008] In a first aspect, the present invention provides a noise assessment simulation method for a centrifugal blower, the method specifically comprising:

[0009] Mechanical vibration noise is calculated by multibody dynamics model and coupled with electromagnetic excitation. At the same time, unsteady CFD simulation is performed to extract aerodynamic sound sources. The vibration sound source data and aerodynamic sound source data are transmitted to the acoustic solver via real-time interface.

[0010] Based on vibration and aerodynamic sound source data, BEM is used to generate an acoustic mesh for the volute structure in the acoustic solver and AML technology is applied. At the same time, the total reflection boundary of the pipe and the AML surface of the outlet are configured to form an acoustic simulation environment.

[0011] In the acoustic simulation environment, vibration sound source data is mapped to equivalent body forces and aerodynamic sound source data is phase corrected. The wave equation is solved by FEM to output the sound pressure level spectrum.

[0012] Based on the sound pressure level spectrum, a silencer structure is embedded in the flow field model. The characteristics of the silencer element are characterized by the impedance boundary. The transmission loss is correlated with the sound power level of the whole machine to form a visual evaluation model of the noise reduction effect.

[0013] A Simulink dynamic simulation platform was built, which calls the acoustic solver in real time in response to changes in rotational speed and load, and links the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

[0014] Secondly, the present invention provides a noise assessment simulation system for a centrifugal blower, the system specifically comprising:

[0015] The first evaluation simulation module is used to calculate mechanical vibration noise and couple electromagnetic excitation through a multibody dynamics model, while performing unsteady CFD simulation to extract aerodynamic sound sources and transmitting the vibration sound source data and aerodynamic sound source data to the acoustic solver via a real-time interface.

[0016] The second evaluation simulation module is used to generate an acoustic mesh for the volute structure using BEM and apply AML technology in the acoustic solver based on vibration and aerodynamic sound source data. At the same time, it configures the pipe total reflection boundary and the outlet AML surface to form an acoustic simulation environment.

[0017] The third evaluation simulation module is used to map vibration sound source data into equivalent body forces and perform phase correction on aerodynamic sound source data in an acoustic simulation environment, and output the sound pressure level spectrum by solving the wave equation through FEM.

[0018] The fourth evaluation simulation module is used to embed the silencer structure into the flow field model based on the sound pressure level spectrum, characterize the characteristics of the silencer element through the impedance boundary, correlate the transmission loss with the sound power level of the whole machine, and form a visual evaluation model of the noise reduction effect.

[0019] The fifth evaluation simulation module is used to build a Simulink dynamic simulation platform. It calls the acoustic solver in real time in response to changes in rotational speed and load, and links with the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

[0020] Thirdly, the present invention provides a computer device comprising: a memory and a processor, and a computer program stored in the memory, wherein when the computer program is executed on the processor, it implements the noise assessment simulation method for a centrifugal blower as described in any of the above methods.

[0021] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the noise assessment simulation method for a centrifugal blower as described in any of the above methods.

[0022] Compared with the prior art, the present invention has at least one of the following technical effects:

[0023] 1. This invention realizes accurate simulation of the entire process of centrifugal blower from multi-source noise generation to evaluation, noise reduction and dynamic characteristic analysis, providing a comprehensive and reliable basis for noise control and optimization design.

[0024] 2. This invention fully considers the coupling relationship between vibration noise and aerodynamic noise, and can more accurately simulate the actual noise generation process of centrifugal blowers, thereby improving the accuracy of noise source prediction.

[0025] 3. This invention can more accurately simulate the sound wave reflection, scattering and diffraction effects inside irregular casings, and fully considers the influence of the volute structure on the sound field distribution, thereby improving the accuracy of sound field simulation of complex casings and providing a reliable basis for accurately evaluating the radiated noise of centrifugal blowers.

[0026] 4. This invention organically combines muffler performance evaluation with overall machine noise evaluation, which can comprehensively consider the impact of the muffler on the acoustic performance of the whole machine, achieve optimized matching between muffler design and overall machine acoustic performance, and improve the effectiveness of noise reduction measures.

[0027] 5. This invention can simulate the operation of a centrifugal blower under dynamic conditions such as variable speed and variable load, and reflect the impact of the dynamic process on noise characteristics in real time. It improves the dynamic adaptability and accuracy of noise assessment and provides strong support for noise control of centrifugal blowers under different operating conditions. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic flowchart of a noise assessment simulation method for a centrifugal blower provided in an embodiment of the present invention;

[0030] Figure 2 This is a schematic diagram of the structure of a noise assessment simulation system for a centrifugal blower provided in an embodiment of the present invention;

[0031] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0032] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0033] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0034] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0035] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0036] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0037] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0038] In this application embodiment, the entity executing the process includes a terminal device. This terminal device includes, but is not limited to, devices capable of executing the methods disclosed in this application, such as servers, computers, smartphones, and tablets. Figure 1 A flowchart illustrating a noise assessment simulation method for a centrifugal blower according to an embodiment of the present invention is shown below in detail:

[0039] S101 calculates mechanical vibration noise and couples it with electromagnetic excitation through a multibody dynamics model. At the same time, it performs unsteady CFD simulation to extract aerodynamic sound sources and transmits the vibration sound source data and aerodynamic sound source data to the acoustic solver via a real-time interface.

[0040] In this embodiment, the aim is to address the limitations of independent prediction of mechanical vibration and aerodynamic noise in existing centrifugal blower noise assessments. By using a multibody dynamics model and unsteady CFD simulation for collaborative calculation, the coupled analysis of vibration noise and aerodynamic noise is achieved. Furthermore, the data from both types of sound sources are transmitted to the acoustic solver via a real-time interface, providing high-precision input for subsequent acoustic simulations.

[0041] Specifically, based on the three-dimensional geometric model of the centrifugal blower, a simplified multibody system is established in multibody dynamics software (such as SIMPACK or RecurDyn). Electromagnetic torque excitation is applied to the motor drive end, and torque fluctuation curves are obtained through experimental data or electromagnetic simulations (such as Maxwell), inputting them into the multibody model in time series form. Nonlinear excitations such as gear meshing and bearing oil film oscillations are simulated, and dynamic loads are calculated using contact force models (such as the Lankarani-Nikravesh model). A flow field computational domain including the impeller, volute, and inlet is established in CFD software (such as Fluent or Star-CCM+). Sliding mesh or dynamic mesh technology is used to handle impeller rotation, with a hexahedral-dominated unstructured mesh type and near-wall refinement to meet y+ requirements. A turbulence model (such as the SST k-ω model) is selected to capture the turbulent motion within the volute. The time step is set to 1 / 50 to 1 / 100 of the impeller rotation period, with a total physical time covering 3 to 5 impeller rotation periods to eliminate transient effects. The flow field variables (such as pressure and velocity) for each time step are output to a transient data file. Based on Lighthill acoustic analogy theory, dipole sound sources (pressure fluctuations on the impeller surface) and quadrupole sound sources (turbulence within the volute) are extracted from the flow field using the FW-H equation or the generalized Lighthill equation. The sound source data is spatially integrated and time-averaged to generate an aerodynamic sound source data file containing parameters such as sound power level and directivity distribution. Data export modules are written in both multibody dynamics and CFD software to convert the vibration sound source data (frequency-amplitude) and aerodynamic sound source data (spatial coordinates-sound power level) into a unified format (such as HDF5 or XML). Middleware programs (such as Python scripts) are developed to read the two types of data files and parse timestamps or frequency tags to ensure data synchronization. Establish a communication link between the multibody software, CFD software, and acoustic solver (such as VA One or Actran) via TCP / IP protocol or shared memory; verify the data before transmission, for example, check whether the vibration spectrum and the main frequency of aerodynamic noise overlap, if there is a difference, trigger a warning and adjust the simulation parameters (such as time step and mesh density); after transmission, generate a log file in the acoustic solver to record the data reception time, data volume, and integrity status.

[0042] S102, based on vibration and aerodynamic sound source data, uses BEM to generate an acoustic mesh for the volute structure in the acoustic solver and applies AML technology. At the same time, it configures the pipe total reflection boundary and the outlet AML surface to form an acoustic simulation environment.

[0043] In this embodiment, addressing the challenges of accurately simulating the complex structure and sound field distribution of centrifugal blower volutes, a method for acoustic mesh generation and boundary condition configuration combining the boundary element method (BEM) and automatic matching layer (AML) techniques is proposed. By fusing data from mechanical vibration and aerodynamic sound sources, a high-precision acoustic simulation environment is constructed, resolving simulation errors caused by insufficient quantization of sound wave reflection / scattering and unreasonable boundary condition settings in traditional methods. This provides a reliable foundation for subsequent sound pressure level spectrum calculations.

[0044] Specifically, import the 3D CAD model of the centrifugal blower casing (e.g., STEP or IGES format), and perform geometric cleanup in acoustic preprocessing software (e.g., VA One or Actran) to remove minor features such as chamfers and bolt holes. Perform stitching on discontinuous surfaces (e.g., volute tongue, flange connection surfaces) to ensure the model's watertightness. Extract the inner surface of the casing as the acoustic analysis boundary, ignoring the direct influence of external structures (e.g., motor, base) on the internal sound field. Discretize the inner surface of the casing using the boundary element method, selecting a quadrilateral-dominated unstructured mesh type. The element size must meet the acoustic wavelength requirements (typically 1 / 6 to 1 / 10 of the wavelength corresponding to the highest analysis frequency). Locally refine the sound source area (e.g., near the impeller outlet, volute tongue) with a refinement factor of 2 to 3 times to capture high-frequency sound wave propagation details. Generate a mesh quality check report to ensure element skewness < 0.7 and aspect ratio < 5, avoiding distorted elements from affecting computational convergence. AML technology automatically generates a gradient impedance layer to simulate infinite sound field conditions, avoiding the drawbacks of traditional PML (Perfect Matching Layer) which requires manual parameter setting. For the open boundary of the volute outlet, activate the AML option in the acoustic solver, select the "Free Field Radiation" mode, and set the matching layer thickness to 1.5 times the wavelength corresponding to the highest frequency. Set a total reflection boundary for the centrifugal blower intake duct (such as a straight or curved pipe) to simulate the infinite reflection of sound waves within the duct. Select "HardWall" in the acoustic solver to force the sound pressure normal gradient to zero, equivalent to reflection from a rigid wall. If there is sound-absorbing material (such as glass wool) inside the duct, obtain the material's sound absorption coefficient experimentally and replace it with an "Impedance Boundary," inputting the complex acoustic impedance value (e.g., Z=ρc(1+0.1j), where ρc is the characteristic impedance of air). Vibration source data (such as vibration acceleration on the bearing housing surface) output from the multibody dynamics model are mapped to the mesh nodes of the volute structure as equivalent body force (unit volume force). Aerodynamic source data (such as impeller surface dipole sound source) extracted by CFD are applied to the inner surface of the volute in a spatially distributed form. The temporal characteristics of synchronous vibration and aerodynamic sound source are corrected by phase correction. Unit sound pressure excitation (1 Pa) is applied at the pipe inlet to verify the effect of total reflection boundary on the preservation of sound wave energy and ensure that the deviation of the inlet sound pressure level from the theoretical value is <0.5dB.

