Intelligent detection method and system for household low-noise volute

Abstract

The invention relates to the technical field of intelligent detection, in particular to an intelligent detection method and system for a household low-noise volute, and the method comprises the steps: building a digital twinborn model according to geometric data and material characteristics of the volute, obtaining a noise characteristic predicted value through calculating CFD simulation and acoustic simulation, and outputting performance basic data; in the dynamic detection environment, collecting multi-modal data under different working conditions; performing spectral analysis on the collected noise signal to generate a voiceprint map, and judging volute performance through a deep learning model in combination with airflow distribution data and vibration characteristics to obtain an intelligent detection result; generating a design improvement scheme of the volute based on an optimization algorithm; and feeding back the design improvement scheme to the digital twinborn model, updating model parameters, and performing subsequent detection by using the updated digital twinborn model. According to the invention, the problems of single detection working condition and inaccurate noise cause analysis are effectively solved, and dynamic evaluation and optimal design of volute performance are realized.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to an intelligent detection method and system for a household low-noise volute. Background Art

[0002] The volute is a core component widely used in household appliances (such as exhaust fans, air conditioners, etc.), and its performance has an important impact on the airflow efficiency and noise level of the equipment. In the existing technology, the detection of volute performance is mainly achieved through physical experiments, usually with the overall airflow efficiency and sound pressure level (decibel value) as the main measurement indicators.

[0003] However, these traditional detection methods have significant shortcomings. First, the detection conditions are single. Conventional tests only target fixed operating conditions and it is difficult to capture the dynamic performance of the volute in an actual complex operating environment. Second, the existing detection methods lack accurate analysis of the source of noise and cannot effectively distinguish the specific causes of noise, such as problems caused by eddy current turbulence, mechanical vibration or structural resonance. Summary of the invention

[0004] The present invention provides an intelligent detection method and system for a household low-noise volute, thereby effectively solving the problems pointed out in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: An intelligent detection method for a household low-noise volute, comprising: According to the geometric data and material properties of the volute, a digital twin model is established. Through CFD simulation and acoustic simulation, the noise characteristic prediction value is obtained and the basic performance data is output; In a dynamic detection environment, multimodal data under different working conditions are collected, wherein the multimodal data includes noise signals, airflow distribution data and vibration amplitude; Performing spectrum analysis on the collected noise signal to generate a soundprint spectrum, and combining the airflow distribution data and vibration characteristics to judge the volute performance through a deep learning model to obtain an intelligent detection result; According to the intelligent detection results, generating a design improvement plan for the volute based on an optimization algorithm; The design improvement plan is fed back to the digital twin model, the model parameters are updated, and the updated digital twin model is used for subsequent testing.

[0006] Furthermore, a digital twin model is established based on the geometric data and material properties of the volute. By calculating CFD simulation and acoustic simulation, the noise characteristic prediction value is obtained and the basic performance data is output, including: According to the geometric data and material properties of the volute, the internal and external structural characteristics of the volute are obtained, combined with the material parameters, as input data; Based on the input data, a digital twin model is established to construct a virtual model consistent with the actual operating state of the volute; Based on the digital twin model, CFD simulation is used to simulate the airflow behavior inside the volute and output airflow distribution data; According to the airflow distribution data, combined with airflow noise and vibration coupling effects, acoustic simulation is performed to output noise characteristic data; Based on the airflow distribution data and the noise characteristic data, the basic performance data of the volute is summarized and generated.

[0007] Furthermore, the multimodal data is fused by means of feature level fusion.

[0008] Furthermore, the collected noise signal is subjected to spectrum analysis to generate a voiceprint spectrum, including: De-noising the noise signal, filtering out high-frequency and low-frequency interference signals through a bandpass filter, and normalizing the signal; Short-time Fourier transform is used to convert the preprocessed noise signal into time-frequency domain to generate a spectrum diagram of time-frequency distribution; Extracting key features of the noise signal from the spectrum diagram, including main frequency, spectrum energy distribution and its dynamic change characteristics over time; According to the time-frequency distribution of the spectrum diagram, a voiceprint spectrum is generated, with frequency as the horizontal axis and time as the vertical axis.

