CN116467934A - Carbon fiber preform parameter optimization method - Google Patents
Carbon fiber preform parameter optimization method Download PDFInfo
- Publication number
- CN116467934A CN116467934A CN202310309502.1A CN202310309502A CN116467934A CN 116467934 A CN116467934 A CN 116467934A CN 202310309502 A CN202310309502 A CN 202310309502A CN 116467934 A CN116467934 A CN 116467934A
- Authority
- CN
- China
- Prior art keywords
- carbon fiber
- neural network
- parameters
- fiber preform
- needling
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 229920000049 Carbon (fiber) Polymers 0.000 title claims abstract description 47
- 239000004917 carbon fiber Substances 0.000 title claims abstract description 47
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 title claims abstract description 47
- 238000005457 optimization Methods 0.000 title abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 62
- 238000012549 training Methods 0.000 claims abstract description 36
- 239000003016 pheromone Substances 0.000 claims abstract description 31
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 239000000835 fiber Substances 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 abstract 1
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C60/00—Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/26—Composites
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Databases & Information Systems (AREA)
- Chemical & Material Sciences (AREA)
- Crystallography & Structural Chemistry (AREA)
- Nonwoven Fabrics (AREA)
Abstract
The invention relates to a carbon fiber preform parameter optimization method, which comprises the steps of collecting needling process parameters to construct sample data, wherein each sample comprises a group of needling process parameters and the mechanical properties of a corresponding carbon fiber preform; constructing a BP neural network model, and encoding a network weight and a threshold value; optimizing network parameters by adopting an improved ant colony algorithm, and taking a prediction error of a sample as an individual pheromone increment; training and verifying the neural network by using the sample training set and the test set to obtain an optimal neural network weight and a threshold value, and taking the optimal neural network weight and the threshold value as a needling parameter selection model; encoding the needling parameters in a needling parameter definition domain, and randomly generating an input vector of the neural network; and comparing the input mechanical properties of the carbon fiber preform with the predicted values to obtain the optimal needling process parameters. The method can rapidly and accurately select the needling process parameters of the mechanical properties required by the carbon fiber preform.
Description
Technical Field
The invention relates to the technical field of neural networks and needling, in particular to a carbon fiber preform parameter optimization method.
Background
The needling method is a typical mechanical consolidation processing method for manufacturing multi-layer fabric structures such as fiber felts and the like, and has become a carbon fiber preform preparation process with wide application prospect. The mechanical properties of the carbon fiber preform can be affected by the technological parameters such as the needle structure, the number of the needles, the needling density, the needling depth, the thickness of the fiber layer, the fiber layering mode and the like.
At present, common optimization methods of technological parameters comprise a field method, a principal component analysis method, an artificial neural network method, a partial least square method and the like. The neural network method is particularly suitable for modeling and predicting complex nonlinear systems, and is suitable for nonlinear characteristics between needling parameters and mechanical properties of the carbon fiber preform. However, for neural networks containing a large number of parameters, conventional back propagation algorithms tend to fall into local optima.
The optimization method for the neural network parameters also comprises a particle swarm algorithm, an atomic search algorithm, a genetic algorithm, an ant colony algorithm and the like. The traditional ant colony algorithm has low convergence speed, insufficient optimization capability and easy sinking into local optimum, but the selection of technological parameters usually hopes global optimum. Therefore, the patent optimizes the parameters of the neural network based on the improved ant colony algorithm, and provides a carbon fiber preform parameter optimization method.
Disclosure of Invention
The invention discloses a carbon fiber preform parameter optimization method which can accurately obtain optimal needling process parameters required by production according to the mechanical properties of a required needling composite material.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
a method for optimizing parameters of a carbon fiber preform, comprising:
step 1: constructing a sample database, wherein the database comprises needling process parameters and corresponding mechanical properties of the carbon fiber preform;
step 2: preprocessing a database, dividing a sample in the database into a training set and a testing set, and encoding needling process parameters in the database;
step 3: constructing a BP neural network, taking needling process parameters as BP neural network input layer nodes, taking the mechanical properties of the carbon fiber preform as BP neural network output layer nodes, and coding network weights and threshold values;
step 4: optimizing the weight and the threshold of the BP neural network by using an improved ant colony algorithm, setting the maximum iteration times, obtaining the weight and the threshold of the network by using a probability selection function added with information heuristic factors, calculating the output vector of the BP neural network after iteration, and selecting the neural network parameter with the minimum error of the output mechanical property vector and meeting the set error condition;
wherein the probability selection function is as follows:
in the above formula, t is the iteration number,representing the probability of the kth ant for the jth chosen value in the set of i (i=1, 2, 3, … n) parameters; τ ij The concentration of the pheromone of the j-th selected value in the i parameter set is represented; alpha is an information heuristic factor, reflecting the importance of the pheromone;
the error calculation formula is as follows:
in the above, E k (t) is the kth ant in the kth iterationThe error generated is V is the number of mechanical property parameters, Y e For the target output value, Y p Is a predicted output value; set E k ＜ε 1 Indicating that the error setting condition is satisfied, ε 1 Is the maximum error value allowed to be accepted;
the pheromone updating formula for improving the ant colony algorithm is as follows:
in the formula, Q is a fixed value of the pheromone, and rho is a volatile factor of the pheromone;
the adaptive formula of the pheromone volatilization factor is as follows:
step 5: randomly selecting M times in a database to obtain M groups of training sets, and training the BP neural network by adopting the M groups of training sets; when the set neural network training stopping condition is met, verifying the BP neural network by adopting a test set corresponding to the training set, and selecting a group of neural network parameters with the minimum mean square error of the final output vector to obtain a selected model of the needling parameters of the carbon fiber preform;
step 6: inputting the mechanical properties of the target carbon fiber preform by using the selection model of the needling parameters of the carbon fiber preform obtained in the step 5, and reversely obtaining needling process parameters corresponding to the mechanical properties of the target carbon fiber preform by using the BP neural network.
Further, in the step 1, needling process parameters include needling depth, needling density, fiber layering mode, and fiber thickness; the mechanical properties include tensile strength, compressive strength, and shear strength.
Further, in the step 2, 80% of the samples in the database are used as training sets, and the remaining 20% of the samples are used as test sets.
Further, in the step 2, the needling process parameters are encoded as follows: determining the value range of needling process parameters, equally dividing the definition domain F of the value range, enabling the needling process parameters in a database to be equal to the interval median value with the smallest difference value, and prescribing that the 90-degree cross ply of the fibers in the fiber ply mode is assigned to be 1 and the parallel ply of the fibers is assigned to be 0.
Further, the needling depth range in the needling process parameters is set to be [2,10 ]]mm, needling density of [15,55 ]]Needle/cm 2 The thickness of the fiber layer is in the range of [5,30 ]]mm。
Further, in the step 3, the network weight and the threshold are encoded as follows: determining the value ranges of the weight and the threshold, equally dividing the definition domain F of the weight and the threshold, and selecting the median value in the interval as a selected value; setting the number of nodes of an input layer of the BP neural network as a, the number of nodes of an intermediate layer as b, and the number of nodes of an output layer as c; let the number of weights be Nq, nq=a+b+b, let the number of thresholds be Ny, ny=2, and let the number of parameters be n, n=nq+ny.
Further, the weight and threshold are set to be in the range of [0,1].
Further, in the step 4, the number of ant colonies m=50, the maximum number of ant colony iterations t=100, the information heuristic factor α=3, the pheromone volatilizing factor ρ=0.4, ρ are set min =0.1, initial pheromone concentration Tua [ i, j]=0, pheromone fixed value q=50.
Further, in the step 5, the mean square error calculation formula is as follows:
in the above formula, MSE is mean square error, np is training set number or test set number, V is mechanical performance parameter number, Y e For the target output value, Y p Is a predicted output value; let MSE < ε 2 Indicating that the neural network training stop condition is satisfied epsilon 2 Is the maximum mean square error value that is allowed to be accepted.
According to the invention, an improved ant colony algorithm is combined with a BP neural network, an optimization model of a neural network weight and a threshold is established, an information heuristic factor is added into a probability selection function of the ant colony algorithm to prevent the algorithm from sinking into local optimum, meanwhile, a self-adaptive pheromone volatilization factor is added into pheromone updating, the relation between needling process parameters and the mechanical properties of the carbon fiber preform is more accurately described, and the optimized neural network output value is utilized to reversely acquire the accurate needling process parameters corresponding to the mechanical properties of the target carbon fiber preform.
Drawings
FIG. 1 is a flow chart of a carbon fiber preform parameter optimization method of the present invention;
FIG. 2 is a flow chart of an improved ant colony algorithm in the parameter optimization method of the present invention;
FIG. 3 is a flow chart of determining weights and thresholds for neural networks in the parameter optimization method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
The embodiment discloses a carbon fiber preform parameter optimization method, as shown in fig. 1, comprising the following steps:
step 1: constructing a sample database, wherein the database comprises needling process parameters and corresponding mechanical properties of the carbon fiber prefabricated body.
Specifically, a database composed of N groups of different needling process parameters and mechanical properties of the carbon fiber preform is established, wherein the needling process parameters comprise parameters such as needling depth, needling density, fiber layering mode, fiber thickness and the like; the mechanical properties include tensile strength, compressive strength, shear strength and other parameters. N groups of different needling parameters are obtained by a controlled variable method. And obtaining the mechanical properties of the needled carbon fiber preform manufactured by N groups of needling process parameters through experiments, thereby constructing a sample database.
Step 2: preprocessing a database, dividing samples in the database into a training set and a testing set, and coding needling process parameters in the database.
Specifically, 80% of the sample numbers in the N groups of databases are randomly selected as training sets, and the remaining 20% of the sample numbers are selected as test sets. Along with itAnd selecting M times by a machine to obtain M groups of training sets. The needling parameters in the database are encoded, the value range of the needling parameters is determined, the definition domain F is equally divided, the needling parameters in the database are equal to the interval median value with the smallest difference value, the 90-degree cross ply of the fibers in the fiber ply mode is specified to be assigned 1, and the parallel ply of the fibers is assigned 0. In this embodiment, the needling depth range is [2,10 ]]mm, needling density of [15,55 ]]Needle/cm 2 The thickness of the fiber layer is in the range of [5,30 ]]mm。
Step 3: and constructing a BP neural network, taking needling process parameters as BP neural network input layer nodes, taking the mechanical properties of the carbon fiber preform as BP neural network output layer nodes, and encoding network weights and thresholds.
Specifically, let the number of input layer nodes of the BP neural network be a, the number of intermediate layer nodes be b, and the number of output layer nodes be c. Let the number of weights be Nq, nq=a+b+b, let the number of thresholds be Ny, ny=2, let the number of parameters be n, n=nq+ny, and number the parameters i (i=1, 2, 3, … n) in order. And encoding the network weight and the threshold value, defining the value range of the network weight and the threshold value as [0,1], equally dividing the definition domain F, and selecting the median value in the interval as the selected value.
Step 4: optimizing the weight and the threshold of the BP neural network by using an improved ant colony algorithm, setting the maximum iteration times, obtaining the weight and the threshold of the network by using a probability selection function added with information heuristic factors, calculating the output vector of the BP neural network after iteration, and selecting the neural network parameters with minimum error of the output mechanical property vector and meeting the set error condition.
Specifically, as shown in fig. 2, after the neural network weight and the threshold are encoded, the ant colony parameters are initialized, the number of ant colonies m=50, the maximum iteration number of ant colonies t=100, the information heuristic factor α=3, the pheromone volatilization factor ρ=0.4, ρ is set min =0.1, initial pheromone concentration Tua [ i, j]=0, pheromone fixed value q=50.
The probability transfer function is calculated as follows in equation 1:
in the above formula, t is the iteration number,representing the probability of the kth ant for the jth chosen value in the set of i (i=1, 2, 3, … n) parameters; τ ij The concentration of the pheromone of the j-th selected value in the i parameter set is represented; alpha is an information heuristic factor reflecting the importance of the pheromone. And obtaining the network weight and the threshold value selected by the current ant according to the probability selection function, and calculating the output vector of the neural network.
Then the error of the output vector target value and the predicted value is calculated as follows in equation 2:
in the above, E k (t) is the error generated by the kth ant in the t iteration, V is the number of mechanical performance parameters, Y e For the target output value, Y p Is a predicted output value; set E k ＜ε 1 Indicating that the error setting condition is satisfied, ε 1 To a maximum error value that is acceptable.
Secondly, judging whether the iteration number reaches the set maximum iteration number, if t<T, the ant colony algorithm continues to iterate, then the probability transfer function is recalculated after updating the ant pheromone according to the following formula 3, if t=T, E is taken out k And (3) the neural network weight and the threshold value corresponding to the minimum value in (t).
The pheromone updating formula for improving the ant colony algorithm is as follows:
in the above formula, Q is a fixed value of pheromone, and ρ is a volatile factor of pheromone.
The adaptive formula of the pheromone volatilization factor is as follows:
step 5: randomly selecting M times in a database to obtain M groups of training sets, and training the BP neural network by adopting the M groups of training sets; when the set neural network training stopping condition is met, verifying the BP neural network by adopting a test set corresponding to the training set, and selecting a group of neural network parameters with the minimum mean square error of the final output vector to obtain a selected model of the needling parameters of the carbon fiber preform.
Specifically, as shown in fig. 3, the BP neural network is trained using M sets of training sets, and the training set mean square error is calculated according to the following formula 5:
in the above formula, MSE is mean square error, np is training set number or test set number, V is mechanical performance parameter number, Y e For the target output value, Y p To predict the output value.
Then, judging whether the calculated mean square error is smaller than the set value, setting MSE < epsilon 2 Indicating that the neural network training stop condition is satisfied epsilon 2 Is the maximum mean square error value that is allowed to be accepted. And after the training stopping condition of the neural network is met, obtaining the trained BP neural network, and retraining if the training stopping condition is not met. After the neural network is trained, the BP neural network is verified by adopting a test set corresponding to the training set. When MSE < ε 2 And outputting a network weight and a threshold value, continuously selecting another training set to train the network, finally obtaining M needling parameter selection models, and selecting one BP neural network with the minimum test mean square error as a carbon fiber preform needling parameter selection model.
Step 6: inputting the mechanical properties of the target carbon fiber preform by using the selection model of the needling parameters of the carbon fiber preform obtained in the step 5, and reversely obtaining needling process parameters corresponding to the mechanical properties of the target carbon fiber preform by using the BP neural network.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (9)
1. A method for optimizing parameters of a carbon fiber preform, comprising:
step 1: constructing a sample database, wherein the database comprises needling process parameters and corresponding mechanical properties of the carbon fiber preform;
step 2: preprocessing a database, dividing a sample in the database into a training set and a testing set, and encoding needling process parameters in the database;
step 3: constructing a BP neural network, taking needling process parameters as BP neural network input layer nodes, taking the mechanical properties of the carbon fiber preform as BP neural network output layer nodes, and coding network weights and threshold values;
step 4: optimizing the weight and the threshold of the BP neural network by using an improved ant colony algorithm, setting the maximum iteration times, obtaining the weight and the threshold of the network by using a probability selection function added with information heuristic factors, calculating the output vector of the BP neural network after iteration, and selecting the neural network parameter with the minimum error of the output mechanical property vector and meeting the set error condition;
wherein the probability selection function is as follows:
in the above formula, t is the iteration number,representing the probability of the kth ant for the jth chosen value in the set of i (i=1, 2, 3, … n) parameters; τ ij Representing i parameterThe concentration of pheromone of the j-th selected value in the number set; alpha is an information heuristic factor, reflecting the importance of the pheromone;
the error calculation formula is as follows:
in the above, E k (t) is the error generated by the kth ant in the t iteration, V is the number of mechanical performance parameters, Y e For the target output value, Y p Is a predicted output value; set E k ＜ε 1 Indicating that the error setting condition is satisfied, ε 1 Is the maximum error value allowed to be accepted;
the pheromone updating formula for improving the ant colony algorithm is as follows:
in the formula, Q is a fixed value of the pheromone, and rho is a volatile factor of the pheromone;
the adaptive formula of the pheromone volatilization factor is as follows:
step 5: randomly selecting M times in a database to obtain M groups of training sets, and training the BP neural network by adopting the M groups of training sets; when the set neural network training stopping condition is met, verifying the BP neural network by adopting a test set corresponding to the training set, and selecting a group of neural network parameters with the minimum mean square error of the final output vector to obtain a selected model of the needling parameters of the carbon fiber preform;
step 6: inputting the mechanical properties of the target carbon fiber preform by using the selection model of the needling parameters of the carbon fiber preform obtained in the step 5, and reversely obtaining needling process parameters corresponding to the mechanical properties of the target carbon fiber preform by using the BP neural network.
2. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 1, needling process parameters include needling depth, needling density, fiber layering mode and fiber thickness; the mechanical properties include tensile strength, compressive strength, and shear strength.
3. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 2, 80% of the samples in the database are used as training sets, and the remaining 20% of the samples are used as test sets.
4. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 2, the needling process parameters are encoded as follows: determining the value range of needling process parameters, equally dividing the definition domain F of the value range, enabling the needling process parameters in a database to be equal to the interval median value with the smallest difference value, and prescribing that the 90-degree cross ply of the fibers in the fiber ply mode is assigned to be 1 and the parallel ply of the fibers is assigned to be 0.
5. The method for optimizing parameters of a carbon fiber preform according to claim 4, wherein: the needling depth range in the needling process parameters is set to be [2,10]mm, needling density of [15,55 ]]Needle/cm 2 The thickness of the fiber layer is in the range of [5,30 ]]mm。
6. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 3, the network weight and the threshold are encoded as follows: determining the value ranges of the weight and the threshold, equally dividing the definition domain F of the weight and the threshold, and selecting the median value in the interval as a selected value; setting the number of nodes of an input layer of the BP neural network as a, the number of nodes of an intermediate layer as b, and the number of nodes of an output layer as c; let the number of weights be Nq, nq=a+b+b, let the number of thresholds be Ny, ny=2, and let the number of parameters be n, n=nq+ny.
7. The method for optimizing parameters of a carbon fiber preform according to claim 6, wherein: the weight and threshold values are set to be in the range of 0, 1.
8. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 4, the number of ant colony m=50, the maximum iteration number of ant colony t=100, the information heuristic factor α=3, the pheromone volatilizing factor ρ=0.4, ρ are set min =0.1, initial pheromone concentration Tua [ i, j]=0, pheromone fixed value q=50.
9. The method for optimizing parameters of a carbon fiber preform according to claim 1, wherein: in the step 5, the mean square error calculation formula is as follows:
in the above formula, MSE is mean square error, np is training set number or test set number, V is mechanical performance parameter number, Y e For the target output value, Y p Is a predicted output value; let MSE < ε 2 Indicating that the neural network training stop condition is satisfied epsilon 2 Is the maximum mean square error value that is allowed to be accepted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310309502.1A CN116467934A (en) | 2023-03-28 | 2023-03-28 | Carbon fiber preform parameter optimization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310309502.1A CN116467934A (en) | 2023-03-28 | 2023-03-28 | Carbon fiber preform parameter optimization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116467934A true CN116467934A (en) | 2023-07-21 |
Family
ID=87178049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310309502.1A Pending CN116467934A (en) | 2023-03-28 | 2023-03-28 | Carbon fiber preform parameter optimization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116467934A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117672435A (en) * | 2024-01-31 | 2024-03-08 | 广元水木新材料科技有限公司 | Automatic fiber yarn layout method and system based on nanofiber preparation |
-
2023
- 2023-03-28 CN CN202310309502.1A patent/CN116467934A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117672435A (en) * | 2024-01-31 | 2024-03-08 | 广元水木新材料科技有限公司 | Automatic fiber yarn layout method and system based on nanofiber preparation |
| CN117672435B (en) * | 2024-01-31 | 2024-04-09 | 广元水木新材料科技有限公司 | Automatic fiber yarn layout method and system based on nanofiber preparation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111985796B (en) | Method for predicting concrete structure durability based on random forest and intelligent algorithm | |
| CN110544011B (en) | Intelligent system combat effectiveness evaluation and optimization method | |
| CN110363344A (en) | Probability integral parameter prediction method based on MIV-GP algorithm optimization BP neural network | |
| CN109919356B (en) | BP neural network-based interval water demand prediction method | |
| CN108388702A (en) | Engineering ceramics electrical discharge machining effect prediction method based on PSO neural networks | |
| CN116467934A (en) | Carbon fiber preform parameter optimization method | |
| CN108460461A (en) | Mars earth shear parameters prediction technique based on GA-BP neural networks | |
| CN112016244B (en) | Durable concrete multi-target mix proportion optimization method based on SVM and intelligent algorithm | |
| CN112765902B (en) | Soft measurement modeling method for COD concentration in rural domestic sewage treatment process based on TentFWA-GD RBF neural network | |
| CN114547974A (en) | Dynamic soft measurement modeling method based on input variable selection and LSTM neural network | |
| CN116542382A (en) | Sewage treatment dissolved oxygen concentration prediction method based on mixed optimization algorithm | |
| CN109241649B (en) | Fiber yarn performance detection method and system based on decision tree model | |
| CN107274016A (en) | The strip exit thickness Forecasting Methodology of the random symmetrical extreme learning machine of algorithm optimization that leapfrogs | |
| CN115115389A (en) | Express customer loss prediction method based on value subdivision and integrated prediction | |
| CN104732067A (en) | Industrial process modeling forecasting method oriented at flow object | |
| CN109940458A (en) | A kind of cutter future wear amount on-line prediction method | |
| CN113971517A (en) | GA-LM-BP neural network-based water quality evaluation method | |
| CN115481727A (en) | Intention recognition neural network generation and optimization method based on evolutionary computation | |
| CN115221793A (en) | Tunnel surrounding rock deformation prediction method and device | |
| CN110110447B (en) | Method for predicting thickness of strip steel of mixed frog leaping feedback extreme learning machine | |
| CN112800673A (en) | Method for predicting blasting block degree based on SA-GA-BP algorithm | |
| CN115394381B (en) | High-entropy alloy hardness prediction method and device based on machine learning and two-step data expansion | |
| CN113537539A (en) | Multi-time-step heat and gas consumption prediction model based on attention mechanism | |
| CN116822742A (en) | Power load prediction method based on dynamic decomposition-reconstruction integrated processing | |
| CN115577626A (en) | Ground settlement prediction method and system based on soil property parameter quantification |
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 |