[0045] S103, in the acoustic simulation environment, maps the vibration sound source data to equivalent body forces and performs phase correction on the aerodynamic sound source data, and outputs the sound pressure level spectrum by solving the wave equation through FEM.

[0046] In this embodiment, by converting vibration data into equivalent body force, correcting the phase of the aerodynamic sound source, and using the finite element method (FEM) to solve the wave equation, high-precision sound pressure level spectrum calculation is achieved, solving the problem of spectrum distortion caused by neglecting coupling effects in traditional methods.

[0047] Specifically, the surface vibration acceleration time-domain data (unit: m / s²) of key structures such as the volute and bearing housing are exported from the multibody dynamics model. The data sampling frequency needs to cover more than twice the upper limit of the analysis frequency band (e.g., when analyzing up to 2000Hz, the sampling frequency should be ≥4000Hz). The vibration data is then transformed in the frequency domain (e.g., Fast Fourier Transform, FFT) to obtain the vibration acceleration amplitude and phase information at each frequency component, and a spectrum file is generated. According to the acoustic finite element theory, the surface vibration acceleration of the structure is converted into an equivalent body force (unit: N / m³), and the calculation formula is: equivalent body force = ρ0·a(t), where ρ0 is the air density (1.225 kg / m³) and a(t) is the time-domain signal of the surface vibration acceleration. In the acoustic solver, the BEM mesh model of the volute structure is imported, and the equivalent body force is loaded onto the mesh nodes in the form of "volume force". The loading method is selected as "frequency domain harmonic loading", and the amplitude and phase of each frequency component are input. Local densification loading is performed on areas with strong vibration (such as near the bearing seat), and the densification factor is set to 2 to 3 times to ensure that the high-frequency vibration energy is accurately transferred to the sound field.

[0048] Data from the impeller surface dipole sound source and the volute inner wall quadrupole sound source are extracted from unsteady CFD simulations and output as sound pressure level spectrum files (in dB) and phase spectrum files (in degrees). The aerodynamic sound source data are phase-calibrated to eliminate the phase delay caused by the time step setting in the CFD calculation (e.g., by comparing the phase difference between the impeller rotation frequency and the main frequency of the sound source and adjusting the phase spectrum offset). In the acoustic solver, the aerodynamic sound source data is converted into sound pressure amplitude (unit: Pa), and the calculation formula is: sound pressure amplitude = 10^(SPL / 20), where SPL is the sound pressure level (dB); according to the phase spectrum file, phase angles are assigned to each frequency component to generate a complex form of sound pressure signal (real part = amplitude × cos(phase), imaginary part = amplitude × sin(phase)); the corrected aerodynamic sound source is loaded onto the inner wall mesh node of the volute in the form of "surface sound pressure", and the loading method is selected as "frequency domain complex loading" to ensure that aerodynamic noise and vibration noise are synchronously coupled in the frequency domain.

[0049] In the acoustic solver, a three-dimensional acoustic finite element model is generated based on the BEM mesh of the volute structure. The mesh type is selected as a hex-dominant hybrid mesh, and the element size must meet the acoustic wavelength requirements (e.g., the highest analysis frequency of 2000Hz corresponds to a wavelength of 0.17m, and the mesh size is ≤0.017m). The sound source area (e.g., impeller outlet, volute tongue) is locally refined with a refinement factor of 1.5 to capture the details of high-frequency sound wave propagation. The model boundary conditions are set as follows: the intake pipe is set as a "total reflection boundary" (simulating infinite sound wave reflection), and the outlet is set as an "AML automatic matching layer" (simulating an infinite free field). Activate the "Frequency Domain Steady-State Analysis" module in the acoustic solver, select the "Helmholtz Wave Equation" as the governing equation, and input the air velocity (343 m / s) and density (1.225 kg / m³). Set the analysis frequency band (e.g., 50~2000Hz), and the frequency step size is 1 / 3 octave (e.g., 100Hz, 125Hz, 160Hz…) to ensure that the spectral resolution meets the requirements for A-weighted sound level evaluation. Start the solver to calculate the sound pressure distribution at each frequency, output the nodal sound pressure amplitude and phase data, and generate a sound pressure level spectrum file (unit: dB). The calculation formula is: SPL = 20·log10(p / p0), where SPL is the sound pressure level, p is the calculated sound pressure (Pa), and p0 is the reference sound pressure (2×10⁻⁻¹). 5 Pa).

[0050] S104, based on the sound pressure level spectrum, embeds a silencer structure into the flow field model, characterizes the silencer element characteristics through impedance boundaries, and correlates the transmission loss with the overall sound power level to form a visual evaluation model for noise reduction effect.

[0051] In this embodiment, by embedding a muffler structure in the flow field model, the characteristics of the muffler element are simulated using impedance boundaries, and the transmission loss is correlated with the overall sound power level, so as to achieve a visualized and quantitative evaluation of the noise reduction effect and solve the problem of low noise reduction efficiency caused by the fragmented design of traditional methods.

[0052] Specifically, based on the three-dimensional geometric model of the centrifugal blower (such as the volute, impeller, and intake duct), an unstructured mesh is generated in CFD software (such as ANSYS Fluent or Star-CCM+). The mesh type is a hex-dominant hybrid mesh, and the element size must meet the acoustic wavelength requirements (e.g., for a wavelength of 0.17m corresponding to the highest analysis frequency of 2000Hz, the mesh size should be ≤0.017m). Local densification is performed on key areas (such as the impeller outlet, volute tongue, and muffler installation location) with a densification factor of 1.5 to 2 times to ensure that the details of turbulence and sound wave propagation are accurately captured. The intake duct inlet is set as a "velocity inlet" (the flow velocity is set according to the operating conditions, such as 10m / s), the outlet is set as a "pressure outlet" (gauge pressure is set to 0Pa), and the inner wall of the volute is set as a "non-slip wall" (the roughness is set according to the actual material, such as 0.1mm for steel surfaces). Depending on the type of muffler (e.g., resistive, reactive, or composite), locate the installation position in the flow field model (usually the volute outlet or intake duct). Construct the muffler geometric model parametrically, and fuse the muffler model with the host flow field mesh using "shared nodes" or "overlapping meshes" to ensure fluid continuity and avoid calculation divergence due to mesh mismatch.

[0053] Acoustic monitoring points were set at the inlet and outlet sections of the silencer to extract the incident sound power (Wᵢ) and transmitted sound power (Wᵢ). t ), calculate the propagation loss TL = 10·log10(Wᵢ / W t The transmission loss spectrum is generated by calculating the TL value at each frequency through frequency domain analysis (such as FFT). Based on the previously calculated sound pressure level spectrum of the whole machine (without muffler), the sound power level (L_W0) at each frequency is extracted. According to the transmission loss spectrum of the muffler, the sound power level of the whole machine L_W = L_W0-TL is corrected.

[0054] In the CFD post-processing module, the corrected whole-machine sound power level data is imported to generate a three-dimensional sound field distribution cloud map.

[0055] S105 is used to build a Simulink dynamic simulation platform. It responds to changes in rotational speed and load by calling the acoustic solver in real time and links with the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

[0056] In this embodiment, by calling the acoustic solver in real time and linking it with the surge boundary prediction module, accurate prediction of the time-varying characteristics of noise is achieved, solving the evaluation deviation problem caused by fixed operating conditions in traditional methods, and providing technical support for dynamic noise reduction control.

[0057] In some embodiments, step S101 above, which involves calculating mechanical vibration noise using a multibody dynamics model and coupling it with electromagnetic excitation, while simultaneously performing unsteady CFD simulation to extract aerodynamic sound sources, and transmitting the vibration sound source data and aerodynamic sound source data to the acoustic solver via a real-time interface, specifically includes:

[0058] In the multibody dynamics model, the gear meshing stiffness curve and the bearing oil film damping coefficient are set, and the mechanical vibration response of gear meshing stiffness and bearing oil film oscillation is solved by nonlinear contact algorithm.

[0059] The stator-rotor air gap magnetic flux density distribution was calculated in the electromagnetic finite element model, and the electromagnetic force wave spectrum of the main order was extracted by Fourier transform.

[0060] The electromagnetic force wave spectrum is reconstructed into a time-domain excitation force through inverse Fourier transform and then superimposed onto the mechanical vibration response;

[0061] Parallel execution of unsteady CFD simulations was performed to extract transient pressure pulsation data from the impeller and volute surfaces. Based on acoustic analogy theory, the transient pressure pulsation data was converted into dipole and quadrupole sound sources.

[0062] The mechanical vibration response, dipole sound source, and quadrupole sound source are transmitted to the acoustic solver via a real-time data interface.

[0063] In this embodiment, when constructing the multibody dynamics model of the centrifugal blower, detailed parameter settings are performed for two key components: gear meshing and bearing oil film. For gear meshing, the gear meshing stiffness curve is obtained through experimental measurement or theoretical calculation. This curve describes the stiffness variation of the gear at different meshing positions and is accurately set into the multibody dynamics model. For the bearing oil film, its damping coefficient is determined. This coefficient reflects the energy dissipation characteristics of the bearing oil film during vibration and is also set into the model.

[0064] A nonlinear contact algorithm is employed to solve the multibody dynamics model. This algorithm accurately simulates the nonlinear contact behavior between gear teeth during meshing and the complex mechanical properties of the bearing oil film during vibration. Using this algorithm, the mechanical vibration response caused by gear meshing stiffness and bearing oil film oscillation can be calculated. This mechanical vibration response includes information such as the vibration displacement, velocity, and acceleration of the gears and bearings under different operating conditions, forming the basis for subsequent calculations of mechanical vibration noise.

[0065] An electromagnetic finite element model of a centrifugal blower is constructed. This model should accurately describe the geometry, material properties, and electromagnetic parameters of the stator and rotor. Appropriate boundary conditions and excitation sources are set in the model, and the model is solved using electromagnetic finite element analysis software to calculate the magnetic flux density distribution in the stator-rotor air gap. The air gap magnetic flux density distribution reflects the intensity and direction changes of the electromagnetic field inside the motor and is the root cause of electromagnetic force.

[0066] Fourier transform was performed on the calculated air gap magnetic flux density distribution data. Fourier transform is a mathematical method that converts a time-domain signal into a frequency-domain signal, allowing the complex time-varying signal of the air gap magnetic flux density distribution to be decomposed into a superposition of sinusoidal waves with different frequency components. After the Fourier transform, the spectra of the main orders of electromagnetic force waves were extracted. These main orders of electromagnetic force waves are key factors causing motor vibration and noise; their frequency and amplitude information is crucial for accurately assessing the impact of electromagnetic excitation on mechanical vibration.

[0067] The extracted main-order electromagnetic force wave spectrum is subjected to an inverse Fourier transform. The inverse Fourier transform is the reverse process of the Fourier transform, which converts the frequency domain signal back into a time domain signal. Through the inverse Fourier transform, the electromagnetic force wave spectrum is reconstructed into a time-domain excitation force. This time-domain excitation force reflects the change of electromagnetic force over time and has a clear physical meaning and temporal characteristics. The reconstructed time-domain excitation force is superimposed onto the mechanical vibration response calculated earlier using the multibody dynamics model. This superposition considers the additional influence of electromagnetic excitation on mechanical vibration, making the mechanical vibration response more accurately reflect the vibration of the centrifugal blower under various excitations during actual operation. The superimposed mechanical vibration response will serve as the sound source data for mechanical vibration noise, used in subsequent acoustic simulations.