[0009] Furthermore, the airflow distribution data and vibration characteristics are combined to judge the performance of the volute through a deep learning model to obtain intelligent detection results, including: Perform feature-level fusion on the collected airflow distribution data, vibration signals and voiceprint spectra to generate input features for multi-modal fusion; Building a deep learning model, the model including a convolutional neural network and a recurrent neural network, for extracting spatial and temporal series features of input features; The fused multimodal features are input into the deep learning model, and the intelligent detection results of the volute are output through the model.

[0010] Furthermore, according to the intelligent detection results, a design improvement scheme for the volute is generated based on an optimization algorithm, including: Determining a design optimization target according to the volute performance defects identified in the intelligent detection results; Based on the digital twin model, the mapping relationship between the volute geometry parameters, material parameters, and structural parameters and the performance targets is constructed to form a design optimization problem; Solving the design optimization problem using a multi-objective optimization algorithm to generate improved design parameters; The generated improved design parameters are subjected to performance simulation verification, and the optimized volute design scheme is output in combination with the digital twin model.

[0011] Furthermore, the design optimization problem is solved using a multi-objective optimization algorithm to generate improved design parameters, including: A multi-objective optimization function is constructed with the goal of reducing volute noise, reducing vibration amplitude and improving airflow efficiency; The multi-objective optimization function is solved based on a genetic algorithm to search for the optimal design parameter combination that meets the objectives, and a weight adjustment mechanism is introduced to output the optimized design parameters.

[0012] Furthermore, the design improvement scheme is fed back to the digital twin model, the model parameters are updated, and subsequent testing is performed using the updated digital twin model, including: Importing the optimized design parameters in the improved scheme into the twin model, and updating the geometric parameters, material parameters and structural characteristics of the volute; Based on the updated twin model, the airflow behavior, noise characteristics and vibration characteristics of the volute are re-simulated to generate new performance prediction data; Evaluate the improvement effect of the optimization scheme through the newly generated performance prediction data to verify whether the design goals are met; If the design goal is not met, the design parameters are further adjusted according to the intelligent prediction results and the above updating and testing process is repeated.

[0013] An intelligent detection system for a household low-noise volute, the system comprising: The model building module builds a digital twin model based on the geometric data and material properties of the volute, obtains the predicted value of noise characteristics through CFD simulation and acoustic simulation, and outputs basic performance data; A data acquisition module, which collects multimodal data under different working conditions in a dynamic detection environment, wherein the multimodal data includes noise signals, airflow distribution data, and vibration amplitude; The feature analysis module performs spectrum analysis on the collected noise signal to generate a soundprint spectrum, and combines the airflow distribution data and vibration characteristics to judge the performance of the volute through a deep learning model to obtain an intelligent detection result; A design optimization module, which generates a design improvement plan for the volute based on an optimization algorithm according to the intelligent detection results; The feedback update module feeds back the design improvement plan to the digital twin model, updates the model parameters, and uses the updated digital twin model for subsequent testing.

[0014] Furthermore, the design optimization module includes: An optimization target determination unit, which determines a design optimization target according to the volute performance defects identified in the intelligent detection results; The optimization problem modeling unit constructs the mapping relationship between the volute geometry parameters, material parameters, and structural parameters and the performance targets based on the digital twin model to form a design optimization problem; An optimization problem solving unit, which uses a multi-objective optimization algorithm to solve the design optimization problem and generate improved design parameters; The optimized design output unit performs performance simulation verification on the generated improved design parameters and outputs the optimized volute design scheme in combination with the digital twin model.

