CN114117871A - Beam unit structure optimization method based on PSO-BP neural network - Google Patents

Beam unit structure optimization method based on PSO-BP neural network Download PDF

Info

Publication number
CN114117871A
CN114117871A CN202111598711.XA CN202111598711A CN114117871A CN 114117871 A CN114117871 A CN 114117871A CN 202111598711 A CN202111598711 A CN 202111598711A CN 114117871 A CN114117871 A CN 114117871A
Authority
CN
China
Prior art keywords
neural network
pso
unit structure
beam unit
parameters
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
Application number
CN202111598711.XA
Other languages
Chinese (zh)
Inventor
高阳
刘孝保
孙海彬
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.)
Kunming University of Science and Technology
Original Assignee
Kunming University of Science and Technology
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 Kunming University of Science and Technology filed Critical Kunming University of Science and Technology
Priority to CN202111598711.XA priority Critical patent/CN114117871A/en
Publication of CN114117871A publication Critical patent/CN114117871A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
    • 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
    • G06F30/25—Design optimisation, verification or simulation using particle-based methods
    • 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
    • G06N3/045—Combinations of networks
    • 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
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Computer Hardware Design (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method for optimizing a beam unit structure of a PSO-BP neural network, which comprises the following steps: the method comprises the steps of firstly determining the size limit of a beam unit structure under the condition of required rigidity strength by using ABAQUS, extracting a part of initial particles as PSO (particle swarm optimization) from a combination according to required optimization parameters, and optimizing initial weight and threshold values of a BP (back propagation) neural network by using the PSO. And carrying out parametric modeling on the beam unit structure, carrying out programming of secondary development through Anaconda Jupiter, generating an inp file in batches, submitting the file to a solver in batches through a batch file format bat, and finally extracting required data in a result as initial input data of the BP neural network. The beam unit structure can be quickly and efficiently structurally optimized by the combined use of ABAQUS secondary development and optimization algorithm, and meanwhile, the convergence rate of a BP neural network can be improved by the improvement of the algorithm, so that the optimization efficiency of the beam unit structure is improved.