[0068] Parallel execution of unsteady computational fluid dynamics (CFD) simulations. Unsteady CFD simulations can simulate the transient changes in the flow field inside a centrifugal blower, including airflow pulsations caused by impeller rotation and turbulent motion within the volute. During the simulation, appropriate computational domains, boundary conditions, and initial conditions are set to ensure that the simulation results accurately reflect the dynamic characteristics of the actual flow field.

[0069] In unsteady CFD simulations, transient pressure pulsation data are extracted from the impeller and volute surfaces. These transient pressure pulsations are the root cause of aerodynamic noise, reflecting the unstable flow near the impeller and volute surfaces. By appropriately setting monitoring points in the simulation, transient pressure pulsation information at these key locations can be accurately obtained.

[0070] Based on acoustic analogy theory, the extracted transient pressure fluctuation data is transformed into dipole and quadrupole sound sources. Acoustic analogy theory is a theoretical method that connects acoustic problems in a flow field with fluid mechanics problems. This theory can transform complex flow field pressure fluctuation information into aerodynamic sound source models with clear physical meaning. Dipole sound sources are mainly related to pressure fluctuations on the surface of an object, while quadrupole sound sources are related to turbulent motion in the flow field. These two sound source models can accurately describe the generation mechanism of aerodynamic noise inside a centrifugal blower.

[0071] The calculated mechanical vibration response, dipole sound source, and quadrupole sound source are transmitted to the acoustic solver via a real-time data interface. This real-time data interface ensures fast and accurate data transmission between different software modules, avoiding data loss or delay. The acoustic solver receives this sound source data and performs acoustic simulations based on it in subsequent steps to evaluate the noise characteristics of the centrifugal blower.

[0072] Furthermore, the step of setting gear meshing stiffness curves and bearing oil film damping coefficients in the multibody dynamics model, and solving the mechanical vibration response of gear meshing stiffness and bearing oil film oscillations using a nonlinear contact algorithm, specifically includes:

[0073] Based on gear geometry parameters and bearing structural dimensions, a rigid-flexible coupling model including shaft system, gear pair and bearing is established in a multibody dynamics environment;

[0074] The stiffness of a single tooth is calculated by the potential energy method, and the time-varying components caused by the deformation under load are superimposed to generate a stiffness curve containing the meshing fundamental frequency and its harmonics.

[0075] The oil film pressure distribution is solved by the Reynolds equation, and the nonlinear damping coefficient curve related to the rotational speed is extracted.

[0076] Based on the rigid-flexible coupling model, stiffness curves, and nonlinear damping coefficient curves, Hertz contact theory is used to calculate the contact force on the gear tooth surface. The motion equations of the shaft system are solved by numerical integration, and the mechanical vibration response of the key nodes of the shaft system is output.

[0077] In this embodiment, based on the actual gear geometry parameters and bearing structural dimensions of the centrifugal blower, a rigid-flexible coupling model is constructed in a professional multibody dynamics simulation software environment. The gear geometry parameters include key information such as the number of teeth, module, pressure angle, and tooth width, which determine the shape and transmission characteristics of the gear. The bearing structural dimensions include the inner diameter, outer diameter, width, and the diameter and number of balls or rollers, which affect the bearing's load-bearing capacity and motion accuracy.

[0078] The established rigid-flexible coupling model must fully encompass three core components: the shaft system, gear pairs, and bearings. The shaft system, as the key component connecting and supporting gears and bearings, has a significant impact on the overall vibration response due to its stiffness and flexibility characteristics; therefore, its geometry and material properties must be accurately simulated in the model. Gear pairs are the main components transmitting power and motion; the tooth profile and meshing relationship of the gears must be accurately constructed. Bearings support the shaft system and reduce friction; their internal structure and motion must be simulated in detail. For the less rigid parts of the shaft system, flexible body modeling is used to more accurately reflect their deformation during vibration; while for the more rigid parts, rigid body modeling is used to improve computational efficiency.

[0079] The potential energy method is used to calculate the stiffness of a single gear tooth. Based on the principle of deformation energy of a gear under stress, the potential energy method analyzes the elastic deformation of the gear teeth to calculate the stiffness value of a single tooth under different positions and loads. This method considers factors such as the gear's geometry and the material's elastic modulus, and can accurately reflect the stiffness characteristics of a single tooth.

[0080] During gear operation, deformation under load causes time-varying stiffness characteristics. Based on the gear meshing principle and load distribution, the influence of load deformation on the stiffness of a single tooth is analyzed and quantified as a time-varying component. The calculated stiffness of a single tooth is then superimposed with this time-varying component to obtain the variation of the stiffness of a single tooth with time or meshing position, taking into account the influence of load deformation.

[0081] Since multiple teeth engage sequentially during gear meshing, the stiffness of each individual tooth needs to be comprehensively processed. Considering the fundamental meshing frequency and its harmonic components, the stiffness of each individual tooth is superimposed according to the meshing sequence and phase relationship to generate a gear meshing stiffness curve that includes the fundamental meshing frequency and its harmonics. This stiffness curve comprehensively reflects the dynamic changes in gear stiffness during meshing, providing an important basis for subsequent vibration response calculations.

[0082] The pressure distribution of the bearing oil film is solved based on the Reynolds equation. The Reynolds equation is the fundamental equation describing the pressure distribution in a fluid lubrication film, considering the influence of bearing geometry, rotational speed, and lubricating oil viscosity on the oil film pressure. During the solution process, appropriate boundary conditions are set according to the actual bearing structure and operating conditions, such as the inlet and outlet pressures of the oil film and the sealing conditions at both ends of the bearing. The Reynolds equation is solved numerically to obtain the pressure distribution of the bearing oil film at different positions and rotational speeds.

[0083] Based on the obtained oil film pressure distribution, the damping characteristics of the oil film during bearing motion are analyzed. Since the damping characteristics of the oil film are closely related to the rotational speed and exhibit a nonlinear variation at different speeds, the damping coefficients of the oil film at different speeds are extracted and plotted as curves to obtain the nonlinear damping coefficient curve related to the rotational speed. This curve accurately reflects the damping characteristics of the bearing oil film at different speeds, providing key parameters for subsequent vibration response calculations.

[0084] Based on the established rigid-flexible coupling model, the generated gear meshing stiffness curve, and the extracted bearing oil film nonlinear damping coefficient curve, Hertz contact theory is used to calculate the contact force on the gear tooth surface. Hertz contact theory is a classic elastic contact theory that can calculate the pressure distribution and contact force magnitude in the contact area based on the geometry, material properties, and contact load of two elastic bodies. In the calculation process, the gear meshing stiffness curve and the nonlinear damping coefficient curve are used as input parameters, considering the dynamic meshing process of the gear, to accurately calculate the contact force on the gear tooth surface at different times.

[0085] Based on the calculated gear tooth surface contact force, the equations of motion for the shaft system are established according to Newton's second law. These equations describe the motion state of the shaft system under various forces and torques, including displacement, velocity, and acceleration. The numerical integration method is used to solve these equations, which discretizes the continuous motion equations and iteratively calculates the motion state of the shaft system at different times. During the solution process, initial and boundary conditions of the shaft system are considered, such as the initial position and velocity of the shaft system, as well as the bearing support constraints. Finally, the mechanical vibration response of key nodes in the shaft system is output, including curves showing the changes in displacement, velocity, and acceleration over time. These vibration responses directly reflect the vibration characteristics of the shaft system during the operation of the centrifugal blower.

[0086] In this embodiment, the process of setting the gear meshing stiffness curve and bearing oil film damping coefficient in the multibody dynamics model and solving the mechanical vibration response of gear meshing stiffness and bearing oil film oscillation through nonlinear contact algorithm is realized, providing accurate basic data on mechanical vibration for noise assessment simulation of centrifugal blowers.

[0087] Furthermore, the calculation of the stator-rotor air gap magnetic flux density distribution in the electromagnetic finite element model, and the extraction of the electromagnetic force wave spectrum of the main orders via Fourier transform, specifically includes:

[0088] A two-dimensional transient field model was established in electromagnetic finite element software based on the number of stator slots, rotor poles, and winding arrangement parameters of the motor.

[0089] Based on a two-dimensional transient field model, calculation position points are divided within the mechanical angle range of 0°-360°, and the radial air gap magnetic flux distribution at each position point is solved.

[0090] The radial air gap magnetic flux density distribution is converted into a spatial order spectrum by fast Fourier transform, and the main force wave orders with amplitudes exceeding the preset amplitude threshold are screened.

[0091] Based on Maxwell's stress theory, the amplitude and frequency of each major force wave order are calculated to form an electromagnetic force wave spectrum.

[0092] In this embodiment, key parameters of the motor used in the centrifugal blower are collected, including the number of stator slots, the number of rotor poles, and the winding arrangement parameters. The number of stator slots determines the arrangement and quantity of stator windings, which has a significant impact on the magnetic field distribution of the motor; the number of rotor poles affects the number of pole pairs and the speed characteristics of the motor; the winding arrangement parameters cover the number of turns, wire diameter, connection method, etc., and these parameters directly determine the magnetic field strength and distribution generated by the windings.

[0093] Professional electromagnetic finite element software, such as Ansys Maxwell, was selected. Within this software environment, a precise two-dimensional transient field model was established based on the collected motor parameters. During the modeling process, the geometry of the stator and rotor was accurately drawn, including the shape and size of the stator slots, the shape of the rotor poles, and the pole arc coefficient. Simultaneously, the winding layout was rationally set, and the windings were drawn and connected according to the actual winding arrangement parameters to ensure that the model accurately reflects the electromagnetic structure of the motor.

[0094] Define accurate material properties for each component in the model. The stator and rotor are typically made of silicon steel sheets, requiring the setting of magnetic parameters such as permeability and saturation magnetic flux density; the windings generally use copper wire, necessitating the definition of electrical parameters such as conductivity. Furthermore, for the air gap, its relative permeability should be set to 1 to accurately simulate the magnetic properties of air. Accurate material property definitions are crucial for ensuring the reliability of the model's calculation results.

[0095] Based on the established two-dimensional transient field model, calculation position points are uniformly divided within the mechanical angle range of 0° to 360°. The mechanical angle reflects the rotational position of the rotor relative to the stator. By setting calculation points at different mechanical angle positions, the variation of air gap magnetic flux density with rotor position can be comprehensively obtained. The division density of calculation position points can be adjusted according to actual needs. Generally speaking, the denser the division, the more detailed the air gap magnetic flux density distribution information obtained, but the computational workload will also increase accordingly.

[0096] In electromagnetic finite element software, set the transient field solution parameters, including the solution time step and total solution time. The selection of the time step should comprehensively consider the motor speed and the required calculation accuracy, and should generally be less than the time required for the motor to rotate one stator slot pitch to ensure accurate capture of the dynamic changes in air gap magnetic flux density. The total solution time should be long enough to allow the motor to reach a stable operating state, thereby obtaining reliable air gap magnetic flux density distribution data.

[0097] Running the transient field solver program, the software calculates the radial air gap magnetic flux density distribution at each calculation point based on the set parameters and model. Radial air gap magnetic flux density refers to the magnetic flux density component perpendicular to the circumference of the air gap, and it is one of the main factors generating electromagnetic force. The solution results can be presented in the form of contour maps, curves, etc., allowing users to intuitively observe the distribution of air gap magnetic flux density at different locations and times.

[0098] The obtained radial air gap magnetic flux density distribution data are subjected to Fast Fourier Transform (FFT). FFT is an efficient algorithm for converting time-domain signals into frequency-domain signals. It can convert the variation of radial air gap magnetic flux density with mechanical angle into a spatial order spectrum. The spatial order reflects the periodic variation characteristics of the air gap magnetic flux density in space, and the force waves of different orders have different degrees of influence on motor vibration and noise.

[0099] Based on the actual operating conditions of the motor and experience, a reasonable amplitude threshold should be set. This threshold is used to filter out the main force wave orders that have a significant impact on motor vibration and noise. The setting of the amplitude threshold should comprehensively consider factors such as the motor's rated power, speed range, and vibration and noise limits to ensure accurate capture of the main force waves while avoiding excessive interference from irrelevant force waves.

[0100] The spatial order spectrum obtained by the Fast Fourier Transform was analyzed to identify the main force wave orders whose amplitudes exceeded a preset amplitude threshold. These main force wave orders are key factors in generating motor vibration and noise, and further analysis and calculation are required.

[0101] The amplitude and frequency of each major force wave order are calculated based on Maxwell's stress theory. Maxwell's stress theory describes the force exerted by an electromagnetic field on the surface of an object, and it can be used to convert the air gap magnetic flux density distribution into an electromagnetic force distribution. During the calculation, the radial and tangential components of the air gap magnetic flux density, as well as their interaction, must be considered to accurately calculate the electromagnetic force amplitude generated by each major force wave order.

[0102] Based on the motor's rotational speed and the order of the main force waves, the frequencies of each force wave are calculated. Frequency is one of the important characteristic parameters of electromagnetic force waves, and it is closely related to the motor's rotational speed and the order of the force waves. Electromagnetic force waves of different frequencies affect motor vibration and noise in different ways. High-frequency force waves usually cause high-frequency vibration and noise, while low-frequency force waves may lead to low-frequency vibration and structural resonance.

[0103] The calculated amplitudes and frequencies of each major force wave order are presented in graphical form, forming an electromagnetic force wave spectrum. This spectrum visually displays the distribution of electromagnetic force waves at different frequencies and orders, providing crucial information for subsequent analysis of motor vibration and noise characteristics and optimization of motor design.

[0104] In this embodiment, the process of calculating the stator-rotor air gap magnetic flux density distribution in the electromagnetic finite element model and extracting the electromagnetic force wave spectrum of the main order through Fourier transform is realized, providing accurate basic data on electromagnetic force for noise assessment simulation of centrifugal blowers.

[0105] Furthermore, the parallel execution of unsteady CFD simulations extracts transient pressure pulsation data from the impeller and volute surfaces. Based on acoustic analogy theory, the transient pressure pulsation data is converted into dipole and quadrupole sound sources, specifically including:

[0106] Based on the geometric parameters of the impeller and volute, a boundary layer mesh is generated in the CFD preprocessing software, and a large eddy simulation turbulence model is set up.

[0107] Based on the large eddy simulation turbulence model, the computational domain is divided using the MPI domain decomposition technique, and the transient flow field of each subdomain is solved simultaneously, outputting the time-averaged pressure fluctuation data of the impeller surface and the volute tongue region.

[0108] A rotating coordinate system transformation is applied to the time-averaged pressure pulsation data on the impeller surface to compensate for the Doppler effect caused by blade rotation, thereby forming a dipole sound source related to the blade passing frequency.

[0109] Based on the time-averaged pressure pulsation data of the volute tongue region, the Reynolds stress tensor is extracted from the high-turbulence region inside the volute and converted into a quadrupole sound source based on the Lighthill stress tensor theory.

[0110] In this embodiment, geometric parameters of the centrifugal blower impeller and volute are collected, covering key information such as the number of blades, blade shape, impeller diameter, and hub ratio of the impeller, as well as the helical shape, cross-sectional dimensions, and outlet angle of the volute. Using professional 3D modeling software, such as SolidWorks or UG, accurate 3D geometric models of the impeller and volute are constructed based on the collected geometric parameters. This model must fully represent the actual structural features of the impeller and volute, providing an accurate foundation for subsequent mesh generation and flow field simulation.

[0111] The constructed 3D geometric model is imported into CFD preprocessing software, such as ANSYS ICEM CFD or Pointwise. Boundary layer meshes are generated on the impeller blade surface and the inner surface of the volute. The quality of the boundary layer mesh is crucial for accurately capturing the fluid flow characteristics near the wall. Based on the fluid properties and the impeller rotational speed, the height, growth rate, and number of layers of the first boundary layer mesh are appropriately set. Generally, the height of the first layer mesh should meet the dimensionless wall distance y+ requirement; for Large Eddy Simulation (LES), y+ is typically <1 to ensure accurate analysis of the near-wall turbulence structure. Simultaneously, appropriately sized transition meshes are generated in regions far from the wall to ensure a reasonable mesh distribution throughout the computational domain, guaranteeing computational accuracy while controlling the number of meshes and improving computational efficiency.

[0112] In the CFD preprocessing software, select the Large Eddy Simulation (LES) turbulence model. LES is a numerical simulation method that accurately captures large-scale eddy structures in turbulence, providing more detailed transient flow field information compared to the Reynolds-averaged Navier-Stokes (RANS) model. Set the relevant LES parameters, such as the subgrid-scale model. Commonly used subgrid-scale models include the Smagorinsky model and the WALE model. Select an appropriate model based on the specific flow conditions and set the corresponding model parameters, such as the Smagorinsky constant. Simultaneously, set the fluid's physical properties, such as density and viscosity, as well as boundary conditions, including inlet velocity or pressure conditions, outlet pressure conditions, and no-slip conditions on the walls.

[0113] The computational domain is partitioned using Message Passing Interface (MPI) domain decomposition technology. MPI is a programming model for parallel computing that allows data communication and collaborative computation between different computing nodes. In CFD software, the entire computational domain is divided into multiple subdomains based on the amount of computing resources and the characteristics of the computational domain. Each subdomain is assigned to a computing node for solution. The principle of domain decomposition is to balance the computational load across each subdomain as much as possible while minimizing communication between subdomains, thereby improving the efficiency of parallel computing.

[0114] Based on the large eddy simulation (LES) turbulence model, the transient flow field of each subdomain is solved synchronously at each computation node. Appropriate time steps and total computation time are set, with the time step chosen to satisfy the Courant number (CFL) condition to ensure computational stability. Generally, for LES, the time step should be sufficiently small to accurately capture transient changes in turbulence. During the computation, each computation node communicates via MPI to exchange flow field information at the subdomain boundaries, ensuring the continuity and accuracy of the flow field solution throughout the computational domain. After a certain computation time, when the flow field reaches a steady state, the time-averaged pressure fluctuation data of the impeller surface and volute tongue region are output. The time-averaged pressure fluctuation data is obtained by averaging pressure data from multiple time steps, reflecting the average pressure distribution and fluctuation characteristics of the flow field in this region.

[0115] During impeller rotation, pressure pulsations on the blade surface are affected by the Doppler effect. The Doppler effect causes a deviation between the pressure pulsation frequency measured by the observer and the actual passing frequency of the blades. To accurately extract the dipole sound source related to the blade passing frequency, a coordinate transformation needs to be applied to the time-averaged pressure pulsation data of the impeller surface. The principle of the coordinate transformation is to convert the pressure pulsation data from a stationary coordinate system to a coordinate system that rotates with the impeller, thus eliminating the Doppler effect caused by blade rotation.

[0116] In CFD post-processing software or specialized acoustic analysis software, the time-averaged pressure pulsation data of the impeller surface is transformed using a rotating coordinate system. The blade passage frequency is determined based on the impeller's rotational speed and the number of blades. After transformation, the frequency characteristics of the pressure pulsation data are correlated with the blade passage frequency. The transformed pressure pulsation data is considered as the excitation of a dipole sound source. According to the definition of a dipole sound source, it is generated by the interaction between a fluid and an object's surface, and its intensity is proportional to the pressure gradient on the object's surface. Therefore, by processing the impeller surface pressure pulsation data, a dipole sound source correlated with the blade passage frequency can be formed, which accurately reflects the influence of impeller rotation on the sound field.

[0117] Based on time-averaged pressure fluctuation data in the volute tongue region, the Reynolds stress tensor was extracted from the highly turbulent region inside the volute using CFD post-processing software. The Reynolds stress tensor is a crucial physical quantity describing momentum transfer in fluid particles during turbulence, reflecting the intensity and anisotropy of turbulence. In the highly turbulent region, the fluid's turbulent motion is intense, and the components of the Reynolds stress tensor are relatively large. By calculating the Reynolds stress tensor at different locations within the volute tongue region, detailed information about the turbulence in that region can be obtained.

[0118] Based on Lighthill's stress tensor theory, the extracted Reynolds stress tensor is transformed into a quadrupole sound source. Lighthill's stress tensor theory is a theory that links fluid motion with sound radiation, arguing that the generation of a sound source is related to stress changes in the fluid. In turbulent flow, changes in the Reynolds stress tensor cause fluid compression and expansion, thus generating sound radiation. By substituting the Reynolds stress tensor into the Lighthill stress tensor formula, the intensity and distribution of the quadrupole sound source can be calculated. The quadrupole sound source primarily reflects the influence of the non-uniformity and pulsating characteristics within turbulence on the sound field. Unlike dipole sound sources, the intensity of a quadrupole sound source is proportional to the pulsating acceleration of the turbulence.

[0119] In this embodiment, parallel execution of unsteady CFD simulation is realized, transient pressure pulsation data of the impeller and volute surfaces are extracted, and the transient pressure pulsation data is transformed into dipole sound sources and quadrupole sound sources based on acoustic analogy theory, providing accurate aerodynamic sound source basic data for noise assessment simulation of centrifugal blowers.

[0120] In some embodiments, in step S102 above, the step of generating an acoustic mesh for the volute structure using BEM and applying AML technology in the acoustic solver based on vibration and aerodynamic sound source data, while configuring the pipe total reflection boundary and the outlet AML surface to form an acoustic simulation environment, specifically includes:

[0121] Based on the acoustic solver, import vibration sound source data and aerodynamic sound source data;

[0122] For the volute structure, adaptive mesh generation is performed based on its surface curvature characteristics to generate an unstructured boundary element mesh that meets the preset accuracy requirements;

[0123] An AML layer is applied to the inner surface of the volute, the thickness of which is dynamically adjusted according to the highest analysis frequency to absorb acoustic energy with an incident angle of 0°-85°.

[0124] Total reflection boundary conditions are set at the inlet end of the inlet and outlet pipes, and a conical AML surface structure is integrated at the end of the outlet pipe to form a sound propagation path from the closed cavity to the free sound field;

[0125] Based on vibration sound source data, aerodynamic sound source data, unstructured boundary element mesh, AML layer, and sound propagation path, an acoustic simulation environment is formed that includes volute scattering effect and pipeline radiation characteristics.

