A design method and system of a noise reduction orifice plate of a pressure regulating branch of a natural gas station

Abstract

The application discloses a natural gas station pressure regulating branch noise reduction orifice plate design method and system, and relates to the technical field of mechanical design, wherein the method comprises the following steps: sampling structural design parameters; establishing an acoustic simulation model, and calculating the aerodynamic noise value corresponding to the sample points; establishing a Kriging model; if the model error does not meet the error requirement, using an adaptive multi-scale sampling method to obtain new sample points, and repeating the operation until the new Kriging model meets the error requirement; taking the final Kriging model as a target function, and establishing an optimization design model of the noise reduction orifice plate structure; and performing optimization calculation on the optimization design model to obtain optimal structural design parameters. The adaptive multi-scale sampling method can adjust the number of sampling samples according to the change in the accuracy during the Kriging model training process, balance the modeling cost and efficiency, and improve the optimization efficiency of the noise reduction orifice plate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical design, and particularly relates to a design method of a noise reduction orifice plate in a natural gas conveying pipe network. BACKGROUND

[0002] A natural gas station is a transfer station for gas pipeline distribution, pressure regulation and metering. The pressure regulation branch pipeline of the station will generate strong turbulent aerodynamic noise under the condition of large flow and large pressure drop. This noise will not only cause forced vibration of the pipeline and related components, leading to pipeline structure damage, but also form noise pollution, which further endangers the health of the station and the surrounding personnel. At present, the main way to suppress noise is to set a noise reduction orifice plate inside the downstream pipeline to reduce the aerodynamic noise level by using the small hole diffusion principle of the noise reduction orifice plate. However, at present, the design of the noise reduction orifice plate structure is mainly based on experience and relevant standards of the throttle orifice plate. This is mainly because the design of the multi-orifice plate structure for the purpose of noise reduction needs complex fluid simulation and aerodynamic noise calculation to obtain the aerodynamic noise value corresponding to different structure parameters, and then through repeated calculation and optimization to obtain the final optimized noise reduction orifice plate structure. The fluid simulation, aerodynamic noise calculation and other processes involved in the entire optimization design process need to consume a large amount of computing power and time, and a single calculation may take hundreds of hours. The optimization algorithm is used to calculate and optimize the structure parameters of the noise reduction orifice plate, which may need to be repeated thousands of times, which leads to a very time-consuming optimization design process, and the optimization design efficiency is very low, which cannot meet the needs of rapid engineering design.

[0003] To improve the optimization efficiency of the noise reduction orifice plate structure, a surrogate model is introduced into the structural optimization design of the noise reduction orifice plate. The orifice plate structure optimization and flow characteristic analysis based on the surrogate model (such as Kriging, radial basis function network RBF, neural network, etc.) have been carried out. By using the surrogate model to replace the high-cost CFD (computational fluid dynamics) analysis, the efficiency of fluid characteristic analysis and structural optimization can be significantly improved. Among them, the Kriging model has a natural advantage in high-dimensional nonlinear problems due to its good nonlinear fitting ability, accurate unbiased estimation function and unique error estimation function. In the modeling process of the Kriging model, in order to ensure the prediction accuracy of the Kriging model, after establishing the initial Kriging model based on the initial sample, it is necessary to judge whether the initial model meets the accuracy requirement. If the requirement is not met, the point with the maximum error is extracted by the sampling method, and the true response value of the point is obtained through experiment or simulation calculation, and is added to the initial sample to re-establish the Kriging model. The process is iterated until the prediction accuracy of the model meets the accuracy requirement. In this process, the sample filling criterion directly affects the modeling efficiency and prediction accuracy of the Kriging model. The traditional sample filling criterion uses a single-point sampling method, that is, only one sample point is added to the initial sample at a time, which leads to low modeling efficiency of the Kriging model. Based on this, related research results propose a multi-point parallel filling method, that is, multiple sample points are extracted and added to the initial sample at the same time each time to improve the modeling efficiency. However, with the increase of the number of iteration cycles of the Kriging model, the accuracy of the model continues to improve in the later iteration period, and a small number of sample points or even a single sample point can meet the accuracy iteration requirement each time. Adding a fixed number of sample points each time will cause serious waste of computing resources, leading to low modeling efficiency of the Kriging model in the later modeling period, low optimization efficiency of the pressure regulating branch noise reduction orifice plate based on the Kriging model, and difficulty in effectively carrying out the individualized design of the noise reduction orifice plate.

