A method and device for optimizing the robustness of natural vibration characteristics of pipelines
Through multiplication type dimensionality reduction M-DRM method and sequential Kriging agent model, the clamp position is optimized, and the robustness problem under the influence of uncertainty in the design of natural vibration characteristics of aero engine pipelines is solved, and the average of natural frequency of pipelines is realized. Improvement and reduction of dispersion reduces computational costs and enhances dynamic robustness.
Patent Information
- Application Number
- CN202411148315.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-21
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-21
AI Technical Summary
The prior art fails to fully consider uncertainty factors when optimizing the position of the clamp of the aero engine pipeline, resulting in statistically dispersive results of the inherent vibration characteristics of the pipeline, high calculation costs, and lacks efficient and robust optimization design methods.
Using multiplication dimensionality reduction M-DRM method and sequential Kriging agent model optimization method, the clamp position is optimized to maximize the pipeline natural frequency mean and minimize its dispersion through parameterized finite element modeling and uncertainty analysis. , build a robust optimization objective function and a Kriging agent model, and combine an efficient global optimization algorithm to achieve robust optimization of the inherent vibration characteristics of the pipeline.
The stability of the inherent vibration characteristics of the pipeline to external fluctuations is improved, the calculation cost of uncertainty analysis is reduced, the optimization design efficiency of dynamic robustness is enhanced, and the dimensional disaster problem of statistical moment calculation in traditional methods is solved.
Smart Images

Figure CN118673628B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of aviation engine engineering, and in particular relates to a method and a device for optimizing the robustness of natural vibration characteristics of a pipeline. Background Art
[0002] Aircraft engine pipelines are mainly used to transport fuel, lubricating oil, air and other media, and are an important part of the engine's external system. Most pipelines are installed on the surface of the engine casing through structures such as clamps and brackets. When the natural frequency of the pipeline system is the same or similar to the engine excitation frequency, it will cause large structural vibrations, resulting in pipeline oil seepage, leakage, breakage and other faults. The clamp is the connecting support structure between the pipeline and the casing, and its installation position has a great influence on the dynamic characteristics of the pipeline system. For this reason, carrying out the clamp position design on the basis of completing the pipeline obstacle avoidance and connection laying function design is an effective method to adjust the pipeline dynamic characteristics.
[0003] The existing traditional clamp position optimization design method is a deterministic optimization design method that does not fully consider the influence of various uncertain factors faced in actual situations. In the actual external piping system, the clamp stiffness, material parameters and clamp support position fluctuate around the nominal value due to uncertain factors such as installation deviation and pipe properties, which is noise. Under the influence of the above noise parameters, the dynamic design results of the inherent vibration characteristics of the piping system have statistical dispersion. It is urgent to carry out the dynamic robustness optimization design research of the inherent vibration characteristics of the engine external piping system under the influence of uncertain factors.
[0004] The robustness optimization of the natural vibration characteristics of the piping system is based on the statistical results of the mean and variance of the natural frequency of the piping. A new round of system uncertainty analysis needs to be carried out during each iterative update of the design parameters. The natural vibration of the piping system involves complex and time-consuming finite element calculations. A large amount of system uncertainty re-analysis means a high computational cost for the robustness optimization of the external piping dynamics. Therefore, an efficient robustness optimization design method that can regulate the natural vibration characteristics of the piping system is urgently needed. Summary of the invention
[0005] In response to the shortcomings of the prior art, the present application proposes a method and device for optimizing the robustness of the natural vibration characteristics of pipelines, which actively considers the uncertainty factors affecting the natural vibration characteristics of pipelines during the design stage, and optimizes the clamp position to maximize the mean of the natural frequency of the pipeline and minimize the dispersion, thereby achieving the effect of suppressing uncertainty factors such as installation deviation, pipe material dispersion, dimensional tolerance, etc., thereby improving the stability of the natural vibration characteristics of the pipeline against external fluctuation factors.
[0006] In a first aspect, the present application proposes a method for optimizing the robustness of natural vibration characteristics of a pipeline, comprising:
[0007] Step S1: performing parametric finite element modeling on the pipeline system outside the engine, and establishing a clamp-pipeline natural vibration characteristic analysis model;
[0008] Step S2: Determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness, and the clamp position assembly deviation;
[0009] Step S3: construct a robustness optimization objective function according to the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and establish a robustness optimization model for the natural vibration characteristics of the pipeline;
[0010] Step S4: performing Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples, and based on the clamp-pipeline natural vibration characteristic analysis model, using the clamp position samples as input, using the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculating the mean and variance of the first-order natural frequency of the corresponding pipeline system;
[0011] Step S5: Substitute the value of the robustness optimization objective function calculated in step S3 according to the mean and variance of the first-order natural frequency of the corresponding pipeline system, and form an initial sample data set with the clamp position sample and the value of the robustness optimization objective function calculated;
[0012] Step S6: establishing a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtaining the mean and variance predicted by the Kriging proxy model;
[0013] Step S7: Based on the mean and variance predicted by the Kriging surrogate model, an efficient global optimization strategy based on the expected lift function is constructed;
[0014] Step S8: Determine whether the iteration termination condition of the efficient global optimization strategy is met;
[0015] Step S9: If the iteration termination condition of the efficient global optimization strategy is met, go to step S10; if the iteration termination condition of the efficient global optimization strategy is not met, add the new clamp position sample to the initial sample data set, repeat steps S4 to S9 in sequence, and update the prediction accuracy of the Kriging proxy model;
[0016] Step S10: Outputting the clamp position with the largest mean and the smallest dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
[0017] The robustness optimization objective function is constructed according to the mean and variance of the first-order natural frequency of the pipeline system, the clamp position is used as the optimization variable, and the constraint conditions of the clamp position are determined to establish the robustness optimization model of the pipeline natural vibration characteristics. The calculation formula is as follows:
[0018]
[0019] in, is the mean of the first-order natural frequency of the clamp at its initial position, is the standard deviation of the first-order natural frequency at the initial position of the clamp, that is, the square root of the variance of the first-order natural frequency; is the mean of the first-order natural frequency corresponding to the optimized variable x, The standard deviation of the first-order natural frequency corresponding to the optimized variable x is the square root of the variance of the first-order natural frequency corresponding to x; is the upper limit of the optimization variable, is the lower bound of the optimization variable, is the weight coefficient of the mean, is the weight coefficient of the variance.
[0020] The clamp-pipeline natural vibration characteristic analysis model uses the clamp position sample as input, uses the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculates the mean and variance of the first-order natural frequency of the corresponding pipeline system, including:
[0021] The clamp position is used as the optimization variable , pipe material properties, clamp stiffness and clamp position assembly deviation are noise variables , the natural vibration characteristic function of the pipeline system is defined as ;
[0022] The M-DRM method is used to express the natural vibration characteristic function of the pipeline system as a multiplication of multiple one-dimensional single variable sub-functions:
[0023]
[0024] in, is defined as a univariate function of the ith variable, , is the anchor vector, and the mean of the noise variable is taken. is the anchor point value of the nth noise variable, n is the dimension of the input noise variable, that is, the dimension of the random variable;
[0025] According to the form of multiplication of multiple one-dimensional univariate sub-functions, the mean and second-order moment of the one-dimensional univariate sub-functions are solved by numerical integration method:
[0026]
[0027] in, is the mean of the one-dimensional univariate subfunction, is the second-order moment of a one-dimensional univariate subfunction, is the kth random variable The probability density function of is the expected operation of the kth subfunction, is the anchor vector that does not contain the k-th anchor value, is the kth one-dimensional univariate subfunction.
[0028] The Gaussian quadrature method is used to solve the mean and second-order moment of one-dimensional univariate subfunctions;
[0029]
[0030] in, is the lth Gaussian node, is the lth Gaussian quadrature coefficient, N represents the number of Gaussian integration points;
[0031] According to the mean and second-order moment of the one-dimensional univariate subfunction, the mean and variance of the first-order natural frequency of the pipeline system are obtained:
[0032]
[0033] in, is the mean of the first-order natural frequency corresponding to the optimized variable x, To optimize the variance of the first-order natural frequency corresponding to the variable x, is the expectation operation of the natural vibration characteristics of the pipeline, k is the kth one-dimensional single variable subfunction, and n is the number of dimensions of the noise variable.
[0034] The mean and variance of the first-order natural frequency of the corresponding pipeline system are substituted into step S3 to calculate the value of the robustness optimization objective function, and the calculation formula is as follows:
[0035]
[0036] in, is the mean of the first-order natural frequency corresponding to the optimized variable x, To optimize the variance of the first-order natural frequency corresponding to the variable x, is the mean of the one-dimensional univariate subfunction, is the second-order moment of a one-dimensional univariate subfunction, is the mean of the first-order natural frequency of the clamp at its initial position, is the variance of the first-order natural frequency of the clamp’s initial position, is the weight coefficient of the mean, is the weight coefficient of the variance, for The exponential operation of 1-n, for The 2-n power exponential operation is performed, where n is the dimension of the random variable, k is the kth random variable, c is the anchor point vector, and the mean of the noise variable is taken.
[0037] According to the initial sample data set, a Kriging proxy model based on the clamp position and robustness optimization objective function is established, and the mean and variance predicted by the Kriging proxy model are obtained, and the calculation formula is as follows:
[0038]
[0039] in, , is the mean predicted by the Kriging surrogate model, is the variance predicted by the Kriging surrogate model, is a polynomial regression model, is the basis function of the Kriging model, , is the regression coefficient, is the regression coefficient The generalized least squares estimate of is the random process variance of the Kriging surrogate model, the input variable x is the clamp position, is the response of the objective function corresponding to pipeline robustness optimization, For the clamp position sample The correlation matrix of To predict the clamp position and the correlation matrix between the existing clamp samples, are the relevant model parameters, is the number of sample points.
[0040] The method constructs an efficient global optimization strategy based on the expected lift function based on the mean and variance predicted by the Kriging surrogate model, including:
[0041] Determine the point addition criterion of the expected improvement function according to the expected value of the improvement amount of the objective function;
[0042] Among them, the improvement of the objective function is defined as:
[0043]
[0044] in, is the improvement of the objective function, the input variable x is the clamp position, The predicted value of the objective function for pipeline robustness optimization at any clamp position is subject to normal distribution. is the objective function value of the optimal pipeline robustness optimization corresponding to the current clamp position sample;
[0045] Based on the mean and variance predicted by the Kriging surrogate model, the expected lift function is obtained, and the calculation formula is as follows:
[0046]
[0047] in, is the cumulative distribution function of the standard normal distribution, is the probability density function of the standard normal distribution, is the mean predicted by the Kriging surrogate model, is the standard deviation of the Kriging surrogate model prediction, that is, the square root of the variance predicted by the Kriging surrogate model, is the expected improvement function;
[0048] Construct an efficient global optimization strategy based on the expected improvement function. The calculation formula is as follows:
[0049]
[0050] in, is the upper limit of the optimization variable, To optimize the lower limit of the variable, the iterative termination condition of the global optimization strategy is ,at this time Optimize the optimal solution for clamp position robustness.
[0051] In a second aspect, the present application proposes a robustness optimization device for the natural vibration characteristics of an external pipeline, comprising:
[0052] The finite element model building module is used to perform parametric finite element modeling on the pipeline system outside the engine and establish a clamp-pipeline natural vibration characteristic analysis model;
[0053] The performance index determination module is used to determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness and the clamp position assembly deviation;
[0054] The robustness optimization model building module is used to build a robustness optimization objective function based on the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and build a robustness optimization model for the natural vibration characteristics of the pipeline;
[0055] The uncertainty analysis module is used to perform Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples. Based on the clamp-pipeline natural vibration characteristic analysis model, the clamp position samples are used as input, and the M-DRM method is used to perform natural vibration characteristic uncertainty analysis on the noise variable to calculate the mean and variance of the first-order natural frequency of the corresponding pipeline system;
[0056] The initial sample data set establishment module is used to substitute the mean and variance of the first-order natural frequency of the corresponding pipeline system into the robustness optimization model establishment module, calculate the value of the robustness optimization objective function, and form the initial sample data set with the clamp position sample and the calculated value of the robustness optimization objective function;
[0057] A data prediction module, used to establish a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtain the mean and variance predicted by the Kriging proxy model;
[0058] The global optimization strategy module is used to construct an efficient global optimization strategy based on the expected lift function based on the mean and variance predicted by the Kriging surrogate model;
[0059] The termination condition judgment module is used to judge whether the iterative termination condition of the efficient global optimization strategy is met;
[0060] The optimization iteration module is used to transfer to the optimization result output module if the iteration termination condition of the efficient global optimization strategy is met; if the iteration termination condition of the efficient global optimization strategy is not met, the new clamp position sample is added to the initial sample data set, and the uncertainty analysis module ~ optimization iteration module are repeated in sequence to update the prediction accuracy of the Kriging proxy model;
[0061] The optimization result output module is used to output the clamp position with the maximum mean and the minimum dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
[0062] In a third aspect, the present application proposes an electronic device, comprising: one or more processors, and a memory, wherein the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the robustness optimization method of the inherent vibration characteristics of the external pipeline.
[0063] In a fourth aspect, the present application proposes a computer-readable storage medium storing executable instructions, which, when executed, enable a processor to execute the robustness optimization method of the natural vibration characteristics of an external pipeline.
[0064] In a fifth aspect, the present application proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the robustness optimization method of the inherent vibration characteristics of the external pipeline.
[0065] The present application provides a robustness optimization method and device for the natural vibration characteristics of an external pipeline, which has the following beneficial effects compared with the prior art:
[0066] 1. This application combines the multiplicative dimensionality reduction M-DRM method with the sequential Kriging surrogate model optimization method, takes the maximum mean and minimum variance of the natural vibration characteristics of the external pipeline under the influence of uncertain factors as the optimization goal, optimizes the position of the clamp in the pipeline system, and then improves the mean of the first-order natural frequency of the pipeline system and reduces its dispersion, enhances the stability of the natural vibration characteristics of the pipeline, and provides an important reference for the dynamic robustness optimization design and application of complex pipeline systems under uncertain working conditions.
[0067] 2. This application adopts a single variable M-DRM method based on N Gaussian integration points to reduce the number of calculations of the natural vibration characteristic moment integral of the n-dimensional pipeline system from Reduce to , fundamentally solving the dimensionality curse problem of statistical moment calculation. It overcomes the shortcomings of traditional uncertainty analysis methods (such as Monte Carlo method), large sampling, low calculation efficiency, etc., and greatly improves the efficiency of uncertainty analysis of inherent vibration characteristics of pipeline systems.
[0068] 3. This application establishes a Kriging proxy model of the clamp position optimization variables and the pipeline inherent vibration characteristics robustness optimization objective function, constructs the expected improvement (EI) function of the optimization variable update to sequentially update the Kriging model prediction accuracy, and combines the efficient global optimization (EGO) method to achieve global optimization of the clamp position. The resulting sequential Kriging model optimization method can effectively reduce the huge computational cost caused by the global heuristic algorithm (such as genetic algorithm) frequently calling the pipeline inherent vibration characteristics analysis model, thereby further improving the efficiency of optimizing the robustness of the inherent vibration characteristics of the pipeline system. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 A flow chart of a method for optimizing the robustness of natural vibration characteristics of a pipeline according to an embodiment of the present application;
[0070] Figure 2 A schematic diagram of the engine clamp-pipeline layout of an embodiment of the present application;
[0071] Figure 3 A comparison diagram of the probability density distribution function of the first-order natural frequency of the embodiment of the present application;
[0072] Figure 4A schematic diagram of a robustness optimization device for the inherent vibration characteristics of an external pipeline according to an embodiment of the present application. DETAILED DESCRIPTION
[0073] The specific implementation of the present application is further described in detail below in conjunction with the drawings and examples.
[0074] In view of the problems of the prior art, the present application actively considers the uncertainty factors that affect the natural vibration characteristics of the pipeline system during the design stage, optimizes the position of the clamp to increase the mean of the natural frequency of the pipeline system and reduce its dispersion, thereby improving the robustness of the natural vibration characteristics of the pipeline against external fluctuation factors, and aims to provide a robustness optimization method for the natural vibration characteristics of the external pipeline based on the adjustment of the clamp position suitable for uncertain working conditions. The present application combines the multiplicative dimensional reduction method (M-DRM) uncertainty analysis method, the Kriging surrogate model and the efficient global optimization (EGO) algorithm, and proposes a robustness optimization method for the natural vibration characteristics of the pipeline based on the combination of M-DRM and the sequential Kriging surrogate model. The mean and variance of the first-order natural frequency of the pipeline system in the robustness optimization model of the natural vibration characteristics of the pipeline are obtained by uncertainty analysis based on the M-DRM method, and the optimization part adopts the optimization algorithm based on the sequential Kriging surrogate model for optimization. By optimizing the clamp position, the mean of the natural frequency of the piping system is increased and its dispersion is reduced, thereby enhancing the robustness of the natural vibration characteristics of the pipeline; at the same time, the number of model calls in the robustness optimization process is reduced, thereby improving the efficiency of the robustness optimization of the natural vibration characteristics of the pipeline.
[0075] Embodiment 1:
[0076] This embodiment proposes a robustness optimization method for the natural vibration characteristics of an external pipeline, such as Figure 1 As shown, including:
[0077] Step S1: performing parametric finite element modeling on the pipeline system outside the engine, and establishing a clamp-pipeline natural vibration characteristic analysis model;
[0078] This embodiment combines the attached Figure 2 The pipe-clamp system shown is used to further implement the present application. A section of a space pipe of an aircraft engine in actual engineering is taken as the research object. The pipe section is composed of straight pipes, bends, port joints and intermediate joints, and both ends of the pipe are fixed on the casing accessories. The pipe material is 1Cr18Ni9Ti stainless steel with a density of 7850kg / m 3, elastic modulus is 204GPa, Poisson's ratio is 0.3, pipeline diameter is 18mm, turning radius is 40mm, wall thickness is 1mm, and the pipeline passes through 5 clamps of the 3rd, 5th, 7th, 9th and 10th straight pipe sections Fixed on the receiver, the pipeline has 16 control points , and its specific coordinate positions are shown in Table 1.