[0015] The technical solution of the present invention can achieve the following technical effects: It effectively solves the problems of single detection conditions and inaccurate noise cause analysis, and realizes dynamic evaluation and optimal design of volute performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A schematic diagram of a process for intelligent detection of a household low-noise volute; Figure 2 A flow chart for obtaining noise characteristic prediction values ​​and outputting basic performance data; Figure 3 A schematic diagram of the process of generating a voiceprint map; Figure 4 A schematic diagram of the process for judging the performance of the volute and obtaining intelligent detection results; Figure 5 A schematic diagram of the process for generating a design improvement scheme for the volute; Figure 6 Schematic diagram of the process for generating improved design parameters; Figure 7 A schematic diagram of the process of updating model parameters and performing subsequent testing. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0020] Embodiment 1 like Figure 1 As shown, the present invention provides an intelligent detection method for a household low-noise volute, the method comprising: S1: According to the geometric data and material properties of the volute, a digital twin model is established. By calculating CFD simulation and acoustic simulation, the noise characteristic prediction value is obtained and the basic performance data is output; Specifically, by establishing a digital twin model that is consistent with the actual operating state of the volute, the geometric characteristics and material parameters of the volute are digitally expressed, and CFD simulation and acoustic simulation technology are used to simulate the airflow behavior, vortex distribution and the coupling effect of airflow and structure inside the volute in detail. Through these simulation calculations, the noise characteristics of the volute can be predicted in a virtual environment, including key information such as the main frequency of noise generation and sound pressure level distribution, thereby generating basic performance data of the volute, providing a scientific basis and reference direction for subsequent intelligent detection, performance optimization and design improvement.

[0021] S2: In a dynamic detection environment, collect multimodal data under different working conditions, including noise signals, airflow distribution data and vibration amplitude; Specifically, in a dynamic detection environment, different working conditions of the volute in actual operation are simulated, and multimodal data including noise signals, airflow distribution data and vibration amplitude are collected to fully reflect the true performance of the volute under complex working conditions. These data are not only used to capture the dynamic change characteristics of the volute under different operating conditions, but also provide multi-source data support for training and evaluation of subsequent deep learning models, thereby achieving accurate analysis of volute performance and comprehensive identification of potential problems.

[0022] S3: Perform spectrum analysis on the collected noise signal to generate a soundprint map, and combine the airflow distribution data and vibration characteristics to judge the volute performance through a deep learning model to obtain intelligent detection results; Specifically, the collected noise signals are spectrally analyzed to generate a soundprint atlas to extract key features of the noise signal (such as main frequency, spectral energy distribution and dynamic changes), which are then combined with airflow distribution data and vibration characteristics. Multimodal data analysis is performed through a deep learning model. This process aims to accurately determine the performance status of the volute, including noise level, airflow efficiency and vibration stability, identify potential performance defects or problems, and thus provide comprehensive and intelligent detection results, providing data support for subsequent design improvements.

[0023] S4: Based on the intelligent detection results, a design improvement plan for the volute is generated based on the optimization algorithm; Specifically, the performance defects or deficiencies of the volute identified in the intelligent detection results are used to determine the specific goals that need to be optimized (such as reducing noise, improving airflow efficiency or reducing vibration). By constructing optimization problems related to the geometric parameters, material properties and structural properties of the volute and combining them with optimization algorithms (such as multi-objective optimization algorithms, genetic algorithms, etc.), the volute design parameters are solved and adjusted to generate a design scheme with improved performance.

[0024] S5: Feedback the design improvement plan to the digital twin model, update the model parameters, and use the updated digital twin model for subsequent testing.

[0025] Specifically, the generated design improvement plan is applied to the digital twin model, and the model's geometric parameters, material properties, and operating conditions are updated, so that the performance of the improved design can be verified and evaluated in a virtual environment. By using the updated digital twin model for subsequent testing, the performance of the improvement plan under different working conditions can be quickly simulated, and its optimization effect on noise, airflow efficiency, and vibration characteristics can be evaluated to ensure the feasibility and performance improvement of the design. At the same time, this process realizes a closed loop of optimization feedback, improves the efficiency and accuracy of design iterations, and avoids the large amount of cost and time required for direct physical testing.

[0026] The present invention effectively solves the problems of single detection working condition and inaccurate noise cause analysis, and realizes dynamic evaluation and optimal design of volute performance.

[0027] As a preferred embodiment of the above, Figure 2 As shown, in step S1, a digital twin model is established according to the geometric data and material properties of the volute, and the noise characteristic prediction value is obtained by calculating CFD simulation and acoustic simulation, and basic performance data is output, including: S11: according to the geometric data and material properties of the volute, the internal and external structural characteristics of the volute are obtained, combined with the material parameters, as input data; S12: Based on the input data, a digital twin model is established to construct a virtual model consistent with the actual operating state of the volute; S13: Based on the digital twin model, CFD simulation is used to simulate the airflow behavior inside the volute and output the airflow distribution data; S14: Perform acoustic simulation based on the airflow distribution data, combined with the airflow noise and vibration coupling effects, and output noise characteristic data; S15: Based on the airflow distribution data and the noise characteristic data, basic performance data of the volute is generated.

[0028] Specifically, we first obtain the internal and external structural characteristics of the volute, including key geometric parameters such as flow channel size, inlet and outlet positions, and combine them with material parameters (such as density, elastic modulus, acoustic impedance coefficient, etc.) to use these data as input information to provide basic support for model construction. On this basis, we use digital modeling technology (such as CAD modeling or finite element modeling) to establish a digital twin model that is consistent with the actual operating state of the volute. By introducing the operating boundary conditions of the volute (such as wind speed, air volume, pressure and temperature, etc.), we ensure that the virtual model can dynamically simulate the actual working conditions. Based on the digital twin model, we use CFD (computational fluid dynamics) simulation technology to simulate the airflow behavior inside the volute in detail to capture key dynamic characteristics of the airflow, such as velocity distribution, vortex area and pressure changes. Through simulation calculation, the airflow distribution data is output to provide an input basis for noise analysis; then, based on the airflow distribution data, the coupling effect of airflow noise (such as eddy noise) and structural vibration is taken into consideration, and the acoustic simulation technology is used to analyze the sound pressure level, frequency distribution and noise source location of the noise to generate the noise characteristic data of the volute; finally, the airflow distribution data and noise characteristic data are summarized to generate a complete basic data set of volute performance, which integrates information on noise characteristics (such as main frequency and sound pressure level), airflow efficiency (such as flow stability and flow velocity distribution) and vibration characteristics (such as frequency response).

[0029] As a preferred embodiment of the above embodiment, multimodal data is fused by feature level fusion.

[0030] Specifically, feature-level fusion is selected as the multimodal data fusion method, which is closely integrated with the intelligent detection process of this scheme. It can fully mine the core information of noise signals, airflow distribution data and vibration characteristics, and reflect the multiple dimensions of volute performance. Through feature-level fusion, the key information of these modal data is integrated into a unified feature vector, which not only retains the unique characteristics of each mode, but also improves the comprehensive analysis ability of volute performance through complementarity. In this scheme, feature-level fusion further optimizes the data processing process, eliminates redundant information and noise interference through feature selection and dimensionality reduction, and directly outputs a unified feature vector that meets the input requirements of the deep learning model, and seamlessly connects with the deep learning model. This method is particularly suitable for the multimodal collaborative analysis tasks required by this scheme, and can maintain the stability and robustness of the data under complex working conditions, and improve the accuracy and efficiency of detection. Through feature-level fusion, this scheme realizes the deep integration of multimodal data, so that the synergy of data can be fully utilized, ensuring the accuracy and reliability of intelligent detection results, and is particularly suitable for the evaluation and optimization design of volute performance under complex working conditions.

[0031] As a preferred embodiment of the above, Figure 3 As shown, the collected noise signal is subjected to spectrum analysis to generate a voiceprint spectrum, including: A10: De-noise the noise signal, filter out high-frequency and low-frequency interference signals through a bandpass filter, and normalize the signal at the same time; A20: Use short-time Fourier transform to convert the preprocessed noise signal into time-frequency domain and generate a spectrum diagram of time-frequency distribution; A30: Extract key features of noise signals from the spectrum, including the main frequency, spectrum energy distribution and its dynamic change characteristics over time; A40: Generate a voiceprint spectrum based on the time-frequency distribution of the spectrum, with frequency as the horizontal axis and time as the vertical axis.