[0004] It can be seen that the prior art still needs to be improved and improved. SUMMARY

[0005] The embodiments of the present application provide a natural gas station pressure regulating branch noise reduction orifice plate design method and system to solve the problem of low optimization efficiency of a fixed number of sample points in the prior art.

[0006] In one aspect, the embodiments of the present application provide a natural gas station pressure regulating branch noise reduction orifice plate design method, comprising:

[0007] The value range of each structural design parameter of the natural gas pressure regulating branch noise reduction orifice plate is set, wherein, considering the processability and installation size of the aperture, the existing structural size of the noise reduction orifice plate is taken as an initial value, and the value range of each design parameter is determined by floating the initial value up and down in a corresponding interval;

[0008] Random sampling is performed on each structural design parameter in the value range to obtain n groups of sample points;

[0009] The aerodynamic noise value of the pressure regulating branch after installation of the noise reduction orifice plate is taken as an evaluation index, an acoustic simulation model of the pressure regulating branch is established by using CFD software, and the aerodynamic noise values corresponding to the n groups of sample points are calculated;

[0010] The n groups of sample points and the corresponding aerodynamic noise values are taken as initial sample points, a Kriging model is established, error evaluation is performed on the Kriging model, if the error of the Kriging model does not meet the error requirement, new sample points are obtained by using an adaptive multi-scale sampling method, new aerodynamic noise values corresponding to the new sample points are calculated, the new sample points and the new aerodynamic noise values are added to the initial sample points, a new Kriging model is re-established, error evaluation is performed again, until the new Kriging model established meets the error requirement, and a final Kriging model is obtained; wherein, the adaptive multi-scale sampling method is to adaptively adjust the number of sampling samples according to the accuracy change of the Kriging model;

[0011] An optimization design model of the noise reduction orifice plate structure is established by taking the final Kriging model as a target function and the structural design parameters of the noise reduction orifice plate as design variables;

[0012] Optimization calculation is performed on the optimization design model to obtain optimal structural design parameters.

[0013] On the other hand, the application also provides a natural gas station pressure regulating branch noise reduction orifice plate design system, comprising:

[0014] A parameter setting module is configured to set the value range of each structural design parameter of the natural gas pressure regulating branch noise reduction orifice plate, wherein, considering the processability and installation size of the aperture, the existing structural size of the noise reduction orifice plate is taken as an initial value, and the value range of each design parameter is determined by floating the initial value up and down in a corresponding interval;

[0015] A random sampling module is configured to randomly sample each structural design parameter in the value range to obtain n groups of sample points;

[0016] A modeling simulation module is configured to take the aerodynamic noise value of the pressure regulating branch after installation of the noise reduction orifice plate as an evaluation index, establish an acoustic simulation model of the pressure regulating branch by using CFD software, and calculate the aerodynamic noise values corresponding to the n groups of sample points;

[0017] The evaluation and optimization module is used to establish a Kriging model using n sets of sample points and their corresponding aerodynamic noise values ​​as initial sample points. It then evaluates the error of the Kriging model. If the error of the Kriging model does not meet the error requirements, an adaptive multi-scale sampling method is used to obtain new sample points. The new aerodynamic noise values ​​corresponding to these new sample points are calculated, and the new sample points and their corresponding aerodynamic noise values ​​are added to the initial sample points. A new Kriging model is then established, and the error evaluation is repeated until the established new Kriging model meets the error requirements, resulting in the final Kriging model. The adaptive multi-scale sampling method adjusts the number of sampled samples adaptively based on changes in the accuracy of the Kriging model.

[0018] The model building module is used to establish an optimization design model of the noise reduction perforated plate structure with the final Kriging model as the objective function and the structural design parameters of the noise reduction perforated plate as the design variables.

[0019] The optimization design module is used to perform optimization calculations on the optimization design model to obtain the optimal structural design parameters.

[0020] This application also provides an electronic device, including:

[0021] Controller;

[0022] The memory stores multiple computer instructions, which are connected to the controller via a communication bus. These computer instructions are used to cause the controller to execute the methods described above.

[0023] The noise reduction orifice plate design method and system for the pressure regulating branch of a natural gas station in this application have the following advantages:

[0024] (1) An improved adaptive multi-scale Kriging model was established, which can obtain the noise value of the voltage regulation branch corresponding to any noise reduction orifice plate structure in the design domain with fewer computational samples. This can effectively reduce the computational cost of fluid calculation and acoustic simulation analysis in the process of noise reduction orifice plate optimization design.

[0025] (2) In the process of establishing an adaptive multi-scale Kriging model, an adaptive multi-scale sampling method is proposed, which can adaptively adjust the number of sampling samples according to the changes in model accuracy during the Kriging model training process, so as to achieve a balance between the modeling cost and modeling efficiency of the Kriging model, thereby improving the optimization efficiency of the noise reduction plate.

[0026] (3) Using the established noise reduction orifice plate noise value SPL-KG model as the objective function, an optimization design model for the noise reduction orifice plate structure is constructed. Through optimization and solution calculation, the optimized noise reduction orifice plate structure parameters can be obtained quickly, effectively improving the optimization efficiency of the noise reduction orifice plate in the pressure regulating branch of the natural gas station and meeting the rapid design requirements of the project for the noise reduction orifice plate. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0028] Figure 1 A flowchart illustrating a method for designing a noise reduction orifice plate for a pressure regulating branch in a natural gas station, as provided in this application embodiment.

[0029] Figure 2 This is a schematic diagram illustrating the interpolation process of establishing an SPL-KG model using an adaptive multi-scale sampling method, as provided in an embodiment of this application.

[0030] Figure 3 A schematic diagram of the composition of the noise reduction perforated plate design device provided in the embodiments of this application.

[0031] Figure 4 The results are simulations of the noise field of the voltage regulating branch obtained using the method described in this application.

[0032] Figure 5 The results are simulations of the noise field of the voltage regulation branch obtained using the existing EI method.

[0033] The following are the symbols in the attached diagram: 1. Controller; 2. Memory; 3. Communication bus. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] Figure 1 A flowchart illustrating a design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station, provided as an embodiment of this application. This application provides a design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station, including:

[0036] S1 sets the range of values ​​for various structural design parameters of the noise reduction orifice plate for the natural gas pressure regulating branch.

[0037] For example, the above structural design parameters include the aperture D, plate thickness t, and hole spacing s of the noise reduction perforated plate. Considering the machinability of the aperture and the installation dimensions, the existing structural dimensions of the noise reduction perforated plate are used as the initial values, and the range of values ​​for each design parameter is determined by the corresponding fluctuation range of the initial values, as shown in Table 1.

[0038] Table 1. Range of values ​​for each structural design parameter

[0039]

[0040] S2, within the range of values, randomly sample various structural design parameters to obtain n sets of sample points.

[0041] For example, n sets of sample points are represented as S(D) i , t i , s i ), where D i t i and s i Let i represent the aperture, plate thickness, and hole spacing, respectively, i = 1, 2, …, n. Since the number of design parameters for the noise reduction perforated plate structure is 3, the number of samples n = p(p+1), where p is the number of design parameters, then n = 12, meaning 12 sets of sample points S are obtained through sampling.

[0042] S3. The aerodynamic noise value of the pressure regulating branch after the noise reduction orifice plate is installed is used as the evaluation index. The acoustic simulation model of the pressure regulating branch is established using CFD software, and the aerodynamic noise value corresponding to n sample points is calculated.

[0043] For example, the aerodynamic noise values ​​corresponding to n sample points are represented as SPL. i .

[0044] S4, with n sets of sample points S(D) i , t i , s i ) and the corresponding aerodynamic noise value SPL i Using the initial sample points, a Kriging model is built based on the MATLAB toolbox. The error of the Kriging model is evaluated. If the error of the Kriging model does not meet the error requirements, an adaptive multi-scale sampling method is used to obtain new sample points S. new (D i , t i , s i ), calculate the new sample point S n (D i , t i , si The corresponding new aerodynamic noise value SPL ni , will the new sample point S n (D i , t i , s i ) and new aerodynamic noise value SPL ni The samples are added to the initial sample points, a new Kriging model is rebuilt, and the error is evaluated again until the new Kriging model meets the error requirements, thus obtaining the final Kriging model, which is referred to as SPL-KG. The adaptive multi-scale sampling method is to adaptively adjust the number of samples according to the changes in the accuracy of the Kriging model.

[0045] For example, error evaluation of the Kriging model includes:

[0046] The mean absolute relative error (MAPE) and the goodness of fit (R²) were used as error evaluation indicators, and the LOO (Leave-Out Cross-Validation) method was used to evaluate the error of the Kriging model.

[0047] Specifically, adaptive multi-scale sampling methods include:

[0048] Based on the mean square error sampling criterion, the sample point corresponding to the maximum error is taken as the new sample point, the new sample point is added to the initial sample point, the sample set is updated, and a new Kriging model is built based on the updated sample set.

[0049] Calculate the error index (MAPE) of the new Kriging model. new and R2 new By comparing the error index value of the new Kriging model with the previous error index value, the degree of improvement in the error index of the new Kriging model can be obtained: (MAPE) diff =|MAPE new -MAPE|,(R2) diff =|R2 new -R2|, where (MAPE) diff R² represents the improvement in the average absolute relative error. diff The degree of improvement in fit;

[0050] Determine whether the improvement in the error index meets the threshold condition: (R²) diff >R2_thr and (MAPE) diff >MAPE_thr, where R2_thr and MAPE_thr are the fit threshold and the mean absolute relative error threshold, respectively; if the threshold conditions are met, the multiple largest errors, specifically n iSample points P i Add it to the initial sample points; if the condition is not met, add the sample point P with the largest error. i Add initial sample points.

[0051] Among the above thresholds, the goodness-of-fit threshold and the mean absolute relative error threshold are R²_thr ≥ 0.9 and MAPE ≤ 5%, respectively. The sampling process of the adaptive multi-scale sampling method is as follows: Figure 2 As shown, using this sampling method, an SPL-KG model satisfying the error evaluation index was obtained through 26 interpolations. According to... Figure 2 The number of samples added was counted. A total of 45 new sample points were added in 26 interpolations. Combined with the 12 initial sample points, a final SPL-KG model that meets the accuracy requirements was obtained using 57 sample points. The final error of the model is: R2=0.984, MAPE=3.27%.

[0052] Furthermore, the threshold in the threshold condition is a dynamic value adjusted based on the values ​​of the mean absolute relative error (MAPE) and the goodness of fit (R²).

[0053] Specifically, the methods for adjusting the threshold include:

[0054] Based on the actual requirements for fitting accuracy during optimization design, the initial values ​​of the fitting degree threshold and the average absolute relative error threshold are set, and the dynamic threshold adjustment parameters α and β are set, where 0 < α < 1 and β > 1.

[0055] Determine whether the improvement in the goodness of fit in the error index is greater than the initial value of the goodness of fit threshold, and whether the improvement in the mean absolute relative error in the error index is greater than the initial value of the mean absolute relative error threshold, i.e., (R2). diff >R2_thr and (MAPE) diff >MAPE_thr, if so, tightens the initial values ​​of the fit threshold and mean absolute relative error threshold based on the dynamic threshold adjustment parameters: R2_thr=α(R2_thr) i MAPE_thr=α(MAPE_thr) i , where (R2_thr) i =0.05, (MAPE_thr) i =5%, α=0.9, β=1.1, otherwise relax the initial values ​​of the fit threshold and the mean absolute relative error threshold: R2_thr=β(R2_thr) i MAPE_thr = β(MAPE_thr) i .

[0056] In the embodiments of this application, when the degree of improvement of the error index meets the threshold condition, and multiple sample points with the largest errors are added to the initial sample points, the number of sample points n added to the initial sample points is... i for:

[0057] n i =round(k p ·((R2) diff +(MAPE) diff ))

[0058] Where round is the floor function, k p R² is the dynamic coefficient for the number of sample points. diff The degree of improvement in the error index is represented by the degree of improvement in the goodness of fit (MAPE). diff This refers to the improvement in the average absolute relative error within the degree of improvement of the error index.

[0059] Dynamic coefficient of sample point number k p Represented as:

[0060]

[0061] Where σY is the standard deviation of the current Kriging model's predictions, μY is the mean of the objective function, and n s α represents the total number of current sample points. min and α max Each is a pre-set k p The minimum and maximum values ​​are 0.001 and 0.01, respectively, in this embodiment of the application.

[0062] S5. Using the final Kriging model as the objective function, the structural design parameters of the noise reduction perforated plate as the design variables, and the hole spacing as the constraint condition (s≥1.5D), an optimization design model for the noise reduction perforated plate structure is established.

[0063] S6. Perform optimization calculations on the optimization design model to obtain the optimal structural design parameters S. opt (D opt , t opt , s opt ).

[0064] For example, the optimal structural design parameter obtained in the embodiments of this application is S. opt (12.4, 28.4, 19.5).

[0065] Simultaneously, the optimal structural design parameters obtained from this embodiment and the optimized structural parameters obtained using existing methods are substituted into the acoustic simulation model of the voltage regulating branch in step S3 to obtain the acoustic power cloud maps of the voltage regulating branch obtained using the method of this application and the existing optimization method, respectively.Figure 4 and Figure 5 As shown, the corresponding noise values ​​are 116.40dB and 118.41dB, respectively. The specific comparison results are shown in Table 2. Among them, the existing method is the Kriging model modeling method using the EI sampling strategy, which can be found in the literature "Aerodynamic and structural optimization design of wind turbine airfoils based on EI addition criterion and surrogate model".

[0066] Table 2 Comparison between existing methods and the method of this application

[0067]

[0068] As shown in Table 2, the optimized design parameters obtained using the natural gas pressure regulating branch noise reduction orifice plate optimization design method disclosed in this application not only slightly improve the noise reduction effect compared to existing methods (by 1.7%), but also significantly reduce the computational cost (by 41.23%). In other words, compared to existing methods, the natural gas pressure regulating branch noise reduction orifice plate optimization design method disclosed in this application can quickly obtain the optimized design parameters of the orifice plate structure, significantly reducing computation time and improving the efficiency of orifice plate structure optimization design.

[0069] This application also provides a design system for noise reduction orifice plates in the pressure regulating branch of a natural gas station, including:

[0070] The parameter setting module is used to set the value range of various structural design parameters of the noise reduction orifice plate of the natural gas pressure regulating branch. Taking into account the machinability of the orifice diameter and the installation size, the existing structural size of the noise reduction orifice plate is used as the initial value, and the value range of each design parameter is determined by the corresponding range of fluctuation above and below the initial value.

[0071] The random sampling module is used to randomly sample various structural design parameters within the range of values ​​to obtain n sets of sample points;

[0072] The modeling and simulation module is used to evaluate the aerodynamic noise value of the pressure regulating branch after the noise reduction orifice plate is installed. It uses CFD software to establish an acoustic simulation model of the pressure regulating branch and calculates the aerodynamic noise value corresponding to n sample points.

[0073] The evaluation and optimization module is used to establish a Kriging model using n sets of sample points and their corresponding aerodynamic noise values ​​as initial sample points. It then evaluates the error of the Kriging model. If the error of the Kriging model does not meet the error requirements, an adaptive multi-scale sampling method is used to obtain new sample points, calculate the corresponding new aerodynamic noise values, add the new sample points and new aerodynamic noise values ​​to the initial sample points, re-establish a new Kriging model, and perform error evaluation again. This process continues until the established new Kriging model meets the error requirements, resulting in the final Kriging model. The adaptive multi-scale sampling method adjusts the number of sampled samples adaptively based on changes in the accuracy of the Kriging model.

[0074] The model building module is used to establish an optimization design model of the noise reduction perforated plate structure with the final Kriging model as the objective function and the structural design parameters of the noise reduction perforated plate as the design variables.

[0075] The optimization design module is used to perform optimization calculations on the optimization design model to obtain the optimal structural design parameters.

[0076] like Figure 3 As shown in the illustration, this application also provides an electronic device, including:

[0077] Controller 1;

[0078] The memory 2 stores multiple computer instructions. The memory 2 is connected to the controller 1 via the communication bus 3. The multiple computer instructions are used to make the controller 1 execute the above-described method.

[0079] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0080] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A design method for noise reduction orifice plates in pressure regulating branches of natural gas stations, characterized in that, include: The range of values ​​for various structural design parameters of the noise reduction orifice plate for the natural gas pressure regulating branch is set. Taking into account the machinability of the orifice diameter and the installation dimensions, the existing structural dimensions of the noise reduction orifice plate are used as the initial values, and the range of values ​​for each design parameter is determined by the corresponding range of fluctuation above and below the initial values. Within the range of values, various structural design parameters are randomly sampled to obtain n sets of sample points; The aerodynamic noise value of the pressure regulating branch after the noise reduction orifice plate is installed is used as the evaluation index. The acoustic simulation model of the pressure regulating branch is established using CFD software, and the aerodynamic noise value corresponding to the n sample points is calculated. Using n sets of sample points and their corresponding aerodynamic noise values ​​as initial sample points, a Kriging model is established. The Kriging model is then evaluated for error. If the error of the Kriging model does not meet the error requirements, an adaptive multi-scale sampling method is used to obtain new sample points. The new aerodynamic noise values ​​corresponding to these new sample points are calculated, and the new sample points and their new aerodynamic noise values ​​are added to the initial sample points. A new Kriging model is then established, and the error evaluation is repeated until the established new Kriging model meets the error requirements, thus obtaining the final Kriging model. The adaptive multi-scale sampling method adaptively adjusts the number of sampled samples based on changes in the accuracy of the Kriging model. Using the final Kriging model as the objective function and the structural design parameters of the noise reduction perforated plate as the design variables, an optimization design model for the noise reduction perforated plate structure is established. The optimal structural design parameters are obtained by performing optimization calculations on the aforementioned optimization design model. The adaptive multi-scale sampling method includes: Based on the mean square error sampling criterion, the sample point corresponding to the maximum error is taken as the new sample point, the new sample point is added to the initial sample point, the sample set is updated, and the new Kriging model is established based on the updated sample set. Calculate the error index value of the new Kriging model, compare the error index value of the new Kriging model with the previous error index value, and obtain the degree of improvement of the error index of the new Kriging model. Determine whether the degree of improvement of the error index meets the threshold condition; if the threshold condition is met, add the multiple sample points with the largest error to the initial sample points; if the condition is not met, add the single sample point with the largest error to the initial sample points.

2. The design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station according to claim 1, characterized in that, The error evaluation of the Kriging model includes: The Kriging model was evaluated using the mean absolute relative error (MAPE) and the goodness of fit (R²) as error evaluation indicators, and the LOO cross-validation method was employed.

3. The design method for noise reduction orifice plates in the pressure regulating branch of a natural gas station according to claim 1, characterized in that, The threshold in the threshold condition is a dynamic value adjusted based on the values ​​of mean absolute relative error (MAPE) and goodness of fit (R²).

4. The design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station according to claim 3, characterized in that, The method for adjusting the threshold includes: Based on the actual requirements for fitting accuracy during optimization design, initial values ​​for the fitting degree threshold and the average absolute relative error threshold are set, and dynamic threshold adjustment parameters are set. Determine whether the degree of improvement in the error index is greater than the initial value of the degree of improvement threshold, and whether the degree of improvement in the mean absolute relative error is greater than the initial value of the mean absolute relative error threshold. If so, tighten the initial values ​​of the degree of improvement threshold and the mean absolute relative error threshold according to the dynamic threshold adjustment parameter; otherwise, relax the initial values ​​of the degree of improvement threshold and the mean absolute relative error threshold.

5. The design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station according to claim 1, characterized in that, When the improvement of the error index meets the threshold condition, and multiple sample points with the largest errors are added to the initial sample points, the number of sample points n added to the initial sample points is... i for: n i =round(k p ·((R2) diff +(MAPE) diff )) Where round is the floor function, k p R² is the dynamic coefficient for the number of sample points. diff The degree of improvement in the error index is defined as the degree of improvement in the goodness of fit (MAPE). diff The improvement degree of the average absolute relative error in the aforementioned error index improvement degree.

6. The design method for a noise reduction orifice plate in a pressure regulating branch of a natural gas station according to claim 5, characterized in that, The dynamic coefficient of the number of sample points k p Represented as: Where σY is the standard deviation of the current prediction mean square error of the Kriging model, μY is the mean of the objective function, and n s α is the total number of current sample points. min and α max Each is a pre-set k p The minimum and maximum values.

7. A system applying the design method for noise reduction orifice plates in the pressure regulating branch of a natural gas station according to any one of claims 1-6, characterized in that, include: The parameter setting module is used to set the value range of various structural design parameters of the noise reduction orifice plate of the natural gas pressure regulating branch. Taking into account the machinability of the orifice diameter and the installation size, the existing structural size of the noise reduction orifice plate is used as the initial value, and the value range of each design parameter is determined by the corresponding range of fluctuation above and below the initial value. The random sampling module is used to randomly sample various structural design parameters within the range of the values ​​to obtain n sets of sample points; The modeling and simulation module is used to establish an acoustic simulation model of the pressure regulating branch with the aerodynamic noise value of the pressure regulating branch after the noise reduction orifice plate is installed as the evaluation index, and to calculate the aerodynamic noise value corresponding to the n sets of sample points using CFD software. The evaluation and optimization module is used to establish a Kriging model using n sets of sample points and the corresponding aerodynamic noise values ​​as initial sample points, evaluate the error of the Kriging model, and if the error of the Kriging model does not meet the error requirements, an adaptive multi-scale sampling method is used to obtain new sample points, calculate the new aerodynamic noise values ​​corresponding to the new sample points, add the new sample points and the new aerodynamic noise values ​​to the initial sample points, re-establish a new Kriging model, and perform error evaluation again until the established new Kriging model meets the error requirements, thus obtaining the final Kriging model; wherein, the adaptive multi-scale sampling method adaptively adjusts the number of sampling samples according to the change in the accuracy of the Kriging model; The model building module is used to establish an optimization design model of the noise reduction perforated plate structure, with the final Kriging model as the objective function and the structural design parameters of the noise reduction perforated plate as the design variables. The optimization design module is used to perform optimization calculations on the optimization design model to obtain the optimal structural design parameters.

8. An electronic device, characterized in that, include: Controller (1); The memory (2) stores multiple computer instructions, and the memory (2) is connected to the controller (1) via a communication bus (3). The multiple computer instructions are used to cause the controller (1) to execute the method described in any one of claims 1-6.

Images