[0079] Table 1 Coordinates of pipeline control points
[0080]
[0081] The engine's external piping system is parametrically modeled by finite element analysis. Pipeline information parameters such as pipeline control point coordinates, turning radius, and pipeline geometry are extracted. Pipe joints are equivalently simplified, and clamps are simplified into spring units. APDL parametric finite element modeling of the piping system is performed based on beam units. Parametric design can modify the original clamp-pipe system model at any time, making it easier to analyze the inherent vibration characteristics of the piping system under different design parameters.
[0082] Step S2: Determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness, and the clamp position assembly deviation;
[0083] In this embodiment, according to the requirements of the existing pipeline system vibration design criteria (GJB3054-97), the installation of the clamp should make the natural frequency of the pipeline system far away from the operating frequency of the high and low pressure rotors of the engine, so that it will not be damaged by resonance even under strong excitation conditions. Therefore, the first-order natural frequency of the pipeline system is determined as a robust performance design indicator, and the first-order natural frequency of the pipeline is increased and its dispersion is reduced by adjusting the position of the clamp to achieve the purpose of robust performance design. Figure 2 The natural vibration characteristics of the piping system shown in the figure are subjected to robust optimization design. Noise factors such as pipe material properties, clamp installation position deviation and support stiffness that may have a significant impact on the natural vibration characteristics of the piping system are identified, and appropriate description methods are selected for different types of noise variables, as well as the optimization interval of the design variable clamp position. The specific parameters are shown in Table 2, where parameter 1 and parameter 2 of the design variable correspond to the upper and lower intervals of the clamp position optimization variable proportional coefficient, respectively, and parameter 1 and parameter 2 of the noise variable are the mean and coefficient of variation corresponding to each variable. In the actual assembly process of the pipeline-clamp system, there will inevitably be assembly errors in the clamp position. Considering that the assembly deviation of the clamp center position is ±5mm, the clamps move in their respective straight pipe sections. Therefore, the range of motion of the optimized variable clamp support position is the length of the straight pipe section. In order to consider the actual operation process of the clamp assembly, an assembly gap of t is left at each end point of the straight pipe section during the design phase. To facilitate the representation of the position of the clamp during the optimization process, it is assumed that is the proportional coefficient of the straight pipe section where the clamp is located, then the coordinates of the center position of the clamp can be expressed as:
[0084] (1)
[0085] in, is the center coordinate of the clamp, t is the assembly clearance, in this embodiment , , They are the starting points of the straight pipe section where the clamp is located. and end point Therefore, in the clamp optimization process, only the proportional coefficient of the straight pipe section where the clamp is located needs to be optimized. , then the corresponding clamp coordinate position can be obtained through the proportional coefficient and the vector conversion relationship between the initial and final control points.
[0086] Step S3: construct a robustness optimization objective function according to the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and establish a robustness optimization model for the natural vibration characteristics of the pipeline;
[0087] In this embodiment, the loss function model of the first-order natural frequency maximum characteristic is determined according to the performance index of the natural vibration characteristic of the pipeline, that is, by adjusting the clamp position to increase the pipeline fundamental frequency and reduce the dispersion of the first-order natural frequency, so as to achieve the robustness optimization of the natural vibration characteristics of the pipeline system. The objective function of the robustness optimization of the natural vibration characteristics of the pipeline is constructed according to the mean and variance of the first-order natural frequency of the pipeline system. With the optimization goal of maximizing the mean and minimizing the variance of the first-order natural frequency of the pipeline system, the constraint condition of the design variable clamp position, combined with the response value of the first-order natural frequency of the pipeline initially designed by the clamp position, the robustness optimization model of the natural vibration characteristics of the pipeline is established. The robustness optimization model of the natural vibration characteristics of the pipeline includes: the objective function and constraint conditions of the robustness optimization of the natural vibration characteristics, and the calculation formula is as follows:
[0088] (2)
[0089] in, is the mean of the first-order natural frequency of the clamp at its initial position, is the standard deviation of the first-order natural frequency at the initial position of the clamp, that is, the square root of the variance of the first-order natural frequency; is the mean of the first-order natural frequency corresponding to the optimized variable x, The standard deviation of the first-order natural frequency corresponding to the optimized variable x is the square root of the variance of the first-order natural frequency corresponding to x; is the upper limit of the optimization variable, is the lower bound of the optimization variable, is the weight coefficient of the mean, is the weight coefficient of the variance. In this embodiment, , because this embodiment takes 5 clamp positions, namely: , , , , , so its upper and lower limits are expanded as follows: .
[0090] Table 2 Basic parameters of space pipelines and their distribution
[0091]
[0092] Step S4: performing Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples, and based on the clamp-pipeline natural vibration characteristic analysis model, using the clamp position samples as input, using the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculating the mean and variance of the first-order natural frequency of the corresponding pipeline system;
[0093] In this embodiment, the clamp-pipeline natural vibration characteristic analysis model uses the clamp position sample as input, uses the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculates the mean and variance of the first-order natural frequency of the corresponding pipeline system, including:
[0094] Step S4.1: Use the clamp position as the optimization variable , pipe material properties, clamp stiffness and clamp position assembly deviation are noise variables , the natural vibration characteristic function of the pipeline system is defined as ;
[0095] In this embodiment, in order to carry out uncertainty analysis of the natural vibration characteristics of the pipeline, the natural vibration characteristics of the system are defined as ,in Optimize variables for 5 clamp positions in the pipeline, and are 10 independent noise random variables including the uncertainties of clamp position deviation, support stiffness and material properties.
[0096] Step S4.2: Using the M-DRM method, the natural vibration characteristic function of the pipeline system is expressed as a multiplication of multiple one-dimensional single variable sub-functions:
[0097] In this embodiment, the M-DRM method is used to decompose the high-dimensional problem of pipeline natural vibration characteristics into multiple low-dimensional problems in the form of product. Here, only the M-DRM in the form of single-variable sub-function is considered, and the functional relationship between the original pipeline system natural vibration characteristics and the input variables is approximately:
[0098] (3)
[0099] in, is defined as a univariate function of the ith variable, , is the anchor vector, and the mean of the noise variable is taken. is the anchor point value of the nth noise variable, n is the dimension of the input noise variable, that is, the dimension of the random variable;
[0100] Step S4.3: according to the multiplication form of multiple one-dimensional univariate sub-functions, the mean and second-order moment of the one-dimensional univariate sub-functions are solved by numerical integration method;
[0101] In the prior art, the calculation of statistical moments (mean and variance) for robust optimization involves high-dimensional integration problems, which requires a lot of model reanalysis and has high computational costs. The present application uses the M-DRM method in robust optimization to perform uncertainty analysis to obtain the mean and variance of the inherent vibration characteristics of the pipeline, overcome the dimensionality curse of high-dimensional integral calculations of structural response statistical moments, and achieve efficient calculation of uncertainty analysis of the inherent vibration characteristics of the pipeline. The advantage of using the M-DRM method is that the expected operation of the original high-dimensional inherent vibration characteristic function can be neatly separated in a single variable integral, and the mean and second-order moment of the kth one-dimensional function can be expressed as and , defined as:
[0102] (4)
[0103] in, is the mean of the one-dimensional univariate subfunction, is the second-order moment of a one-dimensional univariate subfunction, is the kth random variable The probability density function of is the expected operation of the kth subfunction, is the anchor vector that does not contain the k-th anchor value, is the kth one-dimensional univariate subfunction.
[0104] Step S4.4: using Gaussian quadrature method to solve the mean and second-order moment of the one-dimensional univariate subfunction;
[0105] In this embodiment, the single variable integral of the decomposed one-dimensional function is solved using the Gaussian quadrature formula, and the calculation formula for the numerical integral is:
[0106] (5)
[0107] in, is the lth Gaussian node, N represents the number of Gaussian integration points, is the lth Gaussian quadrature coefficient, and its specific value is determined by the probability distribution type of the kth noise variable. Generally speaking, when the univariate M-DRM method and Gaussian integration method are used to estimate the output statistical moment of the pipeline natural vibration characteristics, the number of model reanalysis required is , where n is the dimension of the random variable and N is the number of Gaussian integration points.
[0108] Step S4.5: According to the mean and second-order moment of the one-dimensional univariate subfunction, the mean and variance of the first-order natural frequency of the pipeline system are obtained.
[0109] In this embodiment, the multivariate performance function of the natural vibration characteristics of the pipeline can be expressed as the product of multiple one-dimensional functions, and then the mean and variance of the natural vibration characteristics Y of the pipeline system are approximately:
[0110] (6)
[0111] in, is the mean of the first-order natural frequency corresponding to the optimized variable x, To optimize the variance of the first-order natural frequency corresponding to the variable x, is the expected operation of the pipeline's natural vibration characteristics, k is the kth one-dimensional single variable subfunction, and n is the number of dimensions of the noise variable. The number of finite element calls required to perform an uncertainty analysis of the pipeline vibration response using the single variable M-DRM method is , in the case of normally distributed random variables, the five-point Gauss-Hermite quadrature formula can provide accurate numerical integration results. Therefore, compared with traditional uncertainty analysis methods (such as MCS), the M-DRM uncertainty analysis method based on N-point Gaussian quadrature can achieve the calculation results of the mean and variance of the performance response based on a small number of model re-analysis, fundamentally overcoming the dimensionality curse of the statistical feature calculation of the natural vibration characteristics of the pipeline, and laying the foundation for realizing efficient uncertainty analysis of the natural vibration characteristics of the pipeline.
[0112] Step S5: Substitute the value of the robustness optimization objective function calculated in step S3 according to the mean and variance of the first-order natural frequency of the corresponding pipeline system, and form an initial sample data set with the clamp position sample and the value of the robustness optimization objective function calculated;
[0113] Based on the calculation results of the mean and variance of the pipeline natural vibration characteristics given by the above univariate M-DRM method, the objective function of the robustness optimization of the external pipeline natural vibration characteristics can be expressed as:
[0114]
[0115] in, is the mean of the first-order natural frequency corresponding to the optimized variable x, is the mean of the one-dimensional univariate subfunction, is the second-order moment of a one-dimensional univariate subfunction, is the mean of the first-order natural frequency of the clamp at its initial position, is the variance of the first-order natural frequency of the clamp’s initial position, is the weight coefficient of the mean, is the weight coefficient of the variance, for The exponential operation of 1-n, for The 2-n power exponential operation is performed, where n is the dimension of the random variable, k is the kth random variable, c is the anchor point vector, and the mean of the noise variable is taken.
[0116] Step S6: establishing a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtaining the mean and variance predicted by the Kriging proxy model;
[0117] In this embodiment, the sample of the clamp position is And the corresponding pipeline robustness optimization objective function value The initial sample data set is formed, where q is the number of sample points. Based on the above sample set, the initial Kriging proxy model between the clamp position optimization variable and the robust optimization objective function in the robust optimization model of the pipeline natural vibration characteristics is established, and the Kriging model prediction mean and variance of the objective function are obtained as follows:
[0118] (7)
[0119] in, , is the mean predicted by the Kriging surrogate model, is the variance predicted by the Kriging surrogate model, is a polynomial regression model, is the basis function of the Kriging model, , is the regression coefficient, is the regression coefficient The generalized least squares estimate of is the random process variance of the Kriging surrogate model, the input variable x is the clamp position, is the response of the objective function corresponding to pipeline robustness optimization, For the clamp position sample The correlation matrix of To predict the clamp position and the correlation matrix between the existing clamp samples, are the relevant model parameters, is the number of sample points.
[0120] Step S7: Based on the mean and variance predicted by the Kriging surrogate model, an efficient global optimization strategy based on the expected lift function is constructed, including:
[0121] In this embodiment, an EI point addition criterion based on the Kriging model of the objective function of the robustness optimization of the pipeline natural vibration characteristics is constructed, and a new clamp position that makes the objective function smaller is found through the EGO algorithm. Assume that the optimal pipeline vibration response robustness index objective function value corresponding to the current clamp position sample is , and the prediction of the objective function of the robust optimization of the pipeline natural vibration characteristics at any clamp position follows a normal distribution , then the corresponding objective function improvement Defined as:
[0122] (8)
[0123] in, is the improvement of the objective function, the input variable x is the clamp position, The predicted value of the objective function for pipeline robustness optimization at any clamp position is subject to normal distribution. is the objective function value of the optimal pipeline robustness optimization corresponding to the current clamp position sample;
[0124] The expected value of is:
[0125] (9)
[0126] in, is the cumulative distribution function of the standard normal distribution, is the probability density function of the standard normal distribution, is the mean predicted by the Kriging surrogate model, is the standard deviation of the Kriging surrogate model prediction, that is, the square root of the variance predicted by the Kriging surrogate model, is the expected improvement function. The EI function is the core of the sequential Kriging model optimization method and determines the optimization mechanism of the EGO algorithm. From a theoretical perspective, the EI function realizes the sequential update of the Kriging proxy model by continuously balancing the smaller objective function value and the prediction results with greater uncertainty, ensuring that the clamp position optimization process can integrate local search and global optimization. Based on this, an efficient global optimization strategy based on the expected improvement function is constructed, and the calculation formula is as follows:
[0127] (10)
[0128] in, is the upper limit of the optimization variable, To optimize the lower limit of the variable, the iterative termination condition of the global optimization strategy is ,at this time Optimize the optimal solution for clamp position robustness.
[0129] Step S8: Determine whether the iteration termination condition of the efficient global optimization strategy is met;
[0130] Step S9: If the iteration termination condition of the efficient global optimization strategy is met, go to step S10; if the iteration termination condition of the efficient global optimization strategy is not met, add the new clamp position sample to the initial sample data set, repeat steps S4 to S9 in sequence, and update the prediction accuracy of the Kriging proxy model;
[0131] It can be understood that when a new clamp position sample is added to the initial sample data set, there is no need to perform Latin hypercube sampling on multiple clamp positions within a preset range again in the process of repeating step S4 to obtain the clamp position sample, because the first iteration is based on the initial sample of the clamp position obtained by Latin hypercube sampling, so after the initial sample data set, there is no need to perform Latin hypercube sampling again during the iteration process.
[0132] Step S10: Outputting the clamp position with the largest mean and the smallest dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
[0133] In this embodiment, the obtained robust optimization of the natural vibration characteristics of the pipeline is compared and analyzed with the traditional deterministic optimization. The mean, standard deviation and coefficient of variation of the first-order natural frequency of the pipeline system before and after optimization are shown in Table 3. It can be seen that the robust optimization ensures that the mean of the first-order natural frequency is increased by 16.48Hz (+9.01%) compared with the initial clamp position result. The standard deviation of the first-order natural frequency response is reduced by 2.25Hz (-15.72%) compared with the deterministic optimization, and the coefficient of variation is reduced by 0.88 (12.70%). The optimization results show that the dispersion of the natural frequency is reduced while the natural frequency of the pipeline system is increased by robust optimization, so that the influence of random uncertainty factors on the first-order natural frequency response of the pipeline system is greatly reduced, which further shows that the robustness level of the pipeline system dynamics is effectively improved.
[0134] Table 3 Comparison of optimization results of the first-order natural frequency of the piping system
[0135]
[0136] Figure 3 The probability density distribution diagram of the first-order natural frequency of the spatial piping system before and after optimization. By observing the probability density distribution of the deterministic optimization and robust optimization responses, we can more intuitively see the impact of the optimal solutions obtained by the two optimization design methods on the robustness of the natural vibration characteristics of the pipeline. Figure 3 It can be seen that the optimal solutions of the first-order natural frequency output by the two optimization methods are in line with the normal distribution law. However, by comparing the two optimization methods, it can be seen that the probability density distribution of the first-order natural frequency of the optimal solution of the robust optimization method is more concentrated, indicating that its fluctuation degree is smaller and more robust. It proves that the first-order natural frequency response output by the robust optimization design is less disturbed by the small-range fluctuation of the uncertainty noise parameters of the pipeline system, so the robustness of the natural vibration characteristics of the pipeline system is also guaranteed. The results of the robust optimization example of the natural vibration characteristics of the spatial pipeline show that robust optimization can effectively improve the mean value of the first-order natural frequency of the pipeline system, reduce the dispersion of the first-order natural frequency, and improve the stability of the natural vibration characteristics of the pipeline, proving the effectiveness of the robust optimization method of the natural vibration characteristics of the pipeline based on the M-DRM method and the sequential Kriging surrogate model in this application.
[0137] Embodiment 2:
[0138] This embodiment proposes a robustness optimization device for the natural vibration characteristics of an external pipeline, such as Figure 4 As shown, it includes: a finite element model building module, a performance index determination module, a robust optimization model building module, an uncertainty analysis module, an initial sample data set building module, a data prediction module, a global optimization strategy module, a termination condition judgment module, an optimization iteration module and an optimization result output module; the finite element model building module, the performance index determination module, the robust optimization model building module, the uncertainty analysis module, the initial sample data set building module, the data prediction module, the global optimization strategy module, and the termination condition judgment module are connected in sequence, the termination condition judgment module is respectively connected to the optimization iteration module and the optimization result output module, and the optimization iteration module is connected to the uncertainty analysis module;
[0139] The finite element model building module is used to perform parametric finite element modeling on the pipeline system outside the engine and establish a clamp-pipeline natural vibration characteristic analysis model;
[0140] The performance index determination module is used to determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness and the clamp position assembly deviation;
[0141] The robustness optimization model building module is used to build a robustness optimization objective function based on the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and build a robustness optimization model for the natural vibration characteristics of the pipeline;
[0142] The uncertainty analysis module is used to perform Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples. Based on the clamp-pipeline natural vibration characteristic analysis model, the clamp position samples are used as input, and the M-DRM method is used to perform natural vibration characteristic uncertainty analysis on the noise variable to calculate the mean and variance of the first-order natural frequency of the corresponding pipeline system;
[0143] The initial sample data set establishment module is used to substitute the mean and variance of the first-order natural frequency of the corresponding pipeline system into the robustness optimization model establishment module, calculate the value of the robustness optimization objective function, and form the initial sample data set with the clamp position sample and the calculated value of the robustness optimization objective function;
[0144] A data prediction module, used to establish a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtain the mean and variance predicted by the Kriging proxy model;
[0145] The global optimization strategy module is used to construct an efficient global optimization strategy based on the expected lift function based on the mean and variance predicted by the Kriging surrogate model;
[0146] The termination condition judgment module is used to judge whether the iterative termination condition of the efficient global optimization strategy is met;
[0147] The optimization iteration module is used to transfer to the optimization result output module if the iteration termination condition of the efficient global optimization strategy is met; if the iteration termination condition of the efficient global optimization strategy is not met, the new clamp position sample is added to the initial sample data set, and the uncertainty analysis module ~ optimization iteration module are repeated in sequence to update the prediction accuracy of the Kriging proxy model;
[0148] The optimization result output module is used to output the clamp position with the maximum mean and the minimum dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
[0149] Embodiment 3:
[0150] This embodiment proposes an electronic device, including: one or more processors, and a memory, wherein the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the robustness optimization method of the inherent vibration characteristics of the external pipeline.
[0151] The electronic device may be a mobile phone, a computer or a tablet computer, etc., including a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the robustness optimization method of the natural vibration characteristics of the external pipeline as described in the embodiment is implemented. It is understood that the electronic device may also include an input / output (I / O) interface and a communication component.
[0152] The processor is used to execute all or part of the steps in the robustness optimization method for the natural vibration characteristics of the external pipeline as described in the above embodiment. The memory is used to store various types of data, which may include instructions of any application or method in the electronic device, as well as data related to the application.
[0153] The processor can be an application specific integrated circuit (ASIC), a digital signal processor (DSP), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components, and is used to execute the robustness optimization method of the inherent vibration characteristics of the external pipeline described in the above embodiments.
[0154] Embodiment 4:
[0155] This embodiment provides a computer-readable storage medium storing executable instructions. When the instructions are executed and implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
[0156] The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the robustness optimization method for the natural vibration characteristics of the external pipeline described in each embodiment of the present application.
[0157] The aforementioned storage media include: flash memory, hard disk, multimedia card, card-type memory (for example, SD (Secure Digital Memory Card) or DX (Memory Data Register, MDR abbreviation, memory data register) memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, CD, server, APP (Application, abbreviation of application software) application store and other media that can store program verification codes, on which a computer program is stored. When the computer program is executed by a processor, the various steps of the above-mentioned robustness optimization method of the inherent vibration characteristics of the external pipeline can be implemented.
[0158] Embodiment 5:
[0159] This embodiment provides a computer program product, including a computer program or instructions, which, when executed by a processor, implements the robustness optimization method of the natural vibration characteristics of the external pipeline.
[0160] Based on such understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a computer program product.
[0161] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0162] The protection scope of the present application is not limited to the above-mentioned embodiments. Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the scope and spirit of the present disclosure. If these changes and modifications fall within the scope of the claims of the present disclosure and their equivalents, the intention of the present disclosure also includes these changes and modifications.
Claims
1. A method for optimizing the robustness of natural vibration characteristics of a pipeline, characterized in that: include: Step S1: performing parametric finite element modeling on the pipeline system outside the engine, and establishing a clamp-pipeline natural vibration characteristic analysis model; Step S2: Determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness, and the clamp position assembly deviation; Step S3: construct a robustness optimization objective function according to the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and establish a robustness optimization model for the natural vibration characteristics of the pipeline; Step S4: performing Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples, and based on the clamp-pipeline natural vibration characteristic analysis model, using the clamp position samples as input, using the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculating the mean and variance of the first-order natural frequency of the corresponding pipeline system; Step S5: Substitute the value of the robustness optimization objective function calculated in step S3 according to the mean and variance of the first-order natural frequency of the corresponding pipeline system, and form an initial sample data set with the clamp position sample and the value of the robustness optimization objective function calculated; Step S6: establishing a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtaining the mean and variance predicted by the Kriging proxy model; Step S7: Based on the mean and variance predicted by the Kriging surrogate model, an efficient global optimization strategy based on the expected lift function is constructed; Step S8: Determine whether the iteration termination condition of the efficient global optimization strategy is met; Step S9: If the iteration termination condition of the efficient global optimization strategy is met, go to step S10; if the iteration termination condition of the efficient global optimization strategy is not met, add the new clamp position sample to the initial sample data set, repeat steps S4 to S9 in sequence, and update the prediction accuracy of the Kriging proxy model; Step S10: Outputting the clamp position with the largest mean and the smallest dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
2. The pipeline natural vibration characteristic robustness optimization method according to claim 1, characterized in that: The robustness optimization objective function is constructed according to the mean and variance of the first-order natural frequency of the pipeline system, the clamp position is used as the optimization variable, and the constraint conditions of the clamp position are determined to establish the robustness optimization model of the pipeline natural vibration characteristics. The calculation formula is as follows: in, is the mean of the first-order natural frequency of the clamp at its initial position, is the standard deviation of the first-order natural frequency at the initial position of the clamp, that is, the square root of the variance of the first-order natural frequency; is the mean of the first-order natural frequency corresponding to the optimized variable x, The standard deviation of the first-order natural frequency corresponding to the optimized variable x is the square root of the variance of the first-order natural frequency corresponding to x; is the upper limit of the optimization variable, is the lower bound of the optimization variable, is the weight coefficient of the mean, is the weight coefficient of the variance.
3. The pipeline natural vibration characteristic robustness optimization method according to claim 1, characterized in that: The clamp-pipeline natural vibration characteristic analysis model uses the clamp position sample as input, uses the M-DRM method to perform natural vibration characteristic uncertainty analysis on the noise variable, and calculates the mean and variance of the first-order natural frequency of the corresponding pipeline system, including: The natural vibration characteristic function of the piping system is defined by taking the clamp position as the optimization variable, the pipe material properties, the clamp stiffness and the clamp position assembly deviation as the noise variables; The M-DRM method is used to express the natural vibration characteristic function of the pipeline system as a multiplication of multiple one-dimensional single variable sub-functions: According to the multiplication form of multiple one-dimensional univariate sub-functions, the mean and second-order moment of one-dimensional univariate sub-functions are expressed by numerical integration method: The Gaussian quadrature method is used to solve the mean and second-order moment of one-dimensional univariate subfunctions; According to the mean and the second-order moment of the one-dimensional univariate subfunction, the mean and the variance of the first-order natural frequency of the pipeline system are obtained.
4. The pipeline natural vibration characteristic robustness optimization method according to claim 1, characterized in that: The mean and variance of the first-order natural frequency of the corresponding pipeline system are substituted into step S3 to calculate the value of the robustness optimization objective function, and the calculation formula is as follows: in, is the mean of the first-order natural frequency corresponding to the optimized variable x, is the mean of the one-dimensional univariate subfunction, The standard deviation of the first-order natural frequency corresponding to the optimized variable x is the square root of the variance of the first-order natural frequency corresponding to x. is the second-order moment of a one-dimensional univariate subfunction, is the mean of the first-order natural frequency of the clamp at its initial position, is the variance of the first-order natural frequency of the clamp’s initial position, is the weight coefficient of the mean, is the weight coefficient of the variance, for The exponential operation of 1-n, for The 2-n power exponential operation is performed, where n is the dimension of the random variable, k is the kth random variable, c is the anchor point vector, and the mean of the noise variable is taken.
5. The pipeline natural vibration characteristic robustness optimization method according to claim 1, characterized in that: According to the initial sample data set, a Kriging proxy model based on the clamp position and robustness optimization objective function is established, and the mean and variance predicted by the Kriging proxy model are obtained, and the calculation formula is as follows: in, , is the mean predicted by the Kriging surrogate model, is the variance predicted by the Kriging surrogate model, is a polynomial regression model, is the basis function of the Kriging model, , is the regression coefficient, is the regression coefficient The generalized least squares estimate of is the random process variance of the Kriging surrogate model, the input variable x is the clamp position, is the response of the objective function corresponding to pipeline robustness optimization, For the clamp position sample The correlation matrix of To predict the clamp position and the correlation matrix between the existing clamp samples, are the relevant model parameters, is the number of sample points.
6. The pipeline natural vibration characteristic robustness optimization method according to claim 1, characterized in that: The method constructs an efficient global optimization strategy based on the expected lift function based on the mean and variance predicted by the Kriging surrogate model, including: Determine the point addition criterion of the expected improvement function according to the expected value of the improvement amount of the objective function; Among them, the improvement of the objective function is defined as: in, is the improvement of the objective function, the input variable x is the clamp position, The predicted value of the objective function for pipeline robustness optimization at any clamp position is subject to normal distribution. is the objective function value of the optimal pipeline robustness optimization corresponding to the current clamp position sample; Based on the mean and variance predicted by the Kriging surrogate model, the expected lift function is obtained, and the calculation formula is as follows: in, is the cumulative distribution function of the standard normal distribution, is the probability density function of the standard normal distribution, is the mean predicted by the Kriging surrogate model, is the standard deviation of the Kriging surrogate model prediction, that is, the square root of the variance predicted by the Kriging surrogate model, is the expected improvement function; Construct an efficient global optimization strategy based on the expected improvement function. The calculation formula is as follows: in, is the upper limit of the optimization variable, To optimize the lower limit of the variable, the iterative termination condition of the global optimization strategy is ,at this time Optimize the optimal solution for clamp position robustness.
7. A device for optimizing the robustness of natural vibration characteristics of a pipeline, characterized in that: include: The finite element model building module is used to perform parametric finite element modeling on the pipeline system outside the engine and establish a clamp-pipeline natural vibration characteristic analysis model; The performance index determination module is used to determine that the output of the clamp-pipeline natural vibration characteristic analysis model is the first-order natural frequency of the pipeline system, the input of the clamp-pipeline natural vibration characteristic analysis model is the clamp position, and the noise variables are the material properties of the pipe, the clamp stiffness and the clamp position assembly deviation; The robustness optimization model building module is used to build a robustness optimization objective function based on the mean and variance of the first-order natural frequency of the pipeline system, take the clamp position as the optimization variable, determine the constraint conditions of the clamp position, and build a robustness optimization model for the natural vibration characteristics of the pipeline; The uncertainty analysis module is used to perform Latin hypercube sampling on multiple clamp positions within a preset interval to obtain clamp position samples. Based on the clamp-pipeline natural vibration characteristic analysis model, the clamp position samples are used as input, and the M-DRM method is used to perform natural vibration characteristic uncertainty analysis on the noise variable to calculate the mean and variance of the first-order natural frequency of the corresponding pipeline system; The initial sample data set establishment module is used to substitute the mean and variance of the first-order natural frequency of the corresponding pipeline system into the robustness optimization model establishment module, calculate the value of the robustness optimization objective function, and form the initial sample data set with the clamp position sample and the calculated value of the robustness optimization objective function; A data prediction module, used to establish a Kriging proxy model based on the clamp position and robustness optimization objective function according to the initial sample data set, and obtain the mean and variance predicted by the Kriging proxy model; The global optimization strategy module is used to construct an efficient global optimization strategy based on the expected lift function based on the mean and variance predicted by the Kriging surrogate model; The termination condition judgment module is used to judge whether the iterative termination condition of the efficient global optimization strategy is met; The optimization iteration module is used to transfer to the optimization result output module if the iteration termination condition of the efficient global optimization strategy is met; if the iteration termination condition of the efficient global optimization strategy is not met, the new clamp position sample is added to the initial sample data set, and the uncertainty analysis module ~ optimization iteration module are repeated in sequence to update the prediction accuracy of the Kriging proxy model; The optimization result output module is used to output the clamp position with the maximum mean and the minimum dispersion of the first-order natural frequency of the pipeline system as the optimal robustness optimization result.
8. An electronic device, characterized in that: include: One or more processors, and a memory, wherein the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the method for optimizing the robustness of the inherent vibration characteristics of a pipeline as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that: It stores executable instructions, which, when executed, enable the processor to execute the pipeline natural vibration characteristic robustness optimization method described in any one of claims 1 to 6.
10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method for optimizing the robustness of the inherent vibration characteristics of a pipeline as described in any one of claims 1 to 6 is implemented.