[0032] Specifically, first, the high-frequency and low-frequency interference components in the noise signal are denoised through a bandpass filter to filter out irrelevant components introduced by environmental noise or signal acquisition equipment. At the same time, the processed noise signal is normalized to ensure that the signal amplitude range is consistent, thereby eliminating the impact of signal amplitude changes on subsequent analysis; next, the short-time Fourier transform (STFT) is used to convert the preprocessed noise signal from the time domain to the time-frequency domain to generate a spectrum diagram containing time and frequency distribution. The short-time Fourier transform can capture the dynamic change characteristics of the signal in time and frequency, and provide time-frequency resolution for extracting detailed features of the noise signal; then, the generated The key features of the noise signal are extracted from the spectrum diagram, including the main frequency, spectrum energy distribution and its dynamic change characteristics over time. These features can clearly reflect the main sources of noise and their characteristics, such as the noise components caused by eddy currents, vibrations or structural resonance during the operation of the volute; finally, according to the time-frequency distribution of the spectrum diagram, a voiceprint spectrum is generated, with frequency as the horizontal axis and time as the vertical axis, to intuitively display the frequency change law and time characteristics of the noise signal. As a visual signal expression, the voiceprint spectrum can not only reveal the characteristics and causes of noise, but also provide important data support for the input of subsequent deep learning models and the intelligent evaluation of volute performance.

[0033] As a preferred embodiment of the above, Figure 4 As shown in the figure, the airflow distribution data and vibration characteristics are combined to judge the performance of the volute through a deep learning model, and intelligent detection results are obtained, including: B10: Perform feature-level fusion on the collected airflow distribution data, vibration signals and voiceprint spectra to generate input features for multi-modal fusion; B20: Build a deep learning model, which includes convolutional neural network and recurrent neural network, to extract spatial and temporal series features of input features; B30: Input the fused multimodal features into the deep learning model, and output the intelligent detection results of the volute through the model.

[0034] Specifically, the collected airflow distribution data, vibration signals and voiceprint maps are fused at the feature level. By extracting and integrating the key features of different modal data (such as the velocity distribution of airflow, the frequency characteristics of vibration signals and the time-frequency characteristics of voiceprint maps), a unified multimodal fusion feature input vector is generated. This fusion method can not only retain the data characteristics of each modality, but also utilize the complementarity between multimodal data to improve the accuracy of comprehensive performance evaluation. Then, a deep learning model for performance evaluation is constructed. The model adopts a combined structure of convolutional neural network (CNN) and recurrent neural network (RNN), in which CNN is used to extract the spatial features of voiceprint maps and airflow distribution data, such as local patterns and feature distributions; RNN (such as LSTM or GRU) is used to capture the time series dynamic characteristics of vibration signals and airflow distribution data. This model architecture can make full use of the correlation between spatial and temporal features to accurately analyze the performance of the volute. Finally, the fused multimodal features are input into the deep learning model. Through model training and reasoning, the intelligent detection results of the volute are output. The detection results include key performance indicators such as the volute's noise level, airflow efficiency, and vibration stability, as well as the identification of potential performance defects (such as abnormal noise in specific frequency bands, turbulent airflow, or excessive vibration). These results can intuitively reflect the overall operating status and performance of the volute.

[0035] As a preferred embodiment of the above, Figure 5 As shown, step S4, according to the intelligent detection result, generates a design improvement scheme of the volute based on the optimization algorithm, including: S41: Determine the design optimization target according to the volute performance defects identified in the intelligent detection results; S42: Based on the digital twin model, the mapping relationship between the volute geometry parameters, material parameters and structural parameters and the performance targets is constructed to form a design optimization problem; S43: solving the design optimization problem using a multi-objective optimization algorithm to generate improved design parameters; S44: Perform performance simulation verification on the generated improved design parameters, and output the optimized volute design scheme in combination with the digital twin model.

[0036] Specifically, first, based on the volute performance defects identified in the intelligent detection results, specific problems that need to be improved are identified, such as excessive high-frequency noise, vortex turbulence leading to reduced airflow efficiency or abnormal vibration. Based on these problems, the corresponding design optimization goals are determined, such as reducing the main frequency peak of noise, improving the uniformity of airflow flow or reducing the amplitude of structural resonance; then, based on the digital twin model, a mapping relationship between the geometric parameters, material parameters and structural parameters of the volute and the performance goals is established. For example, an association model is established between the flow path curvature, inlet width or material damping characteristics of the volute and performance indicators such as noise characteristics and airflow efficiency. Through this process, the performance objectives are converted into optimizable design parameters to form a specific design optimization problem. Subsequently, a multi-objective optimization algorithm is used to solve the design optimization problem. The optimization algorithm establishes a balance between performance objectives (such as reducing noise and improving efficiency) and generates a set of improved design parameters that meet the target requirements through iterative calculations. These parameters include geometric adjustments (such as volute shape and size optimization) and material improvements (such as the selection of lower noise materials). Finally, the generated improved design parameters are subjected to performance simulation verification. Combined with the digital twin model, the performance of the optimized volute under actual operating conditions is simulated to verify whether the design optimization objectives have been achieved. Based on the simulation verification results, the optimized volute design scheme is output, including detailed geometric parameters, material selection and performance prediction values, to provide a basis for actual manufacturing and subsequent verification.

[0037] As a preferred embodiment of the above, Figure 6 As shown, step S43, using a multi-objective optimization algorithm to solve the design optimization problem and generate improved design parameters, including: S431: Construct a multi-objective optimization function with the goal of reducing volute noise, reducing vibration amplitude and improving airflow efficiency; S432: Solve the multi-objective optimization function based on the genetic algorithm, search for the optimal design parameter combination that meets the objectives, introduce a weight adjustment mechanism, and output the optimized design parameters.

[0038] Specifically, firstly, the optimization objectives are to reduce the noise level, vibration amplitude and airflow efficiency during the operation of the volute, and quantify these performance indicators into mathematical expressions. For example, the noise level can be expressed by the reduction of the sound pressure level (SPL), the vibration amplitude can be expressed by the reduction of the vibration acceleration or displacement amplitude, and the airflow efficiency can be expressed by the uniformity of the airflow distribution or the reduction of the resistance. In the optimization function, each performance objective can exist independently and be allowed to balance each other to form a multi-objective mathematical optimization model. Then, the multi-objective optimization function is solved based on the genetic algorithm. The genetic algorithm is an optimization algorithm that simulates the natural selection process. It continuously iterates and searches for the optimal solution through operations such as population initialization, selection, crossover and mutation. In this process, the design parameters of the volute (such as geometric shape, flow channel size, material properties) are used as optimization variables and matched with the constructed optimization objective function. In order to ensure the scientificity and effectiveness of the optimization results, the algorithm introduces a weight adjustment mechanism. Different weights are assigned to the objectives of reducing noise, reducing vibration and improving airflow efficiency according to actual needs, and the optimization direction is dynamically adjusted. Finally, through multiple rounds of iterations, the genetic algorithm generates the optimal design parameter combination that meets the multi-objective optimization requirements.

[0039] As a preferred embodiment of the above, Figure 7 As shown, step S5, feeding back the design improvement plan to the digital twin model, updating the model parameters, and using the updated digital twin model for subsequent testing, including: S51: Import the optimized design parameters in the improved scheme into the twin model, and update the geometric parameters, material parameters and structural characteristics of the volute; S52: Based on the updated twin model, the airflow behavior, noise characteristics and vibration characteristics of the volute are re-simulated to generate new performance prediction data; S53: Evaluate the improvement effect of the optimization scheme through the newly generated performance prediction data to verify whether the design goal is met; S54: If the design goal is not met, further adjust the design parameters according to the intelligent prediction results and repeat the above updating and testing process.

[0040] Specifically, first, the optimized design parameters generated in the design improvement plan are imported into the digital twin model. The optimized design parameters include the geometric parameters of the volute (such as curvature, flow channel size, etc.), material parameters (such as density, elastic modulus, etc.) and structural characteristics (such as damping coefficient, wall thickness, etc.). The digital twin model is updated by these parameters to ensure that the model can accurately reflect the performance characteristics of the volute after the design improvement; then, based on the updated digital twin model, the operating behavior of the volute under different working conditions is re-simulated, and the airflow behavior of the volute is analyzed through CFD simulation to generate airflow distribution and pressure change data; the noise characteristics are simulated through acoustic simulation, and the sound pressure level, main frequency and distribution characteristics of the noise are analyzed; combined with the vibration characteristics simulation, the dynamic stability and resonance of the structure are evaluated. These simulation data can provide performance prediction results of the optimized design scheme under actual working conditions; then, the improvement effect of the optimized design scheme is evaluated through the newly generated performance prediction data, including whether the noise level is significantly reduced, the uniformity of airflow distribution is improved, and the amplitude of structural vibration is reduced. At the same time, these prediction data are compared with the initial performance data and optimization goals to verify whether the design meets the expected goals; finally, if the optimization scheme does not fully meet the design goals, the design parameters are further adjusted according to the performance prediction results, including the geometry, material selection or structural characteristics, and the new design parameters are fed back to the digital twin model again, and the update and detection process is repeated until the design requirements are met. Through this closed-loop feedback mechanism, it is ensured that the effect of the design improvement is gradually optimized to achieve the optimal state of performance improvement.

[0041] Embodiment 2 Based on the same inventive concept as the intelligent detection method for a household low-noise volute in the aforementioned embodiment, the present invention also provides an intelligent detection system for a household low-noise volute, the system comprising: The model building module builds a digital twin model based on the geometric data and material properties of the volute, obtains the predicted value of noise characteristics through CFD simulation and acoustic simulation, and outputs basic performance data; The data acquisition module collects multimodal data under different working conditions in a dynamic detection environment. The multimodal data includes noise signals, airflow distribution data and vibration amplitude; The feature analysis module performs spectrum analysis on the collected noise signals to generate a soundprint map, and combines the airflow distribution data and vibration characteristics to judge the volute performance through a deep learning model to obtain intelligent detection results; Design optimization module, which generates design improvement schemes for the volute based on optimization algorithms according to intelligent detection results; The feedback update module feeds back the design improvement plan to the digital twin model, updates the model parameters, and uses the updated digital twin model for subsequent testing.

[0042] The above-mentioned detection system in the present invention can effectively realize the intelligent detection method of household low-noise volutes, and the technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0043] As a preferred embodiment of the above, the design optimization module includes: An optimization target determination unit determines the design optimization target based on the volute performance defects identified in the intelligent detection results; The optimization problem modeling unit constructs the mapping relationship between the volute geometry parameters, material parameters, and structural parameters and the performance targets based on the digital twin model to form a design optimization problem; The optimization problem solving unit uses a multi-objective optimization algorithm to solve the design optimization problem and generate improved design parameters; The optimized design output unit is used to perform performance simulation verification on the generated improved design parameters, and the optimized volute design scheme is output in combination with the digital twin model.

[0044] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the optimization effects corresponding to the method in Example 1, which will not be repeated here.

[0045] Although the present application has been described in conjunction with specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined herein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application.

[0046] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these modifications and variations.

Claims

1. An intelligent detection method for household low-noise volute, characterized in that: include: According to the geometric data and material properties of the volute, a digital twin model is established. Through CFD simulation and acoustic simulation, the noise characteristic prediction value is obtained and the basic performance data is output; In a dynamic detection environment, multimodal data under different working conditions are collected, wherein the multimodal data includes noise signals, airflow distribution data and vibration amplitude; Performing spectrum analysis on the collected noise signal to generate a soundprint spectrum, and combining the airflow distribution data and vibration characteristics to judge the volute performance through a deep learning model to obtain an intelligent detection result; According to the intelligent detection results, generating a design improvement plan for the volute based on an optimization algorithm; The design improvement plan is fed back to the digital twin model, the model parameters are updated, and the updated digital twin model is used for subsequent testing.

2. The intelligent detection method for household low-noise volute according to claim 1 is characterized in that: According to the geometric data and material properties of the volute, a digital twin model is established. Through CFD simulation and acoustic simulation, the noise characteristic prediction value is obtained and the basic performance data is output, including: According to the geometric data and material properties of the volute, the internal and external structural characteristics of the volute are obtained, combined with the material parameters, as input data; Based on the input data, a digital twin model is established to construct a virtual model consistent with the actual operating state of the volute; Based on the digital twin model, CFD simulation is used to simulate the airflow behavior inside the volute and output airflow distribution data; According to the airflow distribution data, combined with airflow noise and vibration coupling effects, acoustic simulation is performed to output noise characteristic data; Based on the airflow distribution data and the noise characteristic data, the basic performance data of the volute is summarized and generated.

3. The intelligent detection method for household low-noise volute according to claim 1 is characterized in that: The multimodal data is fused by feature-level fusion.

4. The intelligent detection method for household low-noise volute according to claim 1, characterized in that: Performing spectrum analysis on the collected noise signal to generate a voiceprint spectrum includes: De-noising the noise signal, filtering out high-frequency and low-frequency interference signals through a bandpass filter, and normalizing the signal; Short-time Fourier transform is used to convert the preprocessed noise signal into time-frequency domain to generate a spectrum diagram of time-frequency distribution; Extracting key features of the noise signal from the spectrum diagram, including main frequency, spectrum energy distribution and its dynamic change characteristics over time; According to the time-frequency distribution of the spectrum diagram, a voiceprint spectrum is generated, with frequency as the horizontal axis and time as the vertical axis.

5. The intelligent detection method for household low-noise volute according to claim 1, characterized in that: Combining the airflow distribution data and vibration characteristics, the volute performance is judged through a deep learning model to obtain intelligent detection results, including: Perform feature-level fusion on the collected airflow distribution data, vibration signals and voiceprint spectra to generate input features for multi-modal fusion; Building a deep learning model, the model including a convolutional neural network and a recurrent neural network, for extracting spatial and temporal series features of input features; The fused multimodal features are input into the deep learning model, and the intelligent detection results of the volute are output through the model.

6. The intelligent detection method for household low-noise volute according to claim 1, characterized in that: According to the intelligent detection results, a design improvement scheme for the volute is generated based on an optimization algorithm, including: Determining a design optimization target according to the volute performance defects identified in the intelligent detection results; Based on the digital twin model, the mapping relationship between the volute geometry parameters, material parameters, and structural parameters and the performance targets is constructed to form a design optimization problem; Solving the design optimization problem using a multi-objective optimization algorithm to generate improved design parameters; The generated improved design parameters are subjected to performance simulation verification, and the optimized volute design scheme is output in combination with the digital twin model.

7. The intelligent detection method for household low-noise volute according to claim 6, characterized in that: The design optimization problem is solved using a multi-objective optimization algorithm to generate improved design parameters, including: A multi-objective optimization function is constructed with the goal of reducing volute noise, reducing vibration amplitude and improving airflow efficiency; The multi-objective optimization function is solved based on a genetic algorithm to search for the optimal design parameter combination that meets the objectives, and a weight adjustment mechanism is introduced to output the optimized design parameters.

8. The intelligent detection method for household low-noise volute according to claim 1, characterized in that: Feeding back the design improvement plan to the digital twin model, updating the model parameters, and using the updated digital twin model for subsequent testing, including: Importing the optimized design parameters in the improved scheme into the twin model, and updating the geometric parameters, material parameters and structural characteristics of the volute; Based on the updated twin model, the airflow behavior, noise characteristics and vibration characteristics of the volute are re-simulated to generate new performance prediction data; Evaluate the improvement effect of the optimization scheme through the newly generated performance prediction data to verify whether the design goals are met; If the design goal is not met, the design parameters are further adjusted according to the intelligent prediction results and the above updating and testing process is repeated.

9. An intelligent detection system for low-noise volutes for household use, characterized in that: The system comprises: The model building module builds a digital twin model based on the geometric data and material properties of the volute, obtains the predicted value of noise characteristics through CFD simulation and acoustic simulation, and outputs basic performance data; A data acquisition module, which collects multimodal data under different working conditions in a dynamic detection environment, wherein the multimodal data includes noise signals, airflow distribution data, and vibration amplitude; The feature analysis module performs spectrum analysis on the collected noise signal to generate a soundprint spectrum, and combines the airflow distribution data and vibration characteristics to judge the performance of the volute through a deep learning model to obtain an intelligent detection result; A design optimization module, which generates a design improvement plan for the volute based on an optimization algorithm according to the intelligent detection results; The feedback update module feeds back the design improvement plan to the digital twin model, updates the model parameters, and uses the updated digital twin model for subsequent testing.

10. The intelligent detection system for household low-noise volute according to claim 9, characterized in that: The design optimization module includes: An optimization target determination unit, which determines a design optimization target according to the volute performance defects identified in the intelligent detection results; The optimization problem modeling unit constructs the mapping relationship between the volute geometry parameters, material parameters, and structural parameters and the performance targets based on the digital twin model to form a design optimization problem; An optimization problem solving unit, which uses a multi-objective optimization algorithm to solve the design optimization problem and generate improved design parameters; The optimized design output unit performs performance simulation verification on the generated improved design parameters and outputs the optimized volute design scheme in combination with the digital twin model.

Images