[0126] In this embodiment, professional acoustic simulation software is selected as the acoustic solver, such as LMS Virtual.LabAcoustics or the acoustic module of COMSOL Multiphysics. The selected acoustic solver is started to ensure that the software runs normally and has the function of performing boundary element method (BEM) calculations and automatic matching layer (AML) technology applications.

[0127] Create a new acoustic simulation project in the acoustic solver. Accurately import the pre-acquired vibration and aerodynamic sound source data into the project using the software's data import interface. Vibration source data typically includes information such as the vibration displacement, velocity, or acceleration of the volute structure at different frequencies. This data can be obtained through experimental testing (e.g., laser vibrometer measurement) or structural dynamics simulation. Aerodynamic sound source data covers the intensity, frequency, and location of dipole and quadrupole sound sources, which can be obtained through the unsteady CFD simulation and sound source conversion process described above. When importing data, ensure that the data format is compatible with the solver and correctly set the units and reference coordinate system according to the software requirements.

[0128] The 3D geometric model of the volute is imported into the acoustic solver. During the import process, necessary preprocessing is performed on the model, such as repairing geometric defects (e.g., gaps, overlapping surfaces) to ensure the integrity and closure of the model. Simultaneously, based on the actual structural characteristics of the volute, the model is appropriately simplified by removing details that have little impact on acoustic calculations (e.g., small chamfers, fillets) to reduce the number of meshes and computational load.

[0129] For the volute structure, adaptive meshing is performed based on its surface curvature characteristics. Mesh generation parameters are set in the solver to meet preset accuracy requirements (i.e., λ / 6 accuracy). Here, λ is the wavelength of the sound wave in the medium, which is related to the frequency of the sound wave and the speed of sound in the medium. Based on the highest frequency analyzed, the corresponding minimum wavelength is calculated, and thus the maximum mesh size is determined. In regions with high surface curvature, such as the curved parts and edges of the volute, the mesh is automatically refined to accurately capture the propagation and scattering characteristics of sound waves in these areas; in regions with relatively flat surfaces, the mesh size is appropriately increased to improve computational efficiency.

[0130] A mesh generation algorithm is run to generate an unstructured boundary element mesh that meets the accuracy requirements. Unstructured meshes offer advantages such as high flexibility and the ability to adapt to complex geometries, better conforming to the curved surface structure of the volute. After mesh generation, the mesh quality is checked, including parameters such as distortion rate and aspect ratio, to ensure that the mesh quality meets the requirements of acoustic calculations. For poor-quality meshes, local optimization or re-meshing is performed to guarantee the accuracy of the calculation results.

[0131] Automatic Matching Layer (AML) technology is an effective method for handling acoustic infinite-domain problems. It simulates an infinite sound field by placing a special artificial dielectric layer near the boundary of the computational domain, causing incident sound waves to gradually attenuate within this layer. The AML layer can absorb sound wave energy within a certain range of incident angles, effectively reducing the impact of boundary reflections on the calculation results.

[0132] An AML layer is applied to the inner surface of the volute. In the solver, the AML layer is automatically added to the boundary element mesh of the inner surface of the volute by selecting the appropriate operation options. The thickness of the AML layer is a critical parameter that needs to be dynamically adjusted based on the highest analysis frequency. Generally, the AML layer thickness should be large enough to ensure sufficient absorption of incident sound waves, but not too large to avoid increasing computational load and time. A suitable AML layer thickness should be set based on empirical formulas or software recommendations, combined with the highest analysis frequency.

[0133] The sound absorption range of the AML layer is set to absorb sound wave energy with an incident angle of 0°-85°. In the solver, by adjusting relevant parameters of the AML layer, such as material properties and attenuation coefficient, the absorption effect of sound waves at different incident angles is achieved. This ensures that the AML layer has good sound absorption performance throughout the entire absorption range to accurately simulate the acoustic environment inside the volute.

[0134] Total internal reflection boundary conditions are set at the inlet ends of the inlet and outlet pipes. These conditions simulate the complete reflection of sound waves when they encounter a rigid wall at the pipe inlet. In the solver, the boundary element mesh at the inlet ends of the inlet and outlet pipes is selected, and its boundary conditions are set to total internal reflection. By setting these boundary conditions, the reflection characteristics of sound waves at the pipe inlet can be accurately simulated, providing correct boundary constraints for acoustic simulations.

[0135] A conical AML surface structure is integrated at the end of the outlet pipe. The conical AML surface structure is a special type of AML structure that better simulates the propagation of sound waves from a closed cavity (inside the volute) to a free sound field. In the solver, the geometric model of the conical AML surface structure is created or imported and matched and integrated with the boundary element mesh at the end of the outlet pipe. The shape and dimensions of the conical AML surface structure are ensured to meet design requirements for accurate sound propagation simulation.

[0136] After setting boundary conditions and integrating the conical AML surface structure, the sound propagation path is verified. The propagation process is visualized in the solver to check whether the sound wave can smoothly propagate from inside the volute through the inlet and outlet pipes to the free sound field, and whether it conforms to the expected reflection and absorption characteristics during propagation. If problems are found in the sound propagation path, such as abnormal sound wave reflection or insufficient absorption, the boundary conditions or AML layer parameters are adjusted promptly until the correct sound propagation path is formed.

[0137] Based on imported vibration and aerodynamic sound source data, an unstructured boundary element mesh is generated, an applied AML layer is applied, and sound propagation paths are set. These data and settings are then integrated in the acoustic solver to form an acoustic simulation environment that includes the volute scattering effect and the pipe radiation characteristics. This simulation environment can accurately simulate the propagation and scattering of sound waves inside the volute, as well as the radiation process to the external free sound field through the inlet and outlet pipes.

[0138] Before conducting formal acoustic simulation calculations, the constructed acoustic simulation environment should be verified and pre-run for checks. This includes verifying the correctness of all parameter settings, such as the location and intensity of the sound source data, mesh connectivity and quality, boundary condition type and parameters, and the thickness and absorption range of the AML layer. The solver's pre-check function should be run to identify and resolve any potential errors or problems. Once the acoustic simulation environment is confirmed to be correctly constructed, subsequent acoustic simulation calculations can be performed to obtain acoustic characteristic data of the volute structure, such as sound pressure level and sound power level.

[0139] In this embodiment, based on vibration and aerodynamic sound source data, an acoustic mesh is generated for the volute structure using BEM in the acoustic solver, and AML technology is applied. At the same time, the total reflection boundary of the pipeline and the AML surface of the outlet are configured to form an acoustic simulation environment that includes the volute scattering effect and the pipeline radiation characteristics, providing an accurate and reliable acoustic calculation platform for the noise assessment simulation of centrifugal blowers.

[0140] In some embodiments, step S103 above, specifically mapping vibration source data to equivalent body forces and performing phase correction on aerodynamic source data in an acoustic simulation environment, and outputting the sound pressure level spectrum by solving the wave equation through FEM, specifically includes:

[0141] Based on the grid node coordinates in the acoustic simulation environment, the mechanical vibration response is converted into the node equivalent body force distribution using an inverse distance weighted interpolation algorithm.

[0142] For a dipole sound source, the phase delay caused by the difference in sound wave propagation path is calculated based on the impeller rotational angular velocity and the field point azimuth angle.

[0143] Based on the phase delay, the sound source intensity is phase-compensated by rotating the coordinate system to generate a spatiotemporally synchronized corrected sound source.

[0144] The equivalent body force distribution at the nodes and the modified sound source are jointly loaded into the acoustic finite element model, and the frequency-varying damping coefficient of the volute shell material is also loaded to form the acoustic wave equation.

[0145] The acoustic wave equation is solved using the finite element method, and the time-domain signal of the sound pressure level at the field point is output. The time-domain signal of the sound pressure level at the field point is then converted into a 1 / 3 octave band spectrum and a narrowband spectrum using the fast Fourier transform.

[0146] The characteristic order components in the 1 / 3 octave spectrum and narrowband spectrum are extracted, and the impeller passing frequency and electromagnetic force wave order are correlated to generate a noise contribution quantification report.

[0147] In this embodiment, within the pre-constructed acoustic simulation environment, the completeness and accuracy of the vibration source data are ensured. This data typically includes vibration response information of the mechanical structure at different locations and times, such as displacement, velocity, or acceleration. This data is imported into software supporting acoustic finite element analysis, such as ANSYS Mechanical Acoustics or NASTRAN. Simultaneously, the mesh node coordinates within the acoustic simulation environment are acquired; these coordinates will serve as the reference for subsequent interpolation calculations.

[0148] Inverse distance weighted interpolation is an interpolation method based on spatial distance. Its basic idea is that the physical quantity value at a grid node is influenced by surrounding known data points, with closer data points having a greater influence. By calculating the distance between the grid node and surrounding known vibration response data points, and then performing a weighted average based on the reciprocal of the distance, the equivalent body force distribution at that grid node can be obtained.

[0149] Write or call the inverse distance weighted interpolation algorithm module in the software. For each grid node, search for known vibration response data points within a certain range around it. Calculate the distance between the grid node and each data point. And calculate the weight based on the distance. , where n is a power of the distance, typically n=2. Then, the vibration response of the surrounding data points is weighted and averaged to obtain the equivalent body force value at that grid node. This process is repeated until the equivalent body force distribution of all grid nodes has been calculated.

[0150] For a dipole sound source, it is necessary to determine the impeller rotational angular velocity ω and the field azimuth angle θ. The impeller rotational angular velocity can be obtained through experimental measurement or the technical parameters of the equipment; it reflects the speed of impeller rotation. The field azimuth angle is the angle between the line connecting the sound source to the field point and a certain reference direction, and it needs to be determined based on the geometric layout of the acoustic simulation environment.

[0151] The phase delay caused by the difference in sound wave propagation path is calculated based on the impeller rotational angular velocity and the azimuth angle of the field point. Due to the impeller's rotation, the path length of the sound waves emitted at different times to reach the field point varies, resulting in a phase difference. By analyzing the impeller's rotational motion and the sound wave propagation characteristics, a model is established relating the phase delay to the impeller rotational angular velocity, the field point azimuth angle, and time. In the software, a calculation program is written based on this model, taking into account parameters such as the impeller rotational angular velocity and the field point azimuth angle, to calculate the phase delay.

[0152] Based on the calculated phase delay, phase compensation is performed on the sound source intensity through a rotating coordinate system transformation. The purpose of the rotating coordinate system transformation is to align the sound source's reference coordinate system with the impeller's rotating coordinate system, thus eliminating the phase change caused by impeller rotation. In the software, a phase compensation algorithm is written based on the principle of rotating coordinate system transformation to correct the intensity of the dipole sound source, generating a spatiotemporally synchronized corrected sound source. The corrected sound source accurately reflects the actual propagation characteristics of sound waves considering impeller rotation.

[0153] The calculated equivalent body force distribution at the nodes and the corrected sound source are jointly loaded into the acoustic finite element model. In the software, through the corresponding interface or user interface, the equivalent body force at the nodes is applied as a force term to the corresponding mesh nodes, and the corrected sound source is loaded as a sound source term at the location of the sound source. It is crucial to ensure the accuracy of the loading position and intensity to guarantee the accuracy of subsequent acoustic calculations.

[0154] Obtain the frequency-varying damping coefficient of the volute shell material. The frequency-varying damping coefficient reflects the material's attenuation characteristics of sound waves at different frequencies, and it can be obtained through experimental testing or by consulting material handbooks. In the software, the frequency-varying damping coefficient is loaded as a material property parameter into the corresponding finite element model of the volute shell. Based on the material's frequency-varying characteristics, damping coefficient values ​​are set at different frequencies, enabling the model to accurately simulate the material's absorption and attenuation of sound waves.

[0155] After applying the equivalent body forces at the nodes and correcting the frequency-varying damping coefficients of the sound source and the volute shell material, the software automatically generates the acoustic wave equation based on acoustic finite element theory. The acoustic wave equation describes the propagation of sound waves in a medium, taking into account factors such as the excitation of the sound source, the characteristics of the medium, and boundary conditions. The generated acoustic wave equation will be used for subsequent finite element method solutions.

[0156] The acoustic wave equations are solved using the finite element method (FEM). The FEM is a method that discretizes a continuous solution domain into a finite number of elements, and approximates the solution domain by solving the equations within each element. In the software, appropriate solution parameters are set, such as the time step and total solution time. The time step should meet computational stability requirements, and the total solution time should be determined based on the frequency range and propagation characteristics of the sound waves. The solver is then run to numerically solve the acoustic wave equations, obtaining the time-domain signal of the sound pressure level at the field point.

[0157] The Fast Fourier Transform (FFT) is an efficient algorithm for converting time-domain signals into frequency-domain signals. It decomposes a time-domain signal into a superposition of sine and cosine components at different frequencies, thus obtaining the signal's spectral characteristics. Using FFT, the time-domain signal of the sound pressure level at a field point can be converted into a frequency-domain signal, facilitating the analysis of sound pressure level distributions at different frequencies.

[0158] The FFT module is invoked in the software to input the time-domain signal of the field sound pressure level into the FFT module for processing. Appropriate frequency resolution and sampling points are set to ensure the accuracy and resolution of the spectrum. After FFT transformation, the 1 / 3 octave band spectrum and narrowband spectrum of the field sound pressure level are output. The 1 / 3 octave band spectrum divides the frequency range into multiple 1 / 3 octave bands, which can intuitively display the sound pressure level distribution within different frequency bands; the narrowband spectrum provides more detailed frequency resolution, enabling precise analysis of the sound pressure level at a specific frequency.

[0159] Characteristic order components are extracted from the output 1 / 3 octave band spectrum and narrowband spectrum. These characteristic order components are frequency elements related to the impeller passage frequency and the order of the electromagnetic force wave. The impeller passage frequency is the frequency of the sound pressure level fluctuation generated per revolution of the impeller, which is related to the impeller's rotational speed and the number of blades; the electromagnetic force wave order is related to the electromagnetic characteristics of the motor. By analyzing the frequency components in the spectrum, the peak values ​​corresponding to these characteristic orders are identified, and the sound pressure level data of the characteristic order components are extracted.

[0160] The extracted characteristic order components are correlated with the impeller via frequency and electromagnetic force wave order. This determines whether each characteristic order component is caused by impeller rotation or electromagnetic force waves, and their contribution to the total noise. By establishing a correlation model, the interactions between different characteristic order components and their impact on the total noise are analyzed, providing a basis for noise control and optimization.

[0161] Based on the results of the correlation analysis, a quantitative report on noise contribution is generated. The report should include sound pressure level data for each characteristic order component, its correlation with the impeller passing frequency and electromagnetic force wave order, and its contribution ratio to the total noise. Furthermore, the distribution of noise contribution can be visually displayed using charts, curves, and other formats to facilitate analysis and decision-making by engineering technicians.

[0162] In some embodiments, step S104 above, which involves embedding a silencer structure into the flow field model based on the sound pressure level spectrum, characterizing the silencer element characteristics through impedance boundaries, and correlating transmission loss with the overall sound power level to form a visual evaluation model for noise reduction effect, specifically includes:

[0163] Based on the sound pressure level spectrum, the frequency bands exceeding the noise standard are located in the flow field model;

[0164] For the frequency band where noise exceeds the standard, parametric geometry of the muffler is added in situ to the flow field model, and boundary layer mesh is divided for the perforated pipe and the sound-absorbing material layer to form the muffler structure.

[0165] An equivalent mass-resistance model is applied to the perforated pipe, and complex wavenumber frequency-varying parameters are set for the sound-absorbing material layer;

[0166] Sound pressure monitoring points are set at the inlet and outlet sections of the silencer structure. The sound power level difference in the target frequency band is calculated by the two-load method to obtain the transmission loss spectrum.

[0167] By using the sound power integration algorithm, the sound pressure level spectrum is converted into the sound power level of the whole machine, and the transfer loss spectrum is correlated with the sound power level of the whole machine to establish a quantitative model for noise reduction contribution.

[0168] In this embodiment, the sound pressure level spectrum data is imported into professional flow field analysis software, such as ANSYS Fluent or STAR-CCM+, which possess powerful flow field and sound field coupling analysis capabilities. Appropriate noise limits are set based on the application scenario of the centrifugal blower and industry noise standards. For example, in an industrial environment, noise limits may be specified for different time periods or different areas. The sound pressure level spectrum is divided according to a certain frequency range; common division methods include 1 / 3 octave band or narrowband spectrum, to more accurately analyze the noise situation in different frequency bands. Each divided frequency band is analyzed individually in the flow field analysis software. The sound pressure level of each frequency band is compared with the set noise limit; when the sound pressure level of a frequency band exceeds the noise limit, that frequency band is determined to be a noise-exceeding frequency band. Using the software's visualization function, the noise-exceeding frequency bands are highlighted on the spectrum diagram with different colors or markers, facilitating intuitive identification by engineering technicians.

[0169] For the identified noise-exceeding frequency bands, parametric geometry of the muffler is added in situ to the flow field model. Parametric geometry design allows for rapid changes in the shape and size of the muffler by adjusting design parameters to meet different noise reduction requirements. Based on the characteristics of the noise-exceeding frequency bands and the design principles of the muffler, the main structural form of the muffler is determined, such as a resistive muffler, a reactive muffler, or a resistive-resistive composite muffler.

[0170] Boundary layer meshing was performed on the perforated tubes and sound-absorbing material layer in the muffler. Boundary layer meshing can more accurately simulate the flow characteristics of fluid near the wall. For the perforated tubes, the fluid flow and sound wave propagation around the perforations need to be considered, therefore a fine mesh is required near the perforations. For the sound-absorbing material layer, due to its complex internal structure, sound waves undergo multiple reflections and absorptions during propagation, so a sufficiently fine mesh is also needed to capture the sound wave propagation process. Using appropriate mesh generation tools in flow field analysis software, high-quality boundary layer meshes were generated based on the muffler's geometry and flow characteristics.

[0171] The perforated pipes and sound-absorbing material layers, pre-gridded, are integrated into the parametric geometry of the muffler to form a complete muffler structure. Accurate connections between components are ensured to avoid gaps or overlaps, guaranteeing the muffler's performance in actual operation. During integration, the compatibility of the muffler structure with the flow field model is checked to ensure the muffler can be correctly embedded in the flow field without affecting normal fluid flow.

[0172] An equivalent mass-drag model is applied to the perforated tube to characterize its sound wave propagation characteristics. The equivalent mass-drag model considers the effects of perforation ratio, pore size, and plate thickness on sound wave reflection and absorption. In the flow field analysis software, the parameters of the equivalent mass-drag model are set according to the geometric parameters and material properties of the perforated tube. These parameters can be obtained through experimental measurements or theoretical calculations to ensure that the model accurately simulates the acoustic performance of the perforated tube in actual operation.

[0173] Complex wavenumber frequency-varying parameters are set for the sound-absorbing material layer to reflect its sound absorption characteristics at different frequencies. The complex wavenumber contains information such as the sound velocity and attenuation coefficient of the sound-absorbing material, and it is closely related to frequency. Complex wavenumber data for the sound-absorbing material at different frequencies are obtained through experimental testing or by consulting the material's technical manual. In flow field analysis software, these frequency-varying parameters are set as properties of the sound-absorbing material layer, enabling the software to accurately simulate the absorption of sound waves of different frequencies by the material.

[0174] Sound pressure monitoring points should be installed at the inlet and outlet sections of the muffler structure. The locations of the sound pressure monitoring points should be selected to accurately reflect the sound pressure characteristics as sound waves enter and leave the muffler. Generally, the sound pressure monitoring point at the inlet section should be far from the inlet edge of the muffler to avoid the influence of boundary effects; the sound pressure monitoring point at the outlet section should be located at the center of the muffler outlet or at a representative location.

[0175] The two-load method is a commonly used approach for calculating the transmission loss of a muffler. Its basic idea is to apply different load conditions at the muffler's inlet or outlet, measure the corresponding sound pressure data, and then calculate the transmission loss based on acoustic theory. In this embodiment, by changing the boundary conditions at the muffler's inlet or outlet, different load conditions are simulated to obtain sound pressure monitoring point data under different loads.

[0176] Based on the sound pressure monitoring data measured using the two-load method, the sound power level difference within the target frequency band is calculated. The sound power level difference reflects the attenuation capability of the silencer for sound waves. By performing calculations at multiple frequency points, the transmission loss as a function of frequency, i.e., the transmission loss spectrum, is obtained. In the flow field analysis software, a corresponding calculation program is written or called, inputting sound pressure monitoring data and frequency range parameters, to automatically calculate the transmission loss spectrum and display the results in graphical form.

[0177] The sound pressure level spectrum is converted into the overall sound power level of the centrifugal blower using a sound power integration algorithm. This algorithm considers the sound pressure distribution at various locations within the sound field, and by integrating the sound pressure levels, the overall sound power level of the centrifugal blower is obtained. In the flow field analysis software, the sound power calculation function is used, inputting the sound pressure level spectrum data and the geometric information of the flow field model, to perform sound power integration calculations and obtain the overall sound power level of the centrifugal blower.

[0178] A quantified model of noise reduction contribution is established by correlating the transmission loss spectrum with the overall acoustic power level. This model directly reflects the noise reduction effect of the muffler on the overall system noise. By comparing the overall acoustic power level before and after muffler installation, as well as the values ​​of the transmission loss spectrum at different frequencies, the contribution of the muffler to the overall system noise in different frequency bands is analyzed. Different evaluation indicators, such as noise reduction amount and noise reduction rate, can be set in the model to quantify the noise reduction effect of the muffler.

[0179] By utilizing the visualization capabilities of flow field analysis software, the results of the noise reduction contribution quantification model can be presented in an intuitive graphical or animated format. For example, color cloud maps can be used to display the noise reduction effect of the muffler at different frequency bands, with red indicating a significant noise reduction effect and blue indicating a poor noise reduction effect. Animations can also be used to demonstrate the propagation and attenuation of sound waves during the operation of the muffler, enabling engineers to more intuitively understand the noise reduction principle and effect of the muffler.

[0180] In some embodiments, step S105 above, which involves building a Simulink dynamic simulation platform, responding to changes in rotational speed and load by calling the acoustic solver in real time, and linking the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics, specifically includes:

[0181] Based on the aerodynamic characteristics of the fan and the parameters of the motor, a co-simulation platform including the impeller rotor dynamics model, the pipeline flow capacity model and the valve regulation function was built in the Simulink environment.

[0182] The speed command and load resistance are set as time-varying inputs, and the valve opening adjustment signal is generated by the PID controller to drive the dynamic response of the fan system.

[0183] When the change in speed command and load resistance at any operating point exceeds the preset change threshold, the acoustic solver is triggered to recalculate across the entire frequency band.

[0184] For operating points where the change does not exceed the preset change threshold, the noise characteristics are updated using a spectrum interpolation algorithm.

[0185] The surge boundary prediction module is called synchronously. Based on the LSTM network, it analyzes real-time operating data and historical surge records, outputs stability margin threshold, and determines the surge boundary.

[0186] When the real-time operating conditions approach the surge boundary, the anti-surge control logic is executed first and the noise calculation is frozen, and the time-varying noise characteristic curve is output.

[0187] In this embodiment, data related to the aerodynamic characteristics of the fan are collected, covering impeller geometric parameters (such as the number of blades, blade shape, impeller diameter, etc.) and aerodynamic performance curves (including flow-head curves and flow-efficiency curves at different speeds). Simultaneously, detailed motor parameters are obtained, such as rated power, rated speed, and torque-speed characteristics. For the pipeline flow capacity model, information such as the pipeline layout, diameter, length, and roughness is required to determine the flow capacity characteristic parameters of the pipeline. The valve regulation function is defined based on the valve type (such as butterfly valve, ball valve, etc.) and its characteristic curves.

[0188] In the Simulink environment, a rotor dynamics model of the impeller is built using the corresponding module library. This model should be able to simulate the impeller's dynamic characteristics, such as moment of inertia and torque balance, under different speeds and loads. For example, the rotation equations of the impeller are constructed using the integral and gain modules to achieve dynamic calculation of speed and torque. When constructing the pipe network flow-capacity model, fluid mechanics principles and signal processing modules in Simulink are used to simulate the flow and pressure changes of the fluid within the pipe network. The pipe network is viewed as a system composed of multiple flow-capacity elements; by establishing the connections between these elements, flow distribution and pressure transmission are achieved. For the valve control function, the relationship between valve opening and flow rate is defined using lookup tables or mathematical functions based on the valve's characteristic curve. The valve control function is then connected to the pipe network flow-capacity model and the impeller rotor dynamics model to form a complete co-simulation platform.

[0189] After setting up the co-simulation platform, model verification is performed. Known operating parameters (such as fixed speed and load) are input, and the simulation results are compared with theoretical calculations or experimental data. The model's output is checked to ensure it meets expectations; for example, parameters such as flow rate, head, and speed are within reasonable ranges. If the simulation results do not match expectations, the model is debugged. The parameter settings of each module are checked for correctness, the connections are reasonable, and there are any issues such as numerical instability. By gradually adjusting the model parameters and structure, the model can accurately simulate the dynamic characteristics of the wind turbine.

[0190] In the Simulink model, the speed command and load resistance are set as time-varying inputs. A signal generator module can be used to generate different forms of time-varying signals, such as step signals, sinusoidal signals, and random signals, to simulate the changes in speed and load under actual operating conditions. The range and rate of change of the speed command and load resistance are defined to ensure that the input signals can cover various operating conditions the wind turbine may encounter. For example, the speed command can be set to vary between 50% and 120% of the rated speed, and the load resistance can vary between 30% and 150% of the rated load.

[0191] Design a PID controller to generate valve opening adjustment signals. The PID controller adjusts the valve opening based on the deviation between the speed command and the actual speed, using proportional, integral, and derivative actions to control the fan speed. The PID controller parameters are tuned using empirical or trial-and-error methods. First, the proportional coefficient is adjusted to enable the system to respond quickly to changes in the speed command; then, the integral coefficient is adjusted to eliminate steady-state errors; finally, the derivative coefficient is adjusted to improve system stability and anti-interference capability. Through multiple simulation experiments, an optimal set of PID parameters is found, enabling the fan system to respond quickly and stably to changes in speed and load.

[0192] Run the Simulink model and observe the dynamic response of the wind turbine system under different time-varying inputs. Record the variation curves of parameters such as speed, flow rate, head, and valve opening, and analyze the system's response speed, stability, and overshoot performance indicators. If the system's dynamic response is found to be unsatisfactory, further adjust the PID controller parameters or optimize the model structure, such as adding a feedforward control element or adopting an adaptive control strategy, to improve the dynamic performance of the wind turbine system.

[0193] Based on the actual operating conditions of the wind turbine and the accuracy requirements of the acoustic calculations, threshold values ​​for the changes in speed command and load resistance should be set. The magnitude of these threshold values ​​should comprehensively consider the calculation time and accuracy of the acoustic solver. Generally, setting the threshold too small will cause the acoustic solver to trigger frequently, increasing calculation time; setting the threshold too large will reduce the accuracy of noise characteristic updates. For example, the threshold value for the change in speed command can be set to 5% of the rated speed, and the threshold value for the change in load resistance can be set to 10% of the rated load.

[0194] In the Simulink model, a comparison module is used to determine whether the changes in the speed command and load resistance at any operating point exceed a preset threshold. When the changes exceed the threshold, the acoustic solver is triggered to perform a full-frequency recalculation. The acoustic solver can be co-simulated with Simulink or a custom acoustic calculation module can be used within Simulink. After triggering the acoustic solver, the current operating parameters (such as speed, flow rate, and head) are passed to the acoustic solver for full-frequency noise calculation.

[0195] For operating conditions where the change in noise level does not exceed a preset threshold, a spectral interpolation algorithm is used to update the noise characteristics. The spectral interpolation algorithm calculates the noise spectrum under the current operating condition based on known noise spectrum data and changes in operating parameters. In Simulink, the spectral interpolation algorithm can be implemented using an interpolation module or a user-defined function. It takes historical noise spectrum data and current operating parameters as input and outputs the updated noise spectrum. Through the spectral interpolation algorithm, noise characteristics can be updated in real time while maintaining computational efficiency.

[0196] Real-time operating condition data and historical surge records are collected, including parameters such as speed, flow rate, pressure head, and temperature. The collected data undergoes preprocessing, such as normalization and outlier removal, to improve the training performance of the LSTM network. The preprocessed data is then used to train the LSTM network. The LSTM network is a recurrent neural network with memory capabilities, enabling it to process time-series data. Through training, the LSTM network learns the relationship between real-time operating condition data and surge occurrence, thereby accurately predicting surge boundaries.

[0197] In the Simulink model, the surge boundary prediction module is invoked synchronously. Real-time operating data is input into the trained LSTM network, which outputs a stability margin threshold. The stability margin threshold reflects the distance between the current operating point and the surge boundary. When the stability margin threshold is small, it indicates that the current operating point is close to the surge boundary, and the wind turbine system is at risk of surge.

[0198] When real-time operating conditions approach the surge boundary, i.e., when the stability margin threshold is less than the preset safety threshold, the anti-surge control logic is executed first. The anti-surge control logic can employ measures such as venting and recirculation to increase the fan's flow rate and prevent surge. Simultaneously with the execution of the anti-surge control logic, noise calculations are frozen. This is because when the fan approaches the surge boundary, its internal flow state changes drastically, and noise calculations performed at this time may be inaccurate. Freezing noise calculations avoids outputting erroneous time-varying noise characteristic curves.

[0199] The noise characteristic data, updated by acoustic solver calculations or spectral interpolation algorithms, along with information on the execution status of anti-surge control logic, are integrated. The integrated data undergoes further processing, such as filtering and smoothing, to improve data quality and visualization.

[0200] In Simulink, the plotting module or an interface with external plotting software (such as MATLAB) can be used to plot the processed noise data as a time-varying noise characteristic curve. This curve visually displays how noise changes over time and the noise level under different operating conditions. Key information, such as surge boundary points and anti-surge control logic execution points, can be labeled on the curve to help engineers analyze the causes of noise generation and the effectiveness of control measures.

[0201] Reference Figure 2 An embodiment of the present invention provides a noise assessment simulation system 2 for a centrifugal blower, the system 2 specifically comprising:

[0202] The first evaluation simulation module 201 is used to calculate mechanical vibration noise and couple electromagnetic excitation through a multibody dynamics model, while performing unsteady CFD simulation to extract aerodynamic sound sources, and transmitting the vibration sound source data and aerodynamic sound source data to the acoustic solver via a real-time interface.

[0203] The second evaluation simulation module 202 is used to generate an acoustic mesh for the volute structure using BEM and apply AML technology in the acoustic solver based on vibration sound source data and aerodynamic sound source data. At the same time, it configures the pipe total reflection boundary and the outlet AML surface to form an acoustic simulation environment.

[0204] The third evaluation simulation module 203 is used to map vibration sound source data into equivalent body forces and perform phase correction on aerodynamic sound source data in an acoustic simulation environment, and output the sound pressure level spectrum by solving the wave equation through FEM.

[0205] The fourth evaluation simulation module 204 is used to embed the silencer structure into the flow field model based on the sound pressure level spectrum, characterize the characteristics of the silencer element through the impedance boundary, correlate the transmission loss with the sound power level of the whole machine, and form a visual evaluation model of the noise reduction effect.

[0206] The fifth evaluation simulation module 205 is used to build a Simulink dynamic simulation platform. It calls the acoustic solver in real time in response to changes in rotational speed and load, and links with the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

[0207] It is understandable that, such as Figure 1 The content of the noise assessment simulation method embodiment for centrifugal blowers shown is applicable to the noise assessment simulation system embodiment for centrifugal blowers. The specific functions implemented by the noise assessment simulation system embodiment for centrifugal blowers are as follows: Figure 1 The noise assessment simulation method for centrifugal blowers shown is the same as that implemented in this example, and the beneficial effects achieved are the same as those described above. Figure 1 The beneficial effects achieved by the illustrated simulation method for noise assessment of centrifugal blowers are also the same.

[0208] It should be noted that the information interaction and execution process between the above systems are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0209] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0210] Reference Figure 3 The present invention also provides a computer device 3, including: a memory 302 and a processor 301, and a computer program 303 stored in the memory 302. When the computer program 303 is executed on the processor 301, it implements the noise assessment simulation method for a centrifugal blower as described in any of the above methods.

[0211] The computer device 3 may be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will understand that... Figure 3 The computer device 3 is merely an example and does not constitute a limitation on the computer device 3. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0212] The processor 301 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0213] In some embodiments, the memory 302 may be an internal storage unit of the computer device 3, such as a hard disk or memory of the computer device 3. In other embodiments, the memory 302 may be an external storage device of the computer device 3, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 3. Furthermore, the memory 302 may include both internal and external storage units of the computer device 3. The memory 302 is used to store the operating system, applications, boot loader, data, and other programs, such as the program code of the computer program. The memory 302 can also be used to temporarily store data that has been output or will be output.

[0214] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the noise assessment simulation method for a centrifugal blower as described in any of the above methods.

[0215] In this embodiment, if the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0216] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0217] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0218] In the embodiments disclosed in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0219] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

Claims

1. A noise assessment simulation method for a centrifugal blower, characterized in that, The method specifically includes: Mechanical vibration noise is calculated by multibody dynamics model and coupled with electromagnetic excitation. At the same time, unsteady CFD simulation is performed to extract aerodynamic sound sources. The vibration sound source data and aerodynamic sound source data are transmitted to the acoustic solver via real-time interface. Based on vibration and aerodynamic sound source data, BEM is used to generate an acoustic mesh for the volute structure in the acoustic solver and AML technology is applied. At the same time, the total reflection boundary of the pipe and the AML surface of the outlet are configured to form an acoustic simulation environment. In the acoustic simulation environment, vibration sound source data is mapped to equivalent body forces and aerodynamic sound source data is phase corrected. The wave equation is solved by FEM to output the sound pressure level spectrum. Based on the sound pressure level spectrum, a silencer structure is embedded in the flow field model. The characteristics of the silencer element are characterized by the impedance boundary. The transmission loss is correlated with the sound power level of the whole machine to form a visual evaluation model of the noise reduction effect. A Simulink dynamic simulation platform was built, which calls the acoustic solver in real time in response to changes in rotational speed and load, and links the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

2. The method according to claim 1, characterized in that, The process involves calculating mechanical vibration noise using a multibody dynamics model and coupling it with electromagnetic excitation, while simultaneously performing unsteady CFD simulation to extract aerodynamic sound sources. The vibration and aerodynamic sound source data are then transmitted to the acoustic solver via a real-time interface. Specifically, this includes: In the multibody dynamics model, the gear meshing stiffness curve and the bearing oil film damping coefficient are set, and the mechanical vibration response of gear meshing stiffness and bearing oil film oscillation is solved by nonlinear contact algorithm. The stator-rotor air gap magnetic flux density distribution was calculated in the electromagnetic finite element model, and the electromagnetic force wave spectrum of the main order was extracted by Fourier transform. The electromagnetic force wave spectrum is reconstructed into a time-domain excitation force through inverse Fourier transform and then superimposed onto the mechanical vibration response; Parallel execution of unsteady CFD simulations was performed to extract transient pressure pulsation data from the impeller and volute surfaces. Based on acoustic analogy theory, the transient pressure pulsation data was converted into dipole and quadrupole sound sources. The mechanical vibration response, dipole sound source, and quadrupole sound source are transmitted to the acoustic solver via a real-time data interface.

3. The method according to claim 2, characterized in that, The process of setting gear meshing stiffness curves and bearing oil film damping coefficients in a multibody dynamics model, and solving the mechanical vibration response of gear meshing stiffness and bearing oil film oscillations using a nonlinear contact algorithm, specifically includes: Based on gear geometry parameters and bearing structural dimensions, a rigid-flexible coupling model including shaft system, gear pair and bearing is established in a multibody dynamics environment; The stiffness of a single tooth is calculated by the potential energy method, and the time-varying components caused by the deformation under load are superimposed to generate a stiffness curve containing the meshing fundamental frequency and its harmonics. The oil film pressure distribution is solved by the Reynolds equation, and the nonlinear damping coefficient curve related to the rotational speed is extracted. Based on the rigid-flexible coupling model, stiffness curves, and nonlinear damping coefficient curves, Hertz contact theory is used to calculate the contact force on the gear tooth surface. The motion equations of the shaft system are solved by numerical integration, and the mechanical vibration response of the key nodes of the shaft system is output.

4. The method according to claim 2, characterized in that, The calculation of the stator-rotor air gap magnetic flux density distribution in the electromagnetic finite element model, followed by Fourier transform extraction of the electromagnetic force wave spectrum of the main orders, specifically includes: A two-dimensional transient field model was established in electromagnetic finite element software based on the number of stator slots, rotor poles, and winding arrangement parameters of the motor. Based on a two-dimensional transient field model, calculation position points are divided within the mechanical angle range of 0°-360°, and the radial air gap magnetic flux distribution at each position point is solved. The radial air gap magnetic flux density distribution is converted into a spatial order spectrum by fast Fourier transform, and the main force wave orders with amplitudes exceeding the preset amplitude threshold are screened. Based on Maxwell's stress theory, the amplitude and frequency of each major force wave order are calculated to form an electromagnetic force wave spectrum.

5. The method according to claim 2, characterized in that, The parallel execution of unsteady CFD simulations extracts transient pressure pulsation data from the impeller and volute surfaces. Based on acoustic analogy theory, the transient pressure pulsation data is converted into dipole and quadrupole sound sources, specifically including: Based on the geometric parameters of the impeller and volute, a boundary layer mesh is generated in the CFD preprocessing software, and a large eddy simulation turbulence model is set up. Based on the large eddy simulation turbulence model, the computational domain is divided using the MPI domain decomposition technique, and the transient flow field of each subdomain is solved simultaneously, outputting the time-averaged pressure fluctuation data of the impeller surface and the volute tongue region. A rotating coordinate system transformation is applied to the time-averaged pressure pulsation data on the impeller surface to compensate for the Doppler effect caused by blade rotation, thereby forming a dipole sound source related to the blade passing frequency. Based on the time-averaged pressure pulsation data of the volute tongue region, the Reynolds stress tensor is extracted from the high-turbulence region inside the volute and converted into a quadrupole sound source based on the Lighthill stress tensor theory.

6. The method according to claim 2, characterized in that, Based on vibration and aerodynamic sound source data, an acoustic mesh is generated for the volute structure using BEM in the acoustic solver, and AML technology is applied. Simultaneously, the pipe total reflection boundary and the outlet AML surface are configured to form an acoustic simulation environment, specifically including: Based on the acoustic solver, import vibration sound source data and aerodynamic sound source data; For the volute structure, adaptive mesh generation is performed based on its surface curvature characteristics to generate an unstructured boundary element mesh that meets the preset accuracy requirements; An AML layer is applied to the inner surface of the volute, the thickness of which is dynamically adjusted according to the highest analysis frequency to absorb acoustic energy with an incident angle of 0°-85°. Total reflection boundary conditions are set at the inlet end of the inlet and outlet pipes, and a conical AML surface structure is integrated at the end of the outlet pipe to form a sound propagation path from the closed cavity to the free sound field; Based on vibration sound source data, aerodynamic sound source data, unstructured boundary element mesh, AML layer, and sound propagation path, an acoustic simulation environment is formed that includes volute scattering effect and pipeline radiation characteristics.

7. The method according to claim 6, characterized in that, In the acoustic simulation environment, the vibration sound source data is mapped to equivalent body forces and the aerodynamic sound source data is phase-corrected. The sound pressure level spectrum is output by solving the wave equation through FEM. Specifically, this includes: Based on the grid node coordinates in the acoustic simulation environment, the mechanical vibration response is converted into the node equivalent body force distribution using an inverse distance weighted interpolation algorithm. For a dipole sound source, the phase delay caused by the difference in sound wave propagation path is calculated based on the impeller rotational angular velocity and the field point azimuth angle. Based on the phase delay, the sound source intensity is phase-compensated by rotating the coordinate system to generate a spatiotemporally synchronized corrected sound source. The equivalent body force distribution at the nodes and the modified sound source are jointly loaded into the acoustic finite element model, and the frequency-varying damping coefficient of the volute shell material is also loaded to form the acoustic wave equation. The acoustic wave equation is solved using the finite element method, and the time-domain signal of the sound pressure level at the field point is output. The time-domain signal of the sound pressure level at the field point is then converted into a 1 / 3 octave band spectrum and a narrowband spectrum using the fast Fourier transform. The characteristic order components in the 1 / 3 octave band spectrum and narrowband spectrum are extracted, and the impeller passing frequency and electromagnetic force wave order are correlated to generate a noise contribution quantification report.

8. The method according to claim 1, characterized in that, The method, based on the sound pressure level spectrum, embeds a silencer structure into the flow field model, characterizes the silencer element's properties through impedance boundaries, and correlates transmission loss with the overall sound power level to form a visual evaluation model for noise reduction effectiveness. Specifically, this includes: Based on the sound pressure level spectrum, the frequency bands exceeding the noise standard are located in the flow field model; For the frequency band where noise exceeds the standard, the parametric geometry of the muffler is added in situ in the flow field model, and the boundary layer mesh is divided for the perforated pipe and the sound-absorbing material layer to form the muffler structure. An equivalent mass-resistance model is applied to the perforated pipe, and complex wavenumber frequency-varying parameters are set for the sound-absorbing material layer; Sound pressure monitoring points are set at the inlet and outlet sections of the silencer structure. The sound power level difference in the target frequency band is calculated by the two-load method to obtain the transmission loss spectrum. By using the sound power integration algorithm, the sound pressure level spectrum is converted into the sound power level of the whole machine, and the transfer loss spectrum is correlated with the sound power level of the whole machine to establish a quantitative model for noise reduction contribution.

9. The method according to claim 1, characterized in that, The aforementioned construction of the Simulink dynamic simulation platform, which responds to changes in rotational speed and load by calling the acoustic solver in real time and linking it with the surge boundary prediction module to optimize the control logic and output time-varying noise characteristics, specifically includes: Based on the aerodynamic characteristics of the fan and the parameters of the motor, a co-simulation platform including the impeller rotor dynamics model, the pipeline flow capacity model and the valve regulation function was built in the Simulink environment. The speed command and load resistance are set as time-varying inputs, and the valve opening adjustment signal is generated by the PID controller to drive the dynamic response of the fan system. When the change in speed command and load resistance at any operating point exceeds the preset change threshold, the acoustic solver is triggered to recalculate across the entire frequency band. For operating points where the change does not exceed the preset change threshold, the noise characteristics are updated using a spectrum interpolation algorithm. The surge boundary prediction module is called synchronously. Based on the LSTM network, it analyzes real-time operating data and historical surge records, outputs stability margin threshold, and determines the surge boundary. When the real-time operating conditions approach the surge boundary, the anti-surge control logic is executed first and the noise calculation is frozen, and the time-varying noise characteristic curve is output.

10. A noise assessment simulation system for a centrifugal blower, characterized in that, The system specifically includes: The first evaluation simulation module is used to calculate mechanical vibration noise and couple electromagnetic excitation through a multibody dynamics model, while performing unsteady CFD simulation to extract aerodynamic sound sources and transmitting the vibration sound source data and aerodynamic sound source data to the acoustic solver via a real-time interface. The second evaluation simulation module is used to generate an acoustic mesh for the volute structure using BEM and apply AML technology in the acoustic solver based on vibration and aerodynamic sound source data. At the same time, it configures the pipe total reflection boundary and the outlet AML surface to form an acoustic simulation environment. The third evaluation simulation module is used to map vibration sound source data into equivalent body forces and perform phase correction on aerodynamic sound source data in an acoustic simulation environment, and output the sound pressure level spectrum by solving the wave equation through FEM. The fourth evaluation simulation module is used to embed the silencer structure into the flow field model based on the sound pressure level spectrum, characterize the characteristics of the silencer element through the impedance boundary, correlate the transmission loss with the sound power level of the whole machine, and form a visual evaluation model of the noise reduction effect. The fifth evaluation simulation module is used to build a Simulink dynamic simulation platform. It calls the acoustic solver in real time in response to changes in rotational speed and load, and links with the surge boundary prediction module to optimize the control logic to output time-varying noise characteristics.

Images