CN118114545A - Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving - Google Patents
Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving Download PDFInfo
- Publication number
- CN118114545A CN118114545A CN202311547904.1A CN202311547904A CN118114545A CN 118114545 A CN118114545 A CN 118114545A CN 202311547904 A CN202311547904 A CN 202311547904A CN 118114545 A CN118114545 A CN 118114545A
- Authority
- CN
- China
- Prior art keywords
- open
- pore
- maximum stress
- reinforcing plate
- composite material
- 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
- 239000011148 porous material Substances 0.000 title claims abstract description 50
- 239000002131 composite material Substances 0.000 title claims abstract description 38
- 238000000034 method Methods 0.000 title claims abstract description 27
- 230000003014 reinforcing effect Effects 0.000 claims description 31
- 238000013528 artificial neural network Methods 0.000 claims description 22
- 238000004458 analytical method Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 10
- 230000002787 reinforcement Effects 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000003062 neural network model Methods 0.000 claims description 6
- 238000004088 simulation Methods 0.000 claims description 6
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 2
- 238000012546 transfer Methods 0.000 claims description 2
- 238000013507 mapping Methods 0.000 abstract description 2
- 230000006378 damage Effects 0.000 description 9
- 238000005336 cracking Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012805 post-processing Methods 0.000 description 3
- 239000003351 stiffener Substances 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000013515 script Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000004753 textile Substances 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Landscapes
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
Abstract
The invention discloses a method for predicting maximum stress of a composite material open-pore stiffened plate based on data driving, which is used for modeling a stress process of the composite material open-pore stiffened plate based on a DNN (digital network), and for open-pore stiffened plates with different size details, using python parameter modeling to establish a composite material open-pore stiffened plate model, obtaining maximum equivalent stress at an open pore of the composite material open-pore stiffened plate after boundary conditions and loads are applied, so as to establish a high-confidence database, and finally establishing a high-cost-performance high-confidence mapping based on the DNN to obtain a maximum stress value at an open pore stiffened plate to be tested.
Description
Technical Field
The invention relates to a data-driven maximum stress prediction method for a composite material perforated reinforcing plate, and belongs to the technical field of composite material calculation.
Background
The composite material has the characteristics of high specific strength, high specific modulus, fatigue resistance, high temperature resistance, corrosion resistance and the like, and has important application in the key fields of aerospace, chemical industry, vehicles, textile, mechanical manufacturing and the like. In an extreme service environment, the composite material can also be damaged, and the damage forms mainly comprise matrix cracking, fiber cracking and interface cracking. The internal damage seriously degrades the mechanical property of the composite material structure, thereby reducing the strength and the service life of the structure and bringing potential safety hazards. From the perspective of composite material equipment design and reliability high-efficiency evaluation, the rapid prediction technology of the damage behavior of the composite material has high application value.
At present, the damage prediction of the open-pore stiffening plate of the composite material is mainly based on the analysis of equivalent stress. For example, modeling a stress process of a composite material plate, building a composite material plate model for composite material plates with different layering and size details, obtaining data such as strength, damage, fatigue life and the like of the composite material plate after applying boundary conditions and loads, building a high-confidence database, and finally building a mapping relation. The method is oriented to the whole composite material plate, cannot be directly applied to the opening of the reinforcing plate, and the accuracy of the obtained stress value is not high.
Disclosure of Invention
The invention solves the technical problems that: the method for predicting the maximum stress of the open-pore stiffening plate of the composite material based on data driving is provided, the maximum stress of the open-pore stiffening plate of the composite material is directly predicted based on the radius of the open-pore rectangular chamfer, the side length of the open-pore rectangular and the like, and a new way is provided for rapid evaluation and structural design of the performance of the composite material.
The technical scheme of the invention is as follows:
A maximum stress prediction method for a composite material open pore reinforcing plate based on data driving comprises the following steps:
according to the size, the position, the layering angle and the load information of the perforated reinforcement plate, a perforated reinforcement plate model is built based on Abaqus, wherein the side length, the chamfer radius and the rib height of a perforated rectangle are used as variable parameters, and a plurality of perforated reinforcement plate models with different values of the variable parameters are selected;
Based on each open pore reinforcing plate model, carrying out maximum stress simulation of the open pore reinforcing plate based on an Abaqus system to obtain a maximum stress value at an opening; matching the side lengths of all the open rectangles of the model, the chamfer radius of the open rectangles, the heights of the ribs and the calculated corresponding maximum stress values to form a sample data set;
Taking the side length of the open rectangle, the chamfer radius of the open rectangle and the height of the ribs as input, taking the maximum stress value at the opening as output, training the DNN neural network by using a sample data set, and establishing a DNN neural network model for predicting the maximum stress at the opening;
And acquiring the side length of the hole rectangle of the hole stiffening plate to be measured, the chamfer radius of the hole rectangle and the rib height, inputting the side length, the chamfer radius and the rib height of the hole rectangle to be measured into the DNN neural network model, and outputting the maximum stress value of the opening of the hole stiffening plate to be measured.
Preferably, an open pore stiffened plate model is established based on Abaqus, finite element grids of the stiffened plate with holes and the ribs are divided, a coordinate origin is selected at the center point position of the stiffened plate, an x-axis is an external load direction, a y-axis is a direction orthogonal to the x-axis in a laminated plate surface, and a z-axis is a thickness direction of layering and superposition of the stiffened plate; the split reinforcing plates are scattered by adopting the traditional shell units S4R, the ribs and the split plates are connected in a binding mode, full constraint is applied to the model, and the side distribution load of the shell is applied.
Preferably, a parameterized model is established based on Python language, after the geometry, the attribute and the grid parameters of the model are defined, the preprocessing, the analysis and the calculation and the post-processing processes of the open-pore stiffened plate model are realized in a batched and intelligent mode through the Python language, and then different open-pore stiffened plate models are generated in a rapid batch mode through defining the value range of the variable parameter.
Preferably, a plurality of open-pore stiffened plate models with different values of the variable parameters are selected, and the number of the open-pore stiffened plate models is at least 100.
Preferably, the sample data set is divided into a training set, a verification set and a prediction set according to the ratio of 70:15:15, the DNN neural network is trained by the training set, the verification set performs validity check on the trained DNN neural network, and the prediction set performs prediction verification on the DNN neural network model which passes verification.
Preferably, the DNN neural network is provided with an input layer, an output layer and a hidden layer, wherein the input data of the input layer is the side length of an open rectangle, the chamfer radius of the open rectangle and the height of ribs, and the output data of the output layer is the maximum stress value at an opening; the number of layers of the hidden layer, the number of nodes and the transfer function in the hidden layer are set according to the convergence degree of the sample data.
Preferably, the DNN neural network is optimized by adopting an Adam algorithm.
Compared with the prior art, the invention has the advantages that:
(1) According to the method, the open-pore reinforcement is subjected to simulation analysis by parametric modeling and automatic post-processing technology, so that a stress analysis result database of the composite open-pore reinforcement plate is constructed, and the rapid and intelligent prediction of the stress analysis of the composite open-pore reinforcement structure is realized.
(2) The invention can avoid the problems of cost and operation caused by repeated modeling, realizes batch construction, calculation and processing of simulation models and results, and provides a new approach for the design and application of composite material structures.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:
FIG. 1 is a schematic illustration of a geometry of a composite open cell stiffener according to an embodiment of the present invention;
FIG. 2 is a diagram of finite element model information relating to open cell stiffener plates according to an embodiment of the present invention;
FIG. 3 is a topology diagram of a DNN neural network according to an embodiment of the present invention;
FIG. 4 is a graph of maximum stress prediction error for a stiffener plate hole attachment in accordance with an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The invention provides a data-driven method for predicting maximum stress of a composite material perforated reinforcing plate, which is described in detail through an embodiment. The embodiment is an open-pore reinforcing plate, the geometric configuration is shown in fig. 1, the length L of the laminated plate is 465mm, the width W of the laminated plate is 350mm, the open-pore shape is a chamfer rectangle, and the binding of the ribs and the open-pore composite material plate is selected as a research model. The method specifically comprises the following steps:
S1) constructing rapid parametric modeling method combining characteristics of open-pore laminated plates
A numerical model is built on a laminated plate and a reinforcing sheet containing elliptical holes by adopting a parameterized modeling method, a Python modeling script is compiled according to a data file format specified by Abaqus, the method comprises the steps of building a geometric model of an open-pore reinforcing plate, endowing material properties on the basis of the geometric model, dividing a structured grid, and applying corresponding boundary constraint conditions and load working conditions.
The parameterized modeling process is realized based on Python language, the related parameters of geometric dimension, hole opening position, layering angle, load information and grid dimension are set as variables, after the geometric dimension, attribute and grid parameter of the model are defined, the preprocessing, analysis and calculation and post-processing processes of the finite element model are realized in a batched and intelligent mode through Python language, and finally a numerical model database containing 100 groups of different models is rapidly generated in batch through defining the value range of the variables.
In the numerical model database, the finite element meshing strategy of the reinforcing plates and the ribs with holes is shown in fig. 2. The origin of coordinates is selected at the center point of the stiffening plate; the x-axis is the external load direction, and the y-axis is the direction orthogonal to the x-axis in the laminated board surface; the z-axis is the thickness direction of the laminated layers of the reinforcing plates. The composite laminate is discretized using conventional shell elements S4R. The ribs are connected with the composite material in a binding mode, full constraint is applied to the model, and the side distribution load of the shell is applied. Fig. 2 (a) shows the geometry and meshing of the model skin. Fig. 2 (c) shows the meshing of the stiffened plate finite element model. Fig. 2 (b) shows 8 plies laid in the thickness direction, each ply having a thickness of 0.2mm.
S2) constructing a damage analysis result database
After parameterized modeling is achieved, 100 groups of numerical models constructed in the step S1 are submitted to an Abaqus solver for operation by combining Hashin failure criteria. Classifying and grading simulation results obtained by Abaqus calculation, compiling batch Python scripts according to requirements, and extracting corresponding data from a numerical model database and the simulation results to construct a 100-group damage analysis result database.
S3) constructing and training DNN neural network
Based on the database containing 100 groups of damage analysis results constructed in the step S2, a DNN neural network is constructed by Matlab to train the maximum stress analysis result database, and the topological structure of the DNN neural network is composed of 3 parts, namely an input layer, an output layer and an implicit layer respectively as shown in figure 3. The input layer has 3 parameters, namely the side length L of the open rectangle, the chamfer radius R of the open rectangle and the rib height h; the output layer has 1 parameter, which is the maximum stress at the opening that needs to be predicted. And adjusting the convergence degree of the data to achieve the minimum prediction error. The neural network optimization algorithm adopts an Adam algorithm, which is a gradient descent algorithm for optimizing the deep neural network, combines the advantages of momentum and self-adaptive learning rate, and has the advantages of self-adaptive adjustment of learning rate on different parameters, high convergence speed, good adaptability to sparse gradients, deviation correction during parameter updating and the like.
In the training process of the neural network, data in a result database is analyzed for 100 groups of established injuries, wherein the first 70 groups are used as training sets of the neural network, 15 groups of data are used as verification sets, and 15 groups of data are used as prediction sets. Wherein the validation set is used for the trained neural network to avoid data overfitting and to verify the validity of the algorithm. The maximum stress of the 15 predicted sets was predicted using a DNN neural network trained with 70 sets of data. Making the predicted relative error of the dataset into a data diagram shown in fig. 4, wherein fig. 4 (a) is the relation between root mean square error and iteration number; fig. 4 (b) is an error analysis of the data set prediction and fit curves. The root mean square error for the prediction was 0.321. From the prediction result, the data driving technology adopted by the invention can accurately predict the maximum stress of the open-pore reinforcing plate.
S4) maximum stress prediction analysis based on data driving
After training of the DNN neural network is completed through the constructed result database, the side length L of the open rectangle of the open-pore stiffening plate to be predicted, the chamfer radius R of the open rectangle and the rib height h are obtained and input into the DNN neural network, so that the stress prediction condition of the open-pore stiffening plate, such as the maximum stress of the open pore of the stiffening plate, is obtained, and the purpose of predicting and analyzing the stress behavior based on data driving is achieved.
The above examples are only preferred embodiments of the present invention, and ordinary changes and substitutions made by those skilled in the art within the scope of the present invention are intended to be included in the scope of the present invention.
Claims (7)
1. The method for predicting the maximum stress of the open pore reinforcing plate of the composite material based on data driving is characterized by comprising the following steps of:
according to the size, the position, the layering angle and the load information of the perforated reinforcement plate, a perforated reinforcement plate model is built based on Abaqus, wherein the side length, the chamfer radius and the rib height of a perforated rectangle are used as variable parameters, and a plurality of perforated reinforcement plate models with different values of the variable parameters are selected;
Based on each open pore reinforcing plate model, carrying out maximum stress simulation of the open pore reinforcing plate based on an Abaqus system to obtain a maximum stress value at an opening; matching the side lengths of all the open rectangles of the model, the chamfer radius of the open rectangles, the heights of the ribs and the calculated corresponding maximum stress values to form a sample data set;
Taking the side length of the open rectangle, the chamfer radius of the open rectangle and the height of the ribs as input, taking the maximum stress value at the opening as output, training the DNN neural network by using a sample data set, and establishing a DNN neural network model for predicting the maximum stress at the opening;
And acquiring the side length of the hole rectangle of the hole stiffening plate to be measured, the chamfer radius of the hole rectangle and the rib height, inputting the side length, the chamfer radius and the rib height of the hole rectangle to be measured into the DNN neural network model, and outputting the maximum stress value of the opening of the hole stiffening plate to be measured.
2. The method for predicting the maximum stress of the open-pore reinforcing plate of the composite material according to claim 1, wherein an open-pore reinforcing plate model is established based on Abaqus, finite element grids of the reinforcing plate with holes and ribs are divided, a coordinate origin is selected at the center point position of the reinforcing plate, an x-axis is an external load direction, a y-axis is a direction orthogonal to the x-axis in a laminated plate surface, and a z-axis is the thickness direction of laminated layers of the reinforcing plate; the split reinforcing plates are scattered by adopting the traditional shell units S4R, the ribs and the split plates are connected in a binding mode, full constraint is applied to the model, and the side distribution load of the shell is applied.
3. The method for predicting the maximum stress of the open-pore reinforcing plate of the composite material according to claim 1 is characterized in that a parameterized model is established based on Python language, after the geometry, the attribute and the grid parameters of the model are defined, the pretreatment, the analysis and the calculation and the post-treatment processes of the open-pore reinforcing plate model are realized in a batched and intelligent mode through Python language, and then different open-pore reinforcing plate models are generated in a rapid mode through defining the value range of variable parameters.
4. The method for predicting the maximum stress of the open-pore reinforcing plate of the composite material according to claim 1, wherein a plurality of open-pore reinforcing plate models with different values of the variable parameters are selected, and the number of the open-pore reinforcing plate models is at least 100.
5. The method for predicting the maximum stress of the open-pore reinforcing plate of the composite material according to claim 1, wherein the sample data set is divided into a training set, a verification set and a prediction set according to the proportion of 70:15:15, the DNN neural network is trained by the training set, the verification set performs validity check on the trained DNN neural network, and the prediction set performs prediction verification on the DNN neural network model passing verification.
6. The method for predicting the maximum stress of the open-pore reinforcing plate of the composite material according to claim 1, wherein the DNN neural network comprises an input layer, an output layer and a hidden layer, the input data of the input layer is the side length of the open-pore rectangle, the chamfer radius of the open-pore rectangle and the height of the ribs, and the output data of the output layer is the maximum stress value at the opening; the number of layers of the hidden layer, the number of nodes and the transfer function in the hidden layer are set according to the convergence degree of the sample data.
7. The method for predicting the maximum stress of the open pore reinforcing plate of the composite material according to claim 1 or 6, wherein the DNN neural network is optimized by adopting an Adam algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311547904.1A CN118114545A (en) | 2023-11-20 | 2023-11-20 | Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311547904.1A CN118114545A (en) | 2023-11-20 | 2023-11-20 | Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118114545A true CN118114545A (en) | 2024-05-31 |
Family
ID=91219941
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311547904.1A Pending CN118114545A (en) | 2023-11-20 | 2023-11-20 | Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118114545A (en) |
-
2023
- 2023-11-20 CN CN202311547904.1A patent/CN118114545A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kharmanda et al. | Reliability-based topology optimization | |
| CN107451307B (en) | Method for multi-scale calculation of equivalent stiffness matrix of complex composite structure | |
| CN102841978B (en) | Design of curved fiber paths for composite laminates | |
| US8103488B2 (en) | Tire design method | |
| CN107451309B (en) | Method for multi-scale calculation of equivalent thermal expansion coefficient of complex composite material structure | |
| JP2022521907A (en) | High-speed coordinated optimization method for plate-wound shell structure of hybrid fiber composite material | |
| CN111125946B (en) | Method for optimizing structure of boarding body based on MDO technology | |
| CN105335582A (en) | Modeling method for airplane composite material wall plate weight analysis | |
| CN106202731A (en) | Bridge crane multi-flexibl e dynamics structural optimization method | |
| CN108491576B (en) | Optimization design method for reinforcing composite material wing opening | |
| CN110083916A (en) | Based on the mechanical structure Fuzzy fatigue reliability optimization method from structure membership function | |
| Eckrich et al. | Structural topology optimization and path planning for composites manufactured by fiber placement technologies | |
| CN116451375A (en) | Parameterized modeling and optimization design method for box girder of portal crane | |
| CN110399675A (en) | A kind of elevator door multi-objective optimization design of power method based on genetic algorithm | |
| CN109241649A (en) | A kind of fiber filament method for testing performance and system based on decision-tree model | |
| CN101612996A (en) | A kind of plate muscle construction design method | |
| CN115495965A (en) | Method for analyzing time-varying reliability of complex aviation structure under mixed uncertainty | |
| CN113688465B (en) | Aircraft structural strength digital twin method based on combination of load and state | |
| CN118114545A (en) | Maximum stress prediction method for open-pore stiffening plate of composite material based on data driving | |
| Nguyen et al. | Investigation and multi-scale optimization design of woven composite cut-out structures | |
| CN115391947A (en) | Virtual analysis method and system for composite material structure of rail transit vehicle | |
| CN115544746A (en) | Multi-attribute target-driven aluminum auxiliary frame optimization design method and system | |
| CN115547431A (en) | Short-cut fiber composite material fatigue performance prediction method based on machine learning and progressive damage | |
| CN114329805A (en) | Connecting piece multidisciplinary collaborative design optimization method based on self-adaptive agent model | |
| CN114254533A (en) | Method for examining influence and prediction of fatigue vibration on fixed angle of product group component |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |