CN113255235B - Approximate modeling method, device, equipment and medium for complex structure of aircraft - Google Patents
Approximate modeling method, device, equipment and medium for complex structure of aircraft Download PDFInfo
- Publication number
- CN113255235B CN113255235B CN202110716401.7A CN202110716401A CN113255235B CN 113255235 B CN113255235 B CN 113255235B CN 202110716401 A CN202110716401 A CN 202110716401A CN 113255235 B CN113255235 B CN 113255235B
- Authority
- CN
- China
- Prior art keywords
- samples
- clustering
- sample
- prediction
- aircraft
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mathematical Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Mathematics (AREA)
- Evolutionary Biology (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Feedback Control In General (AREA)
Abstract
The application relates to an approximate modeling method, device, equipment and medium for an aircraft complex structure, wherein the method comprises the steps of obtaining a plurality of sample data of the aircraft structure, and clustering each sample data by adopting a K-means clustering method; grouping the clustered samples by adopting a proportional allocation method, and constructing an agent model for predicting the performance of the aircraft structure by adopting a local density-based enhanced anisotropic radial basis method according to the obtained multiple groups of clustered samples; calculating the prediction error of the agent model by adopting a rapid cross validation method according to the groups of clustering grouped samples and calculating the root mean square error of all the prediction errors; the method comprises the steps that the root mean square error is used as a target function, the anisotropic scaling coefficient is used as a design variable, and optimization processing is carried out on the anisotropic scaling coefficient according to the root mean square error to obtain a target scaling coefficient corresponding to the minimum root mean square error; and performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model. Performance indicates higher efficiency.
Description
Technical Field
The application relates to the technical field of aircraft design, in particular to an approximate modeling method, device, equipment and medium for an aircraft complex structure.
Background
Due to extreme working conditions and strict performance requirements, the structure of the complex aircraft needs to have excellent performance in the aspects of strength, rigidity, bearing capacity and the like. The performance of different structures of the complex aircraft is one of the most core and difficult technologies in the structural design of the aircraft, and the main task is to evaluate the bearing capacity of the aircraft in different structural forms by predicting the deformation, damage and other behaviors of the aircraft structure under the working load, so that the structure with light weight and high strength is designed.
The current commonly used structural performance prediction methods of the complex aircraft mainly comprise: (1) and the program simulation method calls commercial finite element software or a high-precision finite element analysis program aiming at the new aircraft structure form to predict the bearing performance of the structure, and the prediction precision is higher. (2) Constructing a prediction model of the structural performance of the aircraft, constructing prediction models of different structural parameters and the bearing performance of the aircraft based on the existing experimental data, and taking the new structural parameters as input to obtain the corresponding predicted bearing performance; typical methods for constructing a model include a radial basis method, a Kriging (Kriging) method, a polynomial response surface method, and the like. The method can avoid time-consuming program simulation, the prediction precision is continuously improved along with the increase of model training data, and the prediction of the aircraft structure performance under the condition of any structure parameter can be realized.
However, in the process of implementing the invention, the inventor finds that the conventional method for predicting the performance of the complex aircraft structure still has the technical problem of low efficiency in predicting the performance of the aircraft structure.
Disclosure of Invention
In view of the above, it is necessary to provide an aircraft complex structure approximate modeling method, an aircraft complex structure approximate modeling device, a computer device, and a computer-readable storage medium, which have high efficiency of predicting aircraft structure performance.
In order to achieve the above purpose, the embodiment of the invention adopts the following technical scheme:
in one aspect, an embodiment of the present invention provides an aircraft complex structure approximate modeling method, including:
acquiring a plurality of sample data of an aircraft structure, and clustering each sample data by adopting a K-means clustering method;
grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustered grouped samples; each group of clustering grouping samples comprises clustering information;
according to the multiple groups of clustering grouping samples, a proxy model for performance prediction of the aircraft structure is constructed by adopting a local density-based enhanced anisotropic radial basis method;
calculating the prediction error of the agent model by adopting a rapid cross validation method according to the groups of clustering grouping samples and calculating the root mean square error of all the prediction errors;
the method comprises the steps that the root mean square error is used as a target function, the anisotropic scaling coefficient is used as a design variable, and optimization processing is carried out on the anisotropic scaling coefficient according to the root mean square error to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing a proxy model;
performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
In another aspect, an aircraft complex structure approximate modeling apparatus is further provided, including:
the cluster processing module is used for acquiring a plurality of sample data of the aircraft structure and clustering each sample data by adopting a K-means clustering method;
the clustering grouping module is used for grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustering grouping samples; each group of clustering grouping samples comprises clustering information;
the agent building module is used for building an agent model for performance prediction of the aircraft structure by adopting a local density-based enhanced anisotropic radial basis method according to the multiple groups of clustering grouping samples;
the prediction error module is used for calculating the prediction errors of the agent model by adopting a rapid cross validation method according to the groups of clustering grouping samples and calculating the root mean square errors of all the prediction errors;
the optimization processing module is used for optimizing the anisotropic scaling coefficient according to the root mean square error by taking the root mean square error as a target function and the anisotropic scaling coefficient as a design variable to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing a proxy model;
the prediction execution module is used for predicting the performance of the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
In still another aspect, a computer device is provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the aircraft complex structure approximate modeling method when executing the computer program.
In yet another aspect, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of any of the above-mentioned aircraft complex structure approximate modeling methods.
One of the above technical solutions has the following advantages and beneficial effects:
according to the aircraft complex structure approximate modeling method, the aircraft complex structure approximate modeling device, the aircraft complex structure approximate modeling equipment and the aircraft complex structure approximate modeling medium, based on a radial basis method, for existing real sample data, a K-means clustering method is adopted to cluster samples, groups of sample sets with equal number are formed in a form of proportional distribution, each group of sample set covers information of each clustered sample, a prediction model (namely a proxy model) is built by applying a sample local density-based enhanced anisotropic radial basis method, a rapid cross validation method and an intelligent optimization algorithm are adopted to select a proper anisotropic scaling coefficient, and a final high-precision prediction model (namely a high-precision performance prediction model) is built.
Due to the fact that anisotropy of the model in different dimensions is considered, the check rate of model precision is accelerated by adopting a rapid cross-validation method, model construction efficiency is greatly improved on the basis of guaranteeing the model precision, rapid and high-precision construction of a complex aircraft structure performance prediction model is achieved, and the purpose of remarkably improving aircraft structure performance prediction efficiency is achieved. Compared with the best technology in the prior art, the method adopts a rapid cross validation method compared with a general model training method, has higher training speed and improves the construction efficiency; aiming at the problem of large prediction error of a general model, a method for enhancing the anisotropic radial basis based on local density is provided, and the reliability of the result is improved.
Drawings
FIG. 1 is a schematic flow chart diagram of a method for approximate modeling of an aircraft complex structure according to one embodiment;
FIG. 2 is a schematic flow chart of an adaptive enhanced radial basis modeling method suitable for approximate modeling of a complex structure of a complex aircraft according to an embodiment;
FIG. 3 is a schematic view of a cylindrical stiffened shell configuration in one embodiment;
FIG. 4 is a schematic diagram illustrating convergence of prediction errors of an approximate modeling model of an aircraft complex structure in one embodiment;
FIG. 5 is a schematic diagram of the prediction accuracy of an aircraft complex structure approximate modeling model constructed by four methods in one embodiment;
FIG. 6 is a schematic block diagram of an aircraft complex structure approximate modeling apparatus according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
In addition, the technical solutions in the embodiments of the present invention may be combined with each other, but it must be based on the realization of those skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination of technical solutions should be considered to be absent and not within the protection scope of the present invention.
In practice, the inventor finds that the current commonly used method for predicting the performance of the complex aircraft structure has the following disadvantages: (1) the method calls commercial finite element software or a high-precision finite element analysis program to predict the bearing performance of the structure, the time-consuming high-precision finite element program needs to be called, simulation is carried out on different aircraft structures, one configuration may need hours or even days to obtain a result, the calculation cost is obviously increased along with the improvement of the result precision, and the method is difficult to bear in the actual optimization design engineering.
(2) Although a complex aircraft structure performance prediction model can be constructed, performance parameters of a new structure form can be obtained quickly compared with the former method, the constructed prediction model is generally poor in prediction accuracy, and a prediction result meeting design requirements is difficult to obtain for a new structure.
Aiming at the technical problem that the prediction efficiency of the structural performance of the aircraft is low in the conventional prediction method of the structural performance of the complex aircraft, the application provides an efficient, rapid and high-precision approximate modeling method of the complex structure of the aircraft. The method comprises the steps of clustering existing sample data by adopting a K-means clustering method, obtaining several groups of sample sets containing each clustering data by a proportional distribution method, constructing a proxy model for predicting structural performance by applying an anisotropic radial basis method based on local density, optimizing model parameters by adopting rapid cross validation and an intelligent optimization algorithm, and completing construction of an aircraft structural performance prediction model, wherein the method can be used for realizing efficient, rapid and high-precision prediction of aircraft structural performance. The prediction method provided by the application is different from a general aircraft structure performance prediction model construction method, the existing real data are processed, the radial basis model modeling technology is improved, and a rapid and high-precision construction method is provided for the aircraft structure performance prediction model.
Referring to fig. 1, in one embodiment, the present invention provides a method for modeling an aircraft complex structure approximately, including the following steps S12 to S22:
and S12, acquiring a plurality of sample data of the aircraft structure, and clustering each sample data by adopting a K-means clustering method.
It can be understood that each sample data of the aircraft structure is the existing real sample data, and can be acquired in a pre-acquisition or uploading mode and the like. The K-means clustering method, which is a K-means clustering algorithm in the art, is a clustering analysis algorithm for iterative solution, and firstly, a K-means clustering method is adopted to perform clustering processing on all existing sample data to form a set of sample data of a plurality of different clustering categories.
In some embodiments, step S12 may specifically include the following sub-steps S122 to S126:
s122, generating all sample points uniformly distributed in a design space by adopting an optimized Latin hypercube method according to the sample data;
s124, randomly selecting centers of K clusters in a design space, respectively calculating Euclidean distance from each sample point to the center of each cluster, and distributing each sample point to the center of the cluster closest to the sample point to obtain samples of the K clusters; k is a positive integer less than the total number of samples of the sample data;
and S126, respectively carrying out cluster center recalculation on the K clustered samples, and outputting K groups of clustered samples obtained by current clustering processing.
It can be understood that for each sample point in the design space, an optimized latin hypercube method existing in the field can be adopted, and the generation is performed based on all the obtained sample data.
Specifically, firstly, an optimized Latin hypercube method is adopted to generate all sample points which are uniformly distributed in a design space, and then, the existing sample points are aimed at
Samples and each sample dimension is
Data collection of
I.e. all sample data, using
Dividing sample data into blocks by mean clustering method
The number of the groups is set to be,
and
are all positive integers.
And normalizing all sample points between the upper and lower boundaries of each dimension to ensure that each dimension of the sample is changed between 0 and 1. Randomly selecting centers of K clusters in design space
Each sample is calculated using the following formula (1)
Euclidean distance to each cluster center, and combining the samples
Assigned to the nearest cluster center, the cluster center is obtained
Individual clustered samples.
In the formula (1), the reaction mixture is,
representing the euclidean distance of two sample points,
and
second representing the calculated sample point and the cluster center, respectively
Dimension variables.
In some embodiments, step S126 may be implemented by the following steps:
carrying out cluster center recalculation on each clustered sample by adopting a set center calculation mode; the central calculation mode is as follows:
wherein the content of the first and second substances,
is shown as
The cluster center of each of the clusters is,
which represents the ith sample point, is,
represents a random selection
The center of each of the clusters is,
is shown as
The number of samples of an individual cluster;
adopting a set convergence criterion to recalculate the clustering center to carry out convergence judgment, and outputting a clustering sample of the current clustering treatment if the convergence judgment is carried out; the convergence criterion is:
wherein the content of the first and second substances,
representing the euclidean distance of two sample points,
representing a set euclidean distance threshold; all samples have been normalized within the design space,
the smaller the value the more accurate the algorithm converges. Alternatively to this, the first and second parts may,
the value was taken to be 0.001.
Specifically, the cluster center is recalculated for each clustered sample by using the formula (2) and convergence judgment is performed in the calculation process. If it is
Then the program performing the calculation converges and outputs the sample of the current cluster.
In other embodiments, if
Then the program performing the calculation does not converge and will
As the next step
And step S124, respectively calculating the Euclidean distance from each sample point to the center of each cluster, allocating each sample point to the center of the cluster with the closest distance, obtaining the sample processing process of K clusters, and performing the next iteration processing.
S14, grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustered grouped samples; each group of cluster grouping samples includes individual cluster information.
In some embodiments, the method specifically comprises the following processing steps:
distributing the K groups of clustering samples obtained by clustering according to proportion to obtain
Grouping the grouped samples;
wherein the number of samples in each group of clustering grouping samples is
And contains the sample information for each of the clusters,
is a positive integer less than the total number of samples of the sample data; wherein the content of the first and second substances,
and the number of samples of each cluster contained in each group of cluster grouping samples is calculated in the following way:
wherein N represents the total number of samples of the sample data,
it is shown that the operation of taking the remainder,
representing ith group of clustered samples
Is internally provided with
The number of samples in a cluster is determined,
indicating a rounding-up operation, allocated
After the group adopts
Make up for the remainder and control
The number of samples of (a) is equal two by two.
In particular obtained for clustering
Group samples, proportionally allocated to obtain new
Group sample
Each group has
And contains partial sample information for each cluster. The proportion setting in the grouping can be flexibly set according to the number of the sample information of each cluster required to be included in each group of cluster grouping samples, and the only limitation is not required in the specification as long as each group of clusters can be groupedThe samples all contain partial sample information of each cluster.
And S16, grouping the samples according to the clusters, and constructing a proxy model for performance prediction of the aircraft structure by adopting a local density-based enhanced anisotropic radial basis method.
It can be understood that the agent model established in this embodiment, that is, the prediction model that can be used for predicting the performance of the aircraft structure, belongs to a model that is not trained before optimization processing.
In some embodiments, the following processing steps S162 to S166 may be specifically included:
s162, grouping samples according to the multiple groups of clusters, and calculating the local density of sample points;
s164, setting the influence volume of each sample point to be inversely proportional to the local density of the sample point, setting the sum of the influence volumes of all the sample points to be equal to 1, and calculating to obtain a reference value of each shape parameter of the radial basis function;
s166, multiplying the reference value of each shape parameter by a set scaling factor in each dimension to obtain each real shape parameter;
constructing a proxy model based on the shape parameters; the proxy model is:
wherein the content of the first and second substances,
which represents the point to be predicted and which is,
which represents the size of the training sample,
the weight coefficient is represented by a coefficient of weight,
representing Euclidean conversion from unknown to known samplesDistance is the basis function of the argument:
wherein the content of the first and second substances,
the parameters of the shape are represented by,
representing the number of samples of each group of clustered grouped samples,
sample points representing the ith cluster grouping samples.
Specifically, before constructing the proxy model, the shape parameters are estimated based on the sample local density: first, the local density of the above cluster sample is calculated using the following formula (6).
In the formula (I), the compound is shown in the specification,
is a shape parameter of a gaussian function, and is set with reference to the following equation (7):
after the local density of the sample points is obtained, the shape parameter of the radial basis function is regarded as the influence radius of the sample point, and the influence range (influence volume) of each sample point is inversely proportional to the local density of the point, namely, the following formula (8):
order postThe sum of the volumes of the point influences is equal to 1, and the reference value of the shape parameter, i.e. the reference size of each shape parameter, can be calculated as shown in the following formula (9)
。
Shape parameters for a reference
In each dimension
Up by a separate scaling factor
To obtain the true shape parameters, as shown in the following formula (10):
a radial basis model, i.e., the above-described proxy model, is constructed using the above equation (4), and the Gaussian basis function shown in the above equation (5) is used as the desired basis function
. And optimizing the scaling factor by adopting an intelligent optimization algorithm. After the shape parameters are determined, the corresponding weight coefficients can be calculated by reversely calculating the above formula (4):
wherein the content of the first and second substances,
coefficient matrix inverse for all samples，
The weight coefficients for all samples. After the weight coefficient is determined, aiming at any sample
The corresponding prediction output can be obtained by equation (4).
And S18, calculating the prediction error of the proxy model by adopting a rapid cross validation method according to the groups of clustering grouping samples and calculating the root mean square error of all the prediction errors.
In some embodiments, the following substeps S182 to S188 may be specifically included:
s182, selecting the ith group of clustering grouping samples
As a test sample and using the clustering grouping samples of the other groups as training samples;
k is the total number of clustering grouping samples;
s184, calculating the prediction error of the agent model based on the training sample on the test sample by adopting a set rapid cross-validation method; the set rapid cross-validation method comprises the following steps:
wherein the content of the first and second substances,
corresponding to test samples in an inverse matrix of a matrix of coefficients representing all samples
Go to,
Overlapping part of column elementsIs divided into
The order of the matrix is such that,
representing the correspondence of test samples
The number of the individual weight coefficients is,
representing a current proxy model (also known as a sub-prediction model) in a test sample
A prediction error of (a);
s186, returning to the step S182, and starting to calculate the next group of clustering grouping samples until the prediction error of each group of clustering grouping samples is obtained through calculation;
s188, calculating the root mean square error of each prediction error; the root mean square error calculation method is as follows:
wherein the content of the first and second substances,
the root mean square error is represented as a function of,
representing the total number of samples of the sample data,
the true output of the sample is represented by,
representing the output of the proxy model prediction.
In particular, the method comprises the following steps of,get
The prediction error of the proxy model constructed by using the group samples as the test samples and the other samples as the training samples on the test samples can be calculated by applying the above formula (12) through a rapid cross-validation method. Steps S182 to S186 are repeated until the prediction errors of all samples are calculated, and the root mean square error values of all the prediction errors are calculated using the above expression (13).
S20, optimizing the anisotropic scaling coefficient according to the root mean square error by taking the root mean square error as a target function and the anisotropic scaling coefficient as a design variable to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used to construct a proxy model.
It can be understood that an intelligent optimization algorithm can be used to optimize the scaling coefficient, and the scaling coefficient which enables the predicted root mean square error to be the minimum is the finally constructed high-precision proxy model, that is, the high-precision performance prediction model in the following.
As shown in fig. 2, the minimum predicted root mean square error and the corresponding scaling factor (i.e. the target scaling factor) are obtained by using an intelligent optimization algorithm, i.e. the root mean square error predicted by the training model is used as the objective function, and the anisotropic scaling factor is used as the design variable for optimization
) And then constructing a high-precision performance prediction model of the original complex simulation model.
S22, performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
Specifically, the existing sample data is clustered, and several sets of sample sets containing each clustering information are obtained through a proportional allocation method. The method comprises the steps of constructing an agent model for predicting the structural performance of the aircraft by using a local density-based enhanced anisotropic radial basis method, calculating a prediction error of constructing the agent model each time by using a rapid cross validation method, and selecting an optimal anisotropic scaling coefficient of the model by optimizing the minimum prediction error of a training sample, thereby completing construction of a prediction model of the structure of the complex aircraft and realizing efficient and high-precision prediction of the structural performance of the complex aircraft.
According to the aircraft complex structure approximate modeling method, based on a radial basis method, for existing real sample data, a K-means clustering method is adopted to cluster samples, groups of sample sets with equal number are formed in a proportional distribution mode, each cluster sample set covers information of the sample, a prediction model (namely a proxy model) is constructed by applying a sample local density-based enhanced anisotropic radial basis method, a rapid cross validation method and an intelligent optimization algorithm are adopted to select a proper anisotropic scaling coefficient, and a final high-precision prediction model (namely a high-precision performance prediction model) is constructed.
Due to the fact that anisotropy of the model in different dimensions is considered, the check rate of model precision is accelerated by adopting a rapid cross-validation method, model construction efficiency is greatly improved on the basis of guaranteeing the model precision, rapid and high-precision construction of a complex aircraft structure performance prediction model is achieved, and the purpose of remarkably improving aircraft structure performance prediction efficiency is achieved. Compared with the best technology in the prior art, the method adopts a rapid cross validation method compared with a general model training method, has higher training speed and improves the construction efficiency; aiming at the problem of large prediction error of a general model, a method for enhancing the anisotropic radial basis based on local density is provided, and the reliability of the result is improved.
It should be understood that although the steps in the flowcharts of fig. 1 and 2 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps of fig. 1 and 2 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
In one embodiment, in order to more intuitively and fully describe the approximate modeling method for the complex structure of the aircraft, a calculation example is given below by taking the design of a reinforced shell structure of a large-scale launch vehicle as an example, and is an example of application description and verification of the performance prediction method provided by the invention.
It should be noted that the implementation examples given in this specification are only illustrative and are not the only limitations of the specific implementation examples of the present invention, and those skilled in the art can use the approximate modeling method for the complex structure of the aircraft provided above in the same manner and under the schematic illustration of the implementation examples provided in the present invention to realize the efficient performance prediction for different aircraft structures.
The configuration of the cylindrical reinforced shell adopting the inner peripheral frame and the outer vertical longitudinal beam is shown in fig. 3, wherein fig. 3 (a) is a schematic diagram of a frame truss reinforced thin-wall reinforced column shell structure, fig. 3 (b) is a schematic diagram of layout parameters of a middle frame, and fig. 3 (c) is a schematic diagram of section parameters. The configuration parameters of the reinforced shell are selected as design variables, and the specific parameters and the variation range thereof are shown in the following table 1:
TABLE 1
Firstly, an optimized Latin hypercube method is adopted to generate 400 sample points which are uniformly distributed in a design space, and
classifying the samples by means of mean clustering method, and then adopting automatic grouping method to make pairs
Class samples are grouped and then enhancement based on local density of samples is appliedAnd constructing a proxy model by an anisotropic radial basis method, calculating a prediction error by applying a rapid cross validation method, optimizing a scaling coefficient by adopting an intelligent optimization algorithm by taking the minimum root mean square error value as an objective function, and obtaining the scaling coefficient corresponding to the minimum root mean square error, namely the scaling coefficient applied by constructing the high-precision proxy model finally. The method comprises the following specific steps:
(1) generating 400 sample points which are uniformly distributed in a design space by adopting an optimized Latin hypercube method;
(2) dividing the samples into 10 classes by adopting a clustering method, dividing the samples into 10 groups by adopting an automatic grouping method, wherein each group comprises 10 classes of samples;
(3) constructing a performance prediction model of the reinforced shell by adopting an enhanced anisotropic radial basis model based on the local density of the sample;
(4) calculating a prediction error of the model and a root mean square error value of the calculation error by adopting a rapid cross validation method;
(5) optimizing the scaling coefficient by applying an intelligent optimization algorithm to obtain the scaling coefficient which enables the root mean square error to be minimum, namely the agent model which needs to be constructed finally;
(6) 5000 sample points which are uniformly distributed are generated in a design domain by adopting a quasi-random sequence, and the accuracy of the constructed performance prediction model is verified.
Case 1:
the first case firstly adopts a K-means clustering method for 400 training samples, 400 samples are clustered into 10 classes, and then 10 groups of samples are obtained through sample distribution according to proportion, wherein each group comprises 40 samples and sample information of each cluster.
And (3) optimizing an anisotropic scaling coefficient by adopting an intelligent particle swarm algorithm and taking a cross validation error (RMSE) value of a training sample as an optimization target, setting an initial population to be 50 considering that the problem dimension is 19 dimensions, setting the search range of the scaling coefficient to be [0.1,300] in each dimension, and setting the convergence condition to be that the maximum iteration step number is equal to 3000 steps or the historical optimal population converges. The resulting convergence curve is shown in fig. 4. As the initial random error of the initial population is larger, the error obtained by adopting the rapid cross validation training model is gradually converged along with the continuous adjustment and optimization of the scaling coefficient, the order of magnitude of 10-30 is reached, and the training of the performance prediction model of the reinforced shell structure is completed.
And (3) generating 5000 sample points which are uniformly distributed in a design space by adopting an optimized Latin hypercube so as to evaluate the precision of the constructed performance prediction model. Based on the anisotropic scaling coefficient obtained by the optimization, a proxy model constructed by enhanced anisotropic radial basis based on the local density of the sample is adopted, the prediction results of 5000 sample points are shown in fig. 5, in order to verify the effectiveness of the prediction method, the prediction method is compared with the approximate modeling method proposed by Kitayama and Rippa and the Kriging method under the same condition by adopting the same training prediction sample, and the root mean square error calculated by other three methods is also shown in fig. 5. In the analysis problem of the cylindrical stiffened shell structure, both the Kitayama method (i.e. KA bar graph in fig. 5) and the Kriging method (i.e. KG bar graph in fig. 5) have certain limitations, and the errors thereof respectively reach 194% and 158% of the predictive method (i.e. BO bar graph in fig. 5) and the Rippa method (i.e. RA bar graph in fig. 5) in the present application, but the predictive accuracy of the predictive method in the present application is slightly better than that of the Rippa method, and the effectiveness of the predictive method in the present application is demonstrated.
Compared with the traditional method, the forecasting method is high in simulation speed, high in performance forecasting precision and capable of saving a large amount of time cost, has higher efficiency and reliability compared with a model with long training time and poor forecasting precision, and can effectively meet the requirement for efficient forecasting of the structural performance of the complex aircraft.
Referring to fig. 6, there is also provided an aircraft complex structure approximate modeling apparatus 100, which includes a cluster processing module 13, a cluster grouping module 15, an agent constructing module 17, a prediction error module 19, an optimization processing module 21, and a prediction executing module 23. The clustering module 13 is configured to obtain a plurality of sample data of the aircraft structure, and perform clustering on each sample data by using a K-means clustering method. The clustering grouping module 15 is used for grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustering grouping samples; each group of cluster grouping samples includes individual cluster information. The agent construction module 17 is configured to construct an agent model for performance prediction of the aircraft structure according to the plurality of groups of cluster grouping samples and by using a local density-based enhanced anisotropic radial basis method. The prediction error module 19 is configured to calculate prediction errors of the proxy model by using a fast cross validation method according to the groups of clustering grouping samples and calculate root mean square errors of all the prediction errors. The optimization processing module 21 is configured to perform optimization processing on the anisotropic scaling factor according to the root mean square error by using the root mean square error as a target function and using the anisotropic scaling factor as a design variable, so as to obtain a target scaling factor corresponding to the minimum root mean square error; the anisotropic scaling factor is used to construct a proxy model. The prediction execution module 23 is used for performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
The aircraft complex structure approximate modeling device 100 is based on a radial basis method through cooperation of modules, clusters samples by adopting a K-means clustering method aiming at existing real sample data, forms several groups of sample sets with equal quantity in a form of proportional distribution, covers each clustered sample information, constructs a prediction model (namely a proxy model) by adopting a sample local density-based enhanced anisotropic radial basis method, selects a proper anisotropic scaling coefficient by adopting a rapid cross validation method and an intelligent optimization algorithm, and constructs a final high-precision prediction model (namely a high-precision performance prediction model).
Due to the fact that anisotropy of the model in different dimensions is considered, the check rate of model precision is accelerated by adopting a rapid cross-validation method, model construction efficiency is greatly improved on the basis of guaranteeing the model precision, rapid and high-precision construction of a complex aircraft structure performance prediction model is achieved, and the purpose of remarkably improving aircraft structure performance prediction efficiency is achieved. Compared with the best technology in the prior art, the method adopts a rapid cross validation method compared with a general model training method, has higher training speed and improves the construction efficiency; aiming at the problem of large prediction error of a general model, a method for enhancing the anisotropic radial basis based on local density is provided, and the reliability of the result is improved.
In one embodiment, the modules of the aircraft complex structure approximate modeling apparatus 100 may also be used to implement the corresponding sub-steps in the embodiments of the aircraft complex structure approximate modeling method.
For specific limitations of the aircraft complex structure approximate modeling apparatus 100, reference may be made to the corresponding limitations of the aircraft complex structure approximate modeling method in the foregoing, and details are not repeated here. The various modules in the aircraft complex structure approximation modeling apparatus 100 described above may be implemented in whole or in part by software, hardware, and combinations thereof. The modules may be embedded in hardware or independent from a device with specific data processing function, or may be stored in a memory of the device in software, so that the processor can invoke and execute operations corresponding to the modules, where the device may be, but is not limited to, a computer device or a computing system for designing an aircraft.
In still another aspect, a computer device is provided, which includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the following steps: acquiring a plurality of sample data of an aircraft structure, and clustering each sample data by adopting a K-means clustering method; grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustered grouped samples; each group of clustering grouping samples comprises clustering information; according to the multiple groups of clustering grouping samples, a proxy model for performance prediction of the aircraft structure is constructed by adopting a local density-based enhanced anisotropic radial basis method; calculating the prediction error of the agent model by adopting a rapid cross validation method according to the groups of clustering grouping samples and calculating the root mean square error of all the prediction errors; the method comprises the steps that the root mean square error is used as a target function, the anisotropic scaling coefficient is used as a design variable, and optimization processing is carried out on the anisotropic scaling coefficient according to the root mean square error to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing a proxy model; performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
In one embodiment, the processor when executing the computer program may also implement the additional steps or sub-steps of the above-described embodiments of the method for approximate modeling of an aircraft complex structure.
In yet another aspect, there is also provided a computer readable storage medium having a computer program stored thereon, the computer program when executed by a processor implementing the steps of: acquiring a plurality of sample data of an aircraft structure, and clustering each sample data by adopting a K-means clustering method; grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustered grouped samples; each group of clustering grouping samples comprises clustering information; according to the multiple groups of clustering grouping samples, a proxy model for performance prediction of the aircraft structure is constructed by adopting a local density-based enhanced anisotropic radial basis method; calculating the prediction error of the agent model by adopting a rapid cross validation method according to the groups of clustering grouping samples and calculating the root mean square error of all the prediction errors; the method comprises the steps that the root mean square error is used as a target function, the anisotropic scaling coefficient is used as a design variable, and optimization processing is carried out on the anisotropic scaling coefficient according to the root mean square error to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing a proxy model; performing performance prediction on the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is a target scaling factor.
In one embodiment, the computer program, when executed by the processor, may further implement the additional steps or sub-steps of the embodiments of the aircraft complex structure approximation modeling method described above.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware related to instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), synchronous link DRAM (Synchlink) DRAM (SLDRAM), Rambus DRAM (RDRAM), and interface DRAM (DRDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for those skilled in the art, various changes and modifications can be made without departing from the spirit of the present application, and all of them fall within the scope of the present application. Therefore, the protection scope of the present patent should be subject to the appended claims.
Claims (10)
1. An approximate modeling method for an aircraft complex structure, characterized by comprising the steps of:
acquiring a plurality of sample data of an aircraft structure, and clustering each sample data by adopting a K-means clustering method;
grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustered grouped samples; each group of clustering grouping samples comprises clustering information;
constructing a proxy model for performance prediction of the aircraft structure by adopting a local density-based enhanced anisotropic radial basis method according to the plurality of groups of clustering grouping samples;
calculating the prediction error of the agent model by adopting a rapid cross validation method and calculating the root mean square error of all the prediction errors according to a plurality of groups of clustering grouping samples;
optimizing the anisotropic scaling coefficient according to the root mean square error by taking the root mean square error as a target function and taking the anisotropic scaling coefficient as a design variable to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing the proxy model;
performing performance prediction on the aircraft structure by using a finally constructed high-precision performance prediction model; the high-precision performance prediction model is the target scaling factor.
2. The aircraft complex structure approximate modeling method according to claim 1, wherein a plurality of sample data of the aircraft structure are acquired, and a K-means clustering method is adopted to perform clustering processing on each sample data, and the method comprises the steps of:
generating all sample points which are uniformly distributed in a design space by adopting an optimized Latin hypercube method according to the sample data;
randomly selecting centers of K clusters in a design space, respectively calculating Euclidean distance from each sample point to the center of each cluster, and allocating each sample point to the center of the cluster closest to the sample point to obtain samples of the K clusters; k is a positive integer less than the total number of samples of the sample data;
and respectively carrying out cluster center recalculation on the K clustered samples, and outputting K groups of clustered samples obtained by current clustering processing.
3. The aircraft complex structure approximate modeling method according to claim 2, characterized in that the step of grouping the clustered samples by using a proportional distribution method to obtain a plurality of groups of clustered and grouped samples comprises:
distributing K groups of clustering samples obtained by clustering according to a proportion to obtain K groups of clustering grouped samples; the number of samples in each group of the clustering grouping samples is
And contains sample information for each cluster, k being a positive integer less than the total number of samples of the sample data; wherein the content of the first and second substances,
and the number of samples of each cluster contained in each group of cluster grouping samples is calculated in the following way:
wherein N represents a total number of samples of the sample data,
it is shown that the operation of taking the remainder,
representing ith group of clustered samples
Is internally provided with
The number of samples in a cluster is determined,
indicating a rounding-up operation, allocated
After the group adopts
Make up for the remainder and control
The number of samples of (a) is equal two by two.
4. The aircraft complex structure approximate modeling method according to any one of claims 1 to 3, characterized in that the step of constructing a proxy model for performance prediction of the aircraft structure by using a local density-based enhanced anisotropic radial basis method according to a plurality of groups of the cluster grouping samples comprises:
according to the groups of clustering grouping samples, local density calculation of sample points is carried out;
setting the influence volume of each sample point to be in inverse proportion to the local density of the sample point, setting the sum of the influence volumes of all the sample points to be equal to 1, and calculating to obtain a reference value of each shape parameter of the radial basis function;
multiplying the reference value of each shape parameter by a set scaling factor in each dimension to obtain each real shape parameter;
building the proxy model based on the shape parameters; the proxy model is as follows:
wherein the content of the first and second substances,
which represents the point to be predicted and which is,
which represents the size of the training sample,
the weight coefficient is represented by a coefficient of weight,
representing the basis functions with the euclidean distance of the unknown sample to the known sample as the argument:
wherein the content of the first and second substances,
the shape parameters are represented by a representation of the shape parameters,
representing the number of samples of each group of the cluster grouping samples,
sample points representing the ith cluster grouping samples.
5. The aircraft complex structure approximate modeling method according to claim 4, characterized in that the step of calculating prediction errors of said agent model and calculating root mean square errors of all said prediction errors using a fast cross-validation method based on a plurality of groups of said cluster grouping samples comprises:
selecting the ith group of clustering grouping samples
As a test sample and using the clustering grouping samples of the rest groups as training samples;
k is the total number of the clustering grouping samples;
calculating the prediction error of the agent model based on the training sample on the test sample by adopting a set rapid cross validation method; the set rapid cross-validation method comprises the following steps:
wherein the content of the first and second substances,
corresponding to said test sample in an inverse matrix of a matrix of coefficients representing all samples
Go to,
Of overlapping parts of column elements
The order of the matrix is such that,
representing the correspondence of test samples
The number of the individual weight coefficients is,
representing the current proxy model in the test sample
The prediction error of (a);
returning to execute the selected ith group of clustering grouping samples
As test specimensTaking the cluster grouping samples of the other groups as training samples, and starting to calculate the next group of cluster grouping samples until the prediction error of each group of cluster grouping samples is calculated;
calculating the root mean square error for each of the prediction errors; the calculation mode of the root mean square error is as follows:
wherein the content of the first and second substances,
the root mean square error is represented as a function of,
representing the total number of samples of said sample data,
the true output of the sample is represented by,
representing the output of the proxy model prediction.
6. The aircraft complex structure approximate modeling method according to claim 2, wherein the step of performing cluster center recalculation on the K clustered samples respectively and outputting K groups of clustered samples obtained by current clustering processing comprises:
carrying out cluster center recalculation on each clustered sample in a set center calculation mode; the center calculation mode is as follows:
wherein the content of the first and second substances,
is shown as
The cluster center of each of the clusters is,
which represents the ith sample point, is,
represents a random selection
The center of each of the clusters is,
is shown as
The number of samples of an individual cluster;
adopting a set convergence criterion to recalculate the clustering center to perform convergence judgment, and outputting a clustering sample of current clustering processing if the convergence judgment is performed; the convergence criterion is:
wherein the content of the first and second substances,
representing the euclidean distance of two sample points,
representing a set euclidean distance threshold.
7. The aircraft complex structure approximate modeling method according to claim 6, wherein the step of performing cluster center recalculation on the K clustered samples respectively and outputting K groups of clustered samples obtained by current clustering processing further comprises:
if it is determined not to converge, it will
As the next step
And respectively calculating the Euclidean distance from each sample point to the center of each cluster, distributing each sample point to the center of the cluster with the nearest distance to obtain K clustered samples, and performing the next iterative calculation.
8. An aircraft complex structure approximate modeling apparatus, comprising:
the system comprises a clustering processing module, a data processing module and a data processing module, wherein the clustering processing module is used for acquiring a plurality of sample data of an aircraft structure and clustering each sample data by adopting a K-means clustering method;
the clustering grouping module is used for grouping the clustered samples by adopting a proportional allocation method to obtain a plurality of groups of clustering grouping samples; each group of clustering grouping samples comprises clustering information;
the agent construction module is used for constructing an agent model for predicting the performance of the aircraft structure by adopting a local density-based enhanced anisotropic radial basis method according to the plurality of groups of clustering grouping samples;
the prediction error module is used for calculating the prediction error of the agent model by adopting a rapid cross validation method according to a plurality of groups of clustering grouping samples and calculating the root mean square error of all the prediction errors;
the optimization processing module is used for optimizing the anisotropic scaling coefficient according to the root mean square error by taking the root mean square error as a target function and taking the anisotropic scaling coefficient as a design variable to obtain a target scaling coefficient corresponding to the minimum root mean square error; the anisotropic scaling factor is used for constructing the proxy model;
the prediction execution module is used for predicting the performance of the aircraft structure by using the finally constructed high-precision performance prediction model; the high-precision performance prediction model is the target scaling factor.
9. A computer device comprising a memory and a processor, said memory storing a computer program, characterized in that said processor, when executing said computer program, implements the steps of the aircraft complex structure approximation modeling method according to any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the aircraft complex structure approximation modeling method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110716401.7A CN113255235B (en) | 2021-06-28 | 2021-06-28 | Approximate modeling method, device, equipment and medium for complex structure of aircraft |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110716401.7A CN113255235B (en) | 2021-06-28 | 2021-06-28 | Approximate modeling method, device, equipment and medium for complex structure of aircraft |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113255235A CN113255235A (en) | 2021-08-13 |
| CN113255235B true CN113255235B (en) | 2021-09-24 |
Family
ID=77189784
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110716401.7A Active CN113255235B (en) | 2021-06-28 | 2021-06-28 | Approximate modeling method, device, equipment and medium for complex structure of aircraft |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255235B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114048544B (en) * | 2021-11-09 | 2022-08-16 | 中国人民解放军国防科技大学 | Aircraft structure response characteristic prediction method, device, equipment and medium |
| CN114200843B (en) * | 2022-02-15 | 2022-04-19 | 中国人民解放军国防科技大学 | Throat-bolt type solid attitude and orbit control thrust rapid prediction method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2378447A1 (en) * | 2010-04-13 | 2011-10-19 | Livermore Software Technology Corporation | Identification of most influential design variables in engineering design optimization |
| CN112070125A (en) * | 2020-08-19 | 2020-12-11 | 西安理工大学 | Prediction method of unbalanced data set based on isolated forest learning |
| CN112231839A (en) * | 2020-10-23 | 2021-01-15 | 北京理工大学 | Intelligent exploration method for peak-tracking sampling of aircraft aiming at high time-consuming constraint |
| CN112529683A (en) * | 2020-11-27 | 2021-03-19 | 百维金科(上海)信息科技有限公司 | Method and system for evaluating credit risk of customer based on CS-PNN |
| CN112613123A (en) * | 2020-12-25 | 2021-04-06 | 成都飞机工业（集团）有限责任公司 | AR three-dimensional registration method and device for aircraft pipeline |
| CN112906295A (en) * | 2021-01-28 | 2021-06-04 | 中国人民解放军国防科技大学 | Method, device and equipment for predicting burning speed of solid rocket engine |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9239956B2 (en) * | 2012-06-28 | 2016-01-19 | Oliver Hein | Method and apparatus for coding of eye and eye movement data |
-
2021
- 2021-06-28 CN CN202110716401.7A patent/CN113255235B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2378447A1 (en) * | 2010-04-13 | 2011-10-19 | Livermore Software Technology Corporation | Identification of most influential design variables in engineering design optimization |
| CN112070125A (en) * | 2020-08-19 | 2020-12-11 | 西安理工大学 | Prediction method of unbalanced data set based on isolated forest learning |
| CN112231839A (en) * | 2020-10-23 | 2021-01-15 | 北京理工大学 | Intelligent exploration method for peak-tracking sampling of aircraft aiming at high time-consuming constraint |
| CN112529683A (en) * | 2020-11-27 | 2021-03-19 | 百维金科(上海)信息科技有限公司 | Method and system for evaluating credit risk of customer based on CS-PNN |
| CN112613123A (en) * | 2020-12-25 | 2021-04-06 | 成都飞机工业（集团）有限责任公司 | AR three-dimensional registration method and device for aircraft pipeline |
| CN112906295A (en) * | 2021-01-28 | 2021-06-04 | 中国人民解放军国防科技大学 | Method, device and equipment for predicting burning speed of solid rocket engine |
Non-Patent Citations (2)
| Title |
|---|
| Global sensitivity analysis using orthogonal augmented radial basis function;Zeping Wu;《Reliability Engineering and System Safety》;20190103;第185卷;第291-302页 * |
| 大型运载火箭加筋柱壳近似建模方法;王志祥;《宇航学报》;20201031;第41卷(第10期);第1267-1279页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113255235A (en) | 2021-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255235B (en) | Approximate modeling method, device, equipment and medium for complex structure of aircraft | |
| Esfandyari et al. | A tuned version of genetic algorithm for efficient test suite generation in interactive t-way testing strategy | |
| CN113343545B (en) | Structure adaptive optimization design method, device, equipment and medium | |
| CN112434448A (en) | Proxy model constraint optimization method and device based on multipoint adding | |
| Ye et al. | Ensemble of surrogate based global optimization methods using hierarchical design space reduction | |
| KR101544457B1 (en) | The method for parameter investigation to optimal design | |
| CN112749495A (en) | Multipoint-point-adding-based proxy model optimization method and device and computer equipment | |
| CN115755954B (en) | Routing inspection path planning method, system, computer equipment and storage medium | |
| CN112734239B (en) | Task planning method, device and medium based on task and resource capability attributes | |
| CN111783209B (en) | Self-adaptive structure reliability analysis method combining learning function and kriging model | |
| CN114154371A (en) | Multi-objective optimization method for reinforced gearbox based on response surface method | |
| CN114297779A (en) | Method, system and equipment for quickly predicting aerodynamic coefficient of aircraft | |
| CN115438589A (en) | Fishing rod adjustability prediction model based on BP neural network and optimization method thereof | |
| CN113821983A (en) | Engineering design optimization method and device based on proxy model and electronic equipment | |
| CN113919484A (en) | Structured pruning method and device based on deep convolutional neural network model | |
| CN112818470A (en) | Optimization method and device of base structure, computer equipment and storage medium | |
| CN109074348A (en) | For being iterated the equipment and alternative manner of cluster to input data set | |
| CN114564787A (en) | Bayesian optimization method, device and storage medium for target-related airfoil design | |
| CN114154615A (en) | Neural architecture searching method and device based on hardware performance | |
| CN112819246A (en) | Energy demand prediction method for optimizing neural network based on cuckoo algorithm | |
| CN115391745B (en) | Rainfall forecast correction method and system based on probability matching average method | |
| CN107038244A (en) | A kind of data digging method and device, a kind of computer-readable recording medium and storage control | |
| KR102559605B1 (en) | Method and apparatus for function optimization | |
| CN114595577A (en) | Staged efficient constraint optimization method and device based on Kriging agent model | |
| CN113128771B (en) | Expensive function optimization method and device for parallel differential evolution algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |