CN116842853A - Missile aerodynamic characteristic prediction model construction method for uncertainty quantization - Google Patents

Missile aerodynamic characteristic prediction model construction method for uncertainty quantization Download PDF

Info

Publication number
CN116842853A
CN116842853A CN202311120878.4A CN202311120878A CN116842853A CN 116842853 A CN116842853 A CN 116842853A CN 202311120878 A CN202311120878 A CN 202311120878A CN 116842853 A CN116842853 A CN 116842853A
Authority
CN
China
Prior art keywords
sample
missile
prediction model
missile aerodynamic
prediction error
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.)
Granted
Application number
CN202311120878.4A
Other languages
Chinese (zh)
Other versions
CN116842853B (en
Inventor
陈坚强
陈江涛
章超
赵炜
张培红
赵娇
肖维
吕罗庚
胡向鹏
沈盈盈
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Computational Aerodynamics Institute of China Aerodynamics Research and Development Center
Original Assignee
Computational Aerodynamics Institute of China Aerodynamics Research and Development Center
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Computational Aerodynamics Institute of China Aerodynamics Research and Development Center filed Critical Computational Aerodynamics Institute of China Aerodynamics Research and Development Center
Priority to CN202311120878.4A priority Critical patent/CN116842853B/en
Publication of CN116842853A publication Critical patent/CN116842853A/en
Application granted granted Critical
Publication of CN116842853B publication Critical patent/CN116842853B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/08—Fluids
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Optimization (AREA)
  • Genetics & Genomics (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Automation & Control Theory (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Pure & Applied Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application provides a missile aerodynamic characteristic prediction model construction method for uncertainty quantization, which comprises the following steps: step 1, acquiring an initial sample, constructing an initial missile aerodynamic characteristic prediction model, and evaluating whether a prediction error of the prediction model meets the requirement. Step 2, when the prediction error does not meet the requirement, determining a sequential sample by adopting a cross-validation criterion, and adding and updating a training sample; step 3, updating a missile aerodynamic characteristic prediction model by using the updated training sample, and evaluating whether a prediction error meets the requirement; and step 4, obtaining a missile aerodynamic characteristic prediction model meeting the prediction error requirement, wherein the missile aerodynamic characteristic prediction model can be used for quantifying uncertainty of missile aerodynamic characteristics. The application can obviously save the computing resources and reduce the computing cost.

Description

Missile aerodynamic characteristic prediction model construction method for uncertainty quantization
Technical Field
The application relates to the field of missile aerodynamic feature analysis, in particular to a missile aerodynamic feature prediction model construction method for uncertainty quantification.
Background
In the real flight process of the missile, the disturbance of the surrounding air environment can cause the fluctuation of the aerodynamic characteristics of the missile, and even cause the failure of tasks or equipment failure when serious. Therefore, in the missile design and development process, uncertainty of incoming flow conditions generated by small changes of the flight environment needs to be fully considered, influence of the uncertainty on the aerodynamic characteristics of the missile is analyzed, and robustness and reliability of equipment are ensured.
In the field of uncertainty quantization, the monte carlo method is the simplest straightforward method. But this method requires a large amount of sample data. If each sample is obtained by CFD calculations, this is computationally expensive for an actual missile. Therefore, development of a missile aerodynamic characteristic rapid prediction method meeting the uncertainty quantization precision requirement is urgently needed to replace CFD calculation and meet the uncertainty quantization requirement.
Disclosure of Invention
Aiming at the problems existing in the prior art, the missile aerodynamic characteristic prediction model construction method for uncertainty quantization is provided, so that the calculation cost of uncertainty quantization analysis can be effectively reduced, and the resource consumption is saved.
The technical scheme adopted by the application is as follows: the missile aerodynamic characteristic prediction model construction method for uncertainty quantization comprises the following steps:
step 1, acquiring an initial sample, constructing an initial missile aerodynamic characteristic prediction model, evaluating whether a prediction error of the prediction model meets requirements, if yes, entering a step 4, otherwise, entering a step 2; the initial sample comprises a training sample and a test sample;
step 2, determining sequential samples by adopting a cross-validation criterion, and adding and updating training samples;
step 3, updating a missile aerodynamic characteristic prediction model by using the updated training sample, evaluating whether a prediction error meets the requirement, if so, entering a step 4, otherwise, entering a step 2;
and step 4, obtaining a missile aerodynamic characteristic prediction model meeting the prediction error requirement, wherein the model is used for quantifying the uncertainty of the missile aerodynamic characteristic.
Further, the step 1 includes the following substeps:
step 101, extracting N training sample inputs and M test sample inputs from the test uncertain input;
102, respectively transmitting N training sample inputs and M test sample inputs to a missile aerodynamic characteristic solver to obtain corresponding missile aerodynamic characteristic concerned outputs, and respectively forming the training samples and the test samples;
step 103, constructing a missile aerodynamic characteristic prediction model by using a training sample;
and 104, evaluating the prediction error of the missile aerodynamic characteristic prediction model by using a test sample, if the prediction error is smaller than a preset prediction error requirement, entering a step 4, otherwise, entering a step 2.
Further, the test uncertainty in step 101 is input as a target of quantitative analysis of aerodynamic characteristics of the missile, such as an incoming flow condition.
Further, the step 2 includes the following substeps:
step 201, sequentially removing one sample from N training samples, and constructing missile aerodynamic characteristic prediction models by using the rest samples to obtain N prediction models;
step 202, constructing a prediction error evaluation function, calculating the minimum value of the prediction error evaluation function through a genetic algorithm, and determining sequential sample input;
step 203, obtaining a sample output corresponding to a sequential sample input through a missile aerodynamic feature solver, forming 1 sequential sample by the sequential sample input and the corresponding sample output, and adding the sequential sample input and the corresponding sample output into a training sample, so that the data size of the training sample is added by 1.
Further, the step 3 includes the following substeps:
step 301, constructing a missile aerodynamic characteristic prediction model by using the updated training sample;
step 302, evaluating a prediction error of a missile aerodynamic characteristic prediction model by using a test sample; if the prediction error is smaller than the preset prediction error requirement, the step 4 is carried out, otherwise, the step 2 is returned.
Further, in the steps 104 and 302, the prediction error calculating method is as follows:
wherein ,for prediction error, M is the number of test samples, < ->In order to correspond to the output of the test sample,indicating that the test sample input is +.>Is used for predicting the model predictive value.
Further, in the step 202, the method for constructing the prediction error estimation function is as follows:
for arbitrary sample inputThe prediction error evaluation function is as follows:
where N is the number of predictive models, n=n,representing the input sample as +.>Is the i-th predictive model of (2)>Representing the input sample as +.>Is a model of initial missile aerodynamic feature prediction.
Compared with the prior art, the beneficial effects of adopting the technical scheme are as follows: the calculation cost required by the missile aerodynamic characteristic prediction model established by the application when reaching the preset prediction error is about 74% of that of the original method for constructing the prediction model based on random sampling, so that the calculation cost is remarkably saved, and the calculation resource is effectively saved; meanwhile, a prediction error evaluation function is constructed based on a cross verification method, training samples are gradually increased in a key area affecting the accuracy of the missile aerodynamic characteristic prediction model until the prediction model meets the accuracy requirement of uncertainty quantification, and the waste of calculation resources in the area with little effect on improving the accuracy of the missile aerodynamic characteristic prediction model is avoided.
Drawings
FIG. 1 is a flow chart of a missile aerodynamic feature prediction model construction method for uncertainty quantization.
Detailed Description
Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar modules or modules having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the application. On the contrary, the embodiments of the application include all alternatives, modifications and equivalents as may be included within the spirit and scope of the appended claims.
As shown in fig. 1, the embodiment of the application provides a missile aerodynamic characteristic prediction model construction method for uncertainty quantization, which comprises the following steps:
step 1, acquiring an initial sample, constructing an initial missile aerodynamic characteristic prediction model, evaluating whether a prediction error of the prediction model meets requirements, if yes, entering a step 4, otherwise, entering a step 2; the initial sample comprises a training sample and a test sample;
step 2, determining sequential samples by adopting a cross-validation criterion, and adding and updating training samples;
step 3, updating a missile aerodynamic characteristic prediction model by using the updated training sample, evaluating whether a prediction error meets the requirement, if so, entering a step 4, otherwise, entering a step 2;
and step 4, obtaining a missile aerodynamic characteristic prediction model meeting the prediction error requirement, wherein the model is used for quantifying the uncertainty of the missile aerodynamic characteristic.
In this embodiment, the test sample data is composed of variables such as flow conditions, etc. required for quantitative analysis of specified aerodynamic characteristics of the missile.
Specifically, the step 1 includes the following steps:
step 101, training samples and test samples are sampled.
Sampling by random sampling method to obtain N training samples of input information. wherein />For incoming flow conditions (Mach number Ma, angle of attackαSideslip angleβEtc.).
Sampling by random sampling method to obtain M test samplesIs input information of (a)。
Step 102, transmitting the input information of each training sample to a missile aerodynamic feature solver to obtain missile aerodynamic feature attention output (such as resistance coefficient C d ) Composing training sample outputs. The input information of each test sample is transmitted to a missile aerodynamic feature solver to obtain missile aerodynamic feature attention output, and the missile aerodynamic feature attention output is formed into test sample output. The ith sample is composed of input information and attention output +.>Composition is prepared.
Step 103, useNConstructing missile aerodynamic characteristic prediction model by using training samples。
Step 104, evaluating the predictive model on the test samplePrediction error of +.>：
wherein ,for prediction error, M is the number of test samples, < ->In order to correspond to the output of the test sample,indicating that the test sample input is +.>Is used for predicting the model predictive value.
To calculate the prediction errorWith a preset prediction error requirement +.>Compare if->Less thanConsidering that the prediction model meets the requirement to terminate the process; if->Greater than or equal to->The model is improved by subsequent steps.
If the prediction error of the initial missile aerodynamic characteristic prediction model does not meet the requirement, the training sample is required to be updated, and the specific process is as shown in the step 2:
and 201, sequentially removing one sample from N training samples, and constructing a missile aerodynamic characteristic prediction model by using the rest N-1 samples.
And 202, constructing a prediction error evaluation function. For arbitrary sample inputA prediction error assessment function is constructed of the form:
where N is the number of predictive models, n=n,representing the input sample as +.>Is the i-th predictive model of (2)>Representing the input sample as +.>Is a model of initial missile aerodynamic feature prediction.
Step 203, solving the minimum value of the prediction error evaluation function through a genetic algorithm to obtain a sequential sample input. Wherein, the genetic algorithm is a classical optimizing algorithm, and the genetic algorithm is adopted by the prior algorithm and is not described in detail herein
Step 204, inputting the sequential samplesTransmitting the result to a missile aerodynamic characteristic solver to obtain the missile aerodynamic characteristic attention output +.>. Will-> and />The sequential samples are added into the training samples, and the data volume of the samplesNAdding 1 to obtain updated N+1 training samples.
After the training sample is updated, the missile aerodynamic characteristic prediction model is built again, and the specific process is as follows:
step 301, constructing a missile aerodynamic characteristic prediction model by using the updated training sample;
step 302, evaluating a prediction error of a missile aerodynamic characteristic prediction model by using a test sample; if forecastThe error is smaller than the preset prediction errorIf the prediction model meets the requirement, the process is terminated, otherwise, the loop is returned to optimize until the prediction error is less than +.>The optimization process is terminated.
In order to better illustrate the method for constructing the prediction model, the uncertainty analysis of the resistance coefficient of a certain missile is taken as an example.
The step 1 comprises the following steps:
(1) Under different incoming flow conditions (Mach number Ma, angle of attackαSideslip angleβ) A total of 364 test sample data were required.
(2) Using random sampling to select 20 initial training sample inputs in a given test design, i.eN=20；
(3) Selecting 30 test sample inputs in a given test design using random sampling, i.eM=30；
(4) Transmitting the input parameters of each training sample to a missile aerodynamic feature solver to obtain the missile resistance coefficient of each sample to form a training sample output；
(5) Transmitting the input parameters of each test sample to a missile aerodynamic characteristic solver to obtain the missile resistance coefficient of each sample to form test sample output；
(6) Constructing a Kriging prediction model of the missile resistance coefficient by using 20 initial training samples;
(7) Assessing prediction error of predictive model over 30 test samples0.2654;
(8) Model prediction error is greater thanDemand valueThus, subsequent model improvements are needed.
The step 2 comprises the following steps:
(1) And sequentially removing one sample from 20 training samples, and constructing a Kerling predictive model of the missile resistance coefficient by using the rest 19 samples. Cycling for 20 times to obtain 20 prediction models, which are recorded as。
(2) A prediction error estimation function is constructed,；
(3) Solving the minimum value of the prediction error evaluation function through a genetic algorithm to obtain the input of the next sample；
(4) Will beTransmitting to missile aerodynamic characteristics solver to obtain resistance coefficient +.>. Will-> and />Adding into training sample, the existing training sample isN=21.
The step 3 comprises the following steps:
(1) Constructing a Kriging prediction model of the missile resistance coefficient by using 21 training samples;
(2) Evaluating prediction errors of a predictive model on test samples0.2256;
(3) The model prediction error is still greater than the required valueThus, step 2 and step 3 are cycled.
After 45 cycles, the prediction error of the prediction modelIs 0.0961, less than the required value +.>At this point the optimization process is terminated.
And finally, 65 training samples are used to obtain the missile resistance coefficient prediction model for uncertainty quantization.
To verify the advantages of the method of the present application, it is compared with the original method of constructing a predictive model based on random sample data. In order to eliminate the influence of random sample selection, the method is repeated 50 times with the original method, and the number of training samples reaching the resistance coefficient prediction error requirement is compared. The comparison results are shown in Table 1. The average number of samples required for the method of the present application was 68.68 and the average number of samples required for the original method was 92.9. The calculated amount of the method is 74% of that of the original method, and the calculation cost is obviously reduced.
TABLE 1 prediction errorSample size under requirement (repeat 50 times)
Test number Original random sampling method The self-adapting method of the application
1 112 91
2 102 128
3 84 61
4 148 42
5 112 36
6 121 44
7 81 80
8 85 65
9 145 65
10 20 49
11 61 61
12 135 129
13 65 67
14 94 45
15 72 55
16 71 66
17 111 85
18 79 63
19 105 80
20 80 47
21 55 52
22 107 92
23 105 54
24 98 78
25 58 74
26 85 114
27 96 79
28 81 61
29 93 89
30 95 39
31 82 78
32 126 57
33 92 63
34 83 79
35 134 72
36 94 67
37 60 90
38 127 79
39 75 30
40 79 81
41 157 49
42 61 55
43 118 53
44 84 69
45 86 57
46 65 93
47 110 58
48 83 95
49 78 53
50 95 65
Table 2 shows the results of a specific test, the initial random sampling method and the model prediction error established by the application, along with the increase of the number of training samples. Also to achieve the resistance coefficient prediction error, the method of the present application requires 65 training samples, whereas the original random sampling modeling method requires 95 training samples.
TABLE 2 prediction error trend for certain experiment
Training sample number Original random sampling method The self-adapting method of the application
20 0.437206245 0.437206245
21 0.43130229 0.335908065
22 0.430921913 0.295174883
23 0.434048742 0.272721716
24 0.33470187 0.356185795
25 0.311551672 0.332658955
26 0.248820493 0.250084946
27 0.337384867 0.245693394
28 0.325768523 0.228764115
29 0.321424776 0.228001873
30 0.232577135 0.207948976
31 0.2312581 0.232819014
32 0.230373288 0.195117628
33 0.232008227 0.196952539
34 0.235353283 0.185788169
35 0.236013367 0.191669412
36 0.230841041 0.162164724
37 0.215850847 0.152520609
38 0.247600361 0.2290124
39 0.247320815 0.230687138
40 0.246539464 0.226359237
41 0.230000515 0.225519685
42 0.230261523 0.223075272
43 0.220049172 0.21908814
44 0.22105031 0.194580524
45 0.220923522 0.191965729
46 0.218586242 0.192240143
47 0.219248825 0.189862812
48 0.215445892 0.189555278
49 0.214344351 0.182269424
50 0.216895305 0.18074156
51 0.21549348 0.135762679
52 0.21479397 0.150032928
53 0.379898375 0.145894741
54 0.325643188 0.141205187
55 0.324956667 0.128811557
56 0.323362353 0.131640448
57 0.323285474 0.130508891
58 0.32430369 0.131385241
59 0.324405442 0.131817684
60 0.323654549 0.130789003
61 0.321994757 0.134603389
62 0.258308519 0.126459196
63 0.253319 0.126038
64 0.282197 0.096092
65 0.371258 0.096986
66 0.399114 0.09651
67 0.391158 0.095719
68 0.26846 0.113937
69 0.267756 0.112052
70 0.269289 0.122219
71 0.267487 0.059062
72 0.292647 0.059461
73 0.229371 0.058254
74 0.248845 0.056757
75 0.268557 0.099374
76 0.248834 0.071998
77 0.272008 0.075459
78 0.283323 0.073683
79 0.216828 0.073577
80 0.166995 0.07388
81 0.166024 0.072474
82 0.16392 0.073375
83 0.206144 0.060349
84 0.206271 0.090678
85 0.282978 0.055405
86 0.284802 0.046476
87 0.284384 0.049677
88 0.284553 0.049754
89 0.269854 0.049174
90 0.283847 0.048113
91 0.180784 0.048204
92 0.269488 0.039767
93 0.249107 0.039014
94 0.098193 0.026789
It should be noted that, in the description of the embodiments of the present application, unless explicitly specified and limited otherwise, the terms "disposed," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; may be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present application will be understood in detail by those skilled in the art; the accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. The components of the embodiments of the present application generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
While embodiments of the present application have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the application, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the application.

Claims (7)

1. The missile aerodynamic characteristic prediction model construction method for uncertainty quantification is characterized by comprising the following steps of:
step 1, acquiring an initial sample, constructing an initial missile aerodynamic characteristic prediction model, evaluating whether a prediction error of the prediction model meets requirements, if yes, entering a step 4, otherwise, entering a step 2; the initial sample comprises a training sample and a test sample;
step 2, determining sequential samples by adopting a cross-validation criterion, and adding and updating training samples;
step 3, updating a missile aerodynamic characteristic prediction model by using the updated training sample, evaluating whether a prediction error meets the requirement, if so, entering a step 4, otherwise, entering a step 2;
and step 4, obtaining a missile aerodynamic characteristic prediction model meeting the prediction error requirement, wherein the model is used for quantifying the uncertainty of the missile aerodynamic characteristic.
2. The missile aerodynamic feature prediction model construction method for uncertainty quantization according to claim 1, wherein the step 1 includes the sub-steps of:
step 101, extracting N training sample inputs and M test sample inputs from the test uncertain input;
102, respectively transmitting N training sample inputs and M test sample inputs to a missile aerodynamic characteristic solver to obtain corresponding missile aerodynamic characteristic concerned outputs, and respectively forming the training samples and the test samples;
step 103, constructing a missile aerodynamic characteristic prediction model by using a training sample;
and 104, evaluating the prediction error of the missile aerodynamic characteristic prediction model by using a test sample, if the prediction error is smaller than a preset prediction error requirement, entering a step 4, otherwise, entering a step 2.
3. The missile aerodynamic feature prediction model construction method for uncertainty quantization according to claim 1 or 2, wherein the experimental uncertainty input in step 101 is the object of the missile aerodynamic feature uncertainty quantization analysis, including the incoming flow conditions.
4. The missile aerodynamic feature prediction model construction method for uncertainty quantization according to claim 2, wherein the step 2 includes the sub-steps of:
step 201, sequentially removing one sample from N training samples, and constructing missile aerodynamic characteristic prediction models by using the rest samples to obtain N prediction models;
step 202, constructing a prediction error evaluation function, calculating the minimum value of the prediction error evaluation function through a genetic algorithm, and determining sequential sample input;
step 203, obtaining a sample output corresponding to a sequential sample input through a missile aerodynamic feature solver, forming 1 sequential sample by the sequential sample input and the corresponding sample output, and adding the sequential sample input and the corresponding sample output into a training sample, so that the data size of the training sample is added by 1.
5. The missile aerodynamic feature prediction model construction method for uncertainty quantization according to claim 4, wherein the step 3 includes the sub-steps of:
step 301, constructing a missile aerodynamic characteristic prediction model by using the updated training sample;
step 302, evaluating a prediction error of a missile aerodynamic characteristic prediction model by using a test sample; if the prediction error is smaller than the preset prediction error requirement, the step 4 is carried out, otherwise, the step 2 is returned.
6. The missile aerodynamic feature prediction model construction method for uncertainty quantization according to claim 5, wherein in the steps 104 and 302, the prediction error calculation method is as follows:
wherein ,for prediction error, M is the number of test samples, < ->For outputting corresponding test sample->Indicating that the test sample input is +.>Is used for predicting the model predictive value.
7. The method for constructing a prediction model of missile aerodynamic characteristics for uncertainty quantization according to claim 4, wherein in the step 202, the method for constructing a prediction error evaluation function is as follows:
for arbitrary sample inputThe prediction error evaluation function is as follows:
where N is the number of predictive models, n=n,representing the input sample as +.>Is the i-th predictive model of (2)>Representing the input sample as +.>Is a model of initial missile aerodynamic feature prediction.
CN202311120878.4A 2023-09-01 2023-09-01 Missile aerodynamic characteristic prediction model construction method for uncertainty quantization Active CN116842853B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311120878.4A CN116842853B (en) 2023-09-01 2023-09-01 Missile aerodynamic characteristic prediction model construction method for uncertainty quantization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311120878.4A CN116842853B (en) 2023-09-01 2023-09-01 Missile aerodynamic characteristic prediction model construction method for uncertainty quantization

Publications (2)

Publication Number Publication Date
CN116842853A true CN116842853A (en) 2023-10-03
CN116842853B CN116842853B (en) 2023-11-28

Family

ID=88174722

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311120878.4A Active CN116842853B (en) 2023-09-01 2023-09-01 Missile aerodynamic characteristic prediction model construction method for uncertainty quantization

Country Status (1)

Country Link
CN (1) CN116842853B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117648846A (en) * 2024-01-30 2024-03-05 中国空气动力研究与发展中心计算空气动力研究所 Image sample generation method for composite material performance prediction modeling

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101633360B1 (en) * 2015-01-12 2016-06-27 한국과학기술원 Sequential sampling method based kriging model for structure condition assessment
WO2021121128A1 (en) * 2020-06-08 2021-06-24 平安科技（深圳）有限公司 Artificial intelligence-based sample evaluation method, apparatus, device, and storage medium
CN114462309A (en) * 2022-01-19 2022-05-10 苏州浪潮智能科技有限公司 Physical simulation model training method, device, equipment and storage medium
CN114757131A (en) * 2022-06-13 2022-07-15 中国空气动力研究与发展中心计算空气动力研究所 Optimization method of proxy model suitable for CFD uncertainty quantification and related equipment

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101633360B1 (en) * 2015-01-12 2016-06-27 한국과학기술원 Sequential sampling method based kriging model for structure condition assessment
WO2021121128A1 (en) * 2020-06-08 2021-06-24 平安科技（深圳）有限公司 Artificial intelligence-based sample evaluation method, apparatus, device, and storage medium
CN114462309A (en) * 2022-01-19 2022-05-10 苏州浪潮智能科技有限公司 Physical simulation model training method, device, equipment and storage medium
CN114757131A (en) * 2022-06-13 2022-07-15 中国空气动力研究与发展中心计算空气动力研究所 Optimization method of proxy model suitable for CFD uncertainty quantification and related equipment

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ABBASZADEH, P等: "The quest for model uncertainty quantification:A hybrid ensemble and variation data assimilation framewor", 《WATER RESOURCES RESEARCH》, pages 2407 - 2431 *
王栋 等: "脉冲 Ｘ 射线辐照铝板气化反冲仿真的不确定度量化方法", 《现代应用物理》, vol. 12, no. 2, pages 020602 - 1 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117648846A (en) * 2024-01-30 2024-03-05 中国空气动力研究与发展中心计算空气动力研究所 Image sample generation method for composite material performance prediction modeling
CN117648846B (en) * 2024-01-30 2024-04-26 中国空气动力研究与发展中心计算空气动力研究所 Image sample generation method for composite material performance prediction modeling

Also Published As

Publication number Publication date
CN116842853B (en) 2023-11-28

Similar Documents

Publication Publication Date Title
Dupuis Exceedances over high thresholds: A guide to threshold selection
CN112434448B (en) Proxy model constraint optimization method and device based on multipoint adding
CN111339488B (en) Slope system failure probability calculation method based on Kriging
CN116842853B (en) Missile aerodynamic characteristic prediction model construction method for uncertainty quantization
CN113259325B (en) Network security situation prediction method for optimizing Bi-LSTM based on sparrow search algorithm
CN111625516B (en) Method, apparatus, computer device and storage medium for detecting data state
WO2021092366A1 (en) Methods and systems for the estimation of the computational cost of simulation
CN111881023B (en) Software aging prediction method and device based on multi-model comparison
CN113221263A (en) Mechanical product structure failure optimization method considering distribution parameter uncertainty
CN115221801B (en) Aircraft uncertainty propagation analysis method and device based on dynamic approximate modeling
CN113627098A (en) CFD model confirmation method and product design method
CN111931983A (en) Precipitation prediction method and system
CN115935834A (en) History fitting method based on deep autoregressive network and continuous learning strategy
CN115982141A (en) Characteristic optimization method for time series data prediction
CN116822214A (en) Multi-credibility data fusion method based on heteroscedastic Gaussian process regression
CN113408070A (en) Method, device and equipment for determining engine parameters and storage medium
CN114627963B (en) Protein data filling method, system, computer device and readable storage medium
CN116595865A (en) Equipment residual life prediction method based on wiener process
CN115392441A (en) Method, apparatus, device and medium for on-chip adaptation of quantized neural network model
CN115389743A (en) Method, medium and system for predicting content interval of dissolved gas in transformer oil
CN109284320B (en) Automatic regression diagnosis method on big data platform
CN110929849B (en) Video detection method and device based on neural network model compression
CN111783236A (en) Turbine casing sensitivity analysis method based on self-adaptive model and subset simulation
CN116780524B (en) Industrial enterprise short-term load prediction method based on LSTM deep learning
CN117131977B (en) Runoff forecasting sample set partitioning method based on misjudgment risk minimum criterion

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