Description

Beam unit structure optimization method based on PSO-BP neural network
Technical Field
The invention relates to the technical field of industrial software design, in particular to a PSO-BP neural network-based beam unit structure optimization method.
Background
With the continuous development of modern science and technology, industrial software and optimization design methods are continuously increased, wherein the optimization design is based on the optimization theory of mathematics and achieves the purpose of optimization through computer industrial software and secondary development programming.
In the prior art, the ABAQUS and the BP neural network are subjected to combined optimization simulation. Firstly, carrying out initial structure design on a beam unit structure, taking part of design parameter combination as an initialization particle of PSO, and optimizing an initial weight and a threshold of a BP neural network. And carrying out parametric modeling on the beam unit structure through an ABAQUS software macro recording function. Performing python secondary development programming in Jupyter, quickly modifying the x. inp file of the required optimization parameters in batches, and extracting the required simulation data after batch submission to a CAE solver. After determining the structures of all layers of the BP neural network, the data can be imported into the neural network for training and prediction. And the time required for optimization by using the PSO-BP is short and the efficiency is high.
Disclosure of Invention
The invention aims to provide a PSO-BP-based beam unit structure optimization method to solve the problems of low optimization efficiency and high repeatability.
In order to achieve the purpose, the invention provides a beam unit structure optimization method based on a PSO-BP neural network.
And determining PSO initial particle swarm according to the beam unit structure initial structure parameters. And optimizing the initial weight and the threshold of the BP neural network by adopting a PSO algorithm. And constructing a prediction model of the PSO-BP neural network according to the obtained initial optimal weight and the threshold. And carrying out parametric modeling on the beam unit structure to be optimized, and establishing a corresponding beam unit structure finite element model after inputting design parameters in a TXT text format.
And carrying out secondary development according to the files obtained by the ABAQUS macro recording function, and combining a corresponding GUI interface to realize the functions of batch generation of the X-inp files required by the ABAQUS and data extraction. After determining parameters of each layer of the BP neural network, inputting the obtained data into the BP neural network for training so as to improve the accuracy of the BP prediction model. And meanwhile, secondary development is carried out in the SW, and the functions of quickly generating three views of the beam unit structure and engineering drawings are realized.
The invention also provides a computer terminal device comprising one or more processors and a memory. A memory coupled to the processor for storing one or more programs; when executed by the one or more processors, cause the one or more processors to implement a PSO-BP based beam cell structure optimization method as in any of the embodiments above.
The present invention further provides a computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the PSO-BP based beam unit structure optimization method according to any of the embodiments.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a PSO-BP-based beam unit structure optimization method according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a PSO algorithm according to another embodiment of the present invention
FIG. 3 is a flow chart showing a PSO-BP neural network optimization beam unit structure according to an embodiment of the present invention
Detailed description of the preferred embodiments
The technical solutions in the present invention will be described clearly and completely with reference to the accompanying drawings, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be understood that the step numbers used herein are for convenience of description only and are not used as limitations on the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The terms "comprises" and "comprising" indicate the presence of the described features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The term "and/or" refers to and includes any and all possible combinations of one or more of the associated listed items.
Referring to fig. 1, the present invention provides a method for optimizing a beam unit structure based on a PSO-BP neural network, including:
s101, determining initial structure parameters, and extracting a part of parameter combinations to serve as initialization particles of a PSO algorithm, so as to achieve the purpose of optimizing optimal BP initial weight and threshold.
Please refer to fig. 2, a joint optimization method based on the PSO-BP neural network and ABAQUS. Firstly, parameterization design is carried out on a beam unit structure, and programming is carried out by using a jupyter programming function of anaconda, so that batch generation of a pre-processing file and batch extraction of post-processing finite element analysis data are realized. Meanwhile, the aim of batch submission of the preprocessed files can be achieved through the bat file format. The beam unit structure can be rapidly modeled through parametric design, a pre-processing file is generated and submitted to a CAE solver for analysis, the operation can be rapidly realized, the time of repeated operation is saved, and the efficiency is improved.
S102, beam unit structure parametric modeling, ABAQUS secondary development, preprocessing batch generation and simulation data extraction. The secondary development can quickly realize the batch generation of the inp file, the structural parameters of the beam unit and the modification of the preprocessing parameters are modified in the file, and a large amount of time wasted on repeated operation can be saved after the secondary development. The post-processing process is developed for the second time, particularly for the part of extracted data, a large amount of data can be generated after the ABAQUS solver is used for solving, only a part of data is needed, the most data, such as stress, strain, overall displacement, plastic deformation and the like, of the maximum joint are obtained, specifically, the needed data can be obtained and used as a training set of a BP neural network for training, a simulated cloud picture can be obtained, and the result of simulation analysis can be checked more intuitively.
The macro recording function of the ABAQUS can record each step of operation in the form of codes, complete codes can be obtained through a complete finite element analysis process, and codes are subjected to parameter modification and partial deletion to obtain codes for secondary development.
After the initial structure parameters of the beam unit structure are determined, three-dimensional modeling is carried out in SW, the parameters can be associated through the equation function of the SW, the model can be quickly updated by modifying the parameters after the model is built, and the three views and the engineering drawing of the updated beam unit structure can be obtained while the time is saved.
S103, training the extracted data as bp neural network training data and acquiring trained parameters
Firstly, the number of input layers, hidden layers and output layers of the neural network is determined according to design parameters, and the relation between neurons of the hidden layers and neurons of the input layers in the network is n2＝2×n1+1。
Meanwhile, the weight and the threshold value after PSO optimization are used for training in the BP neural network, so that the convergence speed of the neural network can be greatly increased. Training can also be directly carried out to check the training effect of the BP neural network which is not optimized and compare with the training effect. After the predicted parameters are obtained, the design parameters are updated in SW.
And S104, carrying out secondary development in SW, connecting related parameters by using self-contained equation functions, updating the model after updating design parameters, and generating a three-view and an engineering drawing of the new model for later use.
Also provided is a computer terminal device including one or more processors and memory. A memory is coupled to the processor for storing one or more programs, which when executed by the one or more processors, cause the one or more processors to implement a PSO-BP based beam cell structure optimization method as in any one of the above embodiments.
The processor is used for controlling the overall operation of the computer terminal equipment so as to complete all or part of the steps of the ABAQUS-based beam unit structure optimization method. The memory is used to store various types of data to support the operation at the computer terminal device, which data may include, for example, instructions for any application or method operating on the computer terminal device, as well as application-related data. The Memory may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk or optical disk.
In an exemplary embodiment, the computer terminal Device may be implemented by one or more Application Specific 1 integrated circuits (AS 1C), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components, and is configured to perform the PSO-BP based beam unit structure optimization method and achieve technical effects consistent with the above methods.
In another exemplary embodiment, there is also provided a computer readable storage medium comprising program instructions which, when executed by a processor, implement the steps of the PSO-BP based beam cell structure optimization method in any of the above embodiments. For example, the computer readable storage medium may be the above-mentioned memory including program instructions executable by a processor of a computer terminal device to perform the above-mentioned PSO-BP based beam cell structure optimization method, and achieve the technical effects consistent with the above-mentioned method.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.

Claims (8)

1. A PSO-BP-based beam unit structure optimization method is characterized by comprising the following steps:
and combining according to preset parameters, and optimizing part of the parameters for the PSO. Obtaining an optimal initial weight and a threshold;
carrying out parametric design on a beam unit structure according to a BP (back propagation) neural network, and carrying out secondary development based on ABAQUS (asynchronous JavaScript and XML) to generate CAE solution files in batches and extract simulation data;
processing the extracted data according to the optimal weight and the threshold value to be used as training data of the BP neural network, and acquiring optimized design parameters;
inputting the trained parameters into SolidWorks for parameter updating, and acquiring updated three-view and engineering drawings.
2. The PSO-BP neural network-based beam unit structure optimization method of claim 1. The method is characterized in that combination is carried out according to preset parameters, and part of the parameters are combined to a PSO for optimization. And obtaining the optimal weight and the threshold. The method comprises the following steps:
and according to the PSO algorithm, taking part of the design parameter combination as an initialization particle of the PSO algorithm, and generating an initial optimal weight and a threshold of the BP neural network after the PSO algorithm is operated.
3. The PSO-BP neural network-based beam unit structure optimization method of claim 2, wherein the secondary development of ABAQUS after the design parameters comprises:
and recording the generation of corresponding codes of the operation steps in the ABAQUS macro function. And modifying the codes in Jupyter to realize batch generation of the inp file and batch submission to a solver for solution through a bat file.
4. The PSO-BP neural network-based beam unit structure optimization method according to claim 3, wherein the data extraction of the obtained simulation file after the solution by the solver comprises:
the extracted simulation data comprise stress, strain, displacement, plastic deformation and other data at the connection position of the beam unit structure, and a data set is constructed after the data are processed.
5. The PSO-BP neural network-based beam unit structure optimization method of claim 4, wherein the optimized weights and thresholds comprise:
and analyzing the design parameters and the BP neural network, determining each layer structure of the BP neural network, and training by the BP neural network to obtain the trained design parameters.
6. The PSO-BP neural network-based beam unit structure optimization method of claim 5, wherein the updated designed structure comprises:
and after the parametric design is carried out in the SW, all design parameters are linked through an equation function, and after the parameters are updated in the past, the parameters are updated in the parametric model to generate a three-view map and an engineering drawing.
7. A computer terminal device, comprising:
one or more processors;
a memory coupled to the processor for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement the ABAQUS based beam cell structure optimization method of any one of claims 1 to 4.
8. A computer-readable storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the ABAQUS-based beam cell structure optimization method according to any of claims 1 to 4.
CN202111598711.XA 2021-12-24 2021-12-24 Beam unit structure optimization method based on PSO-BP neural network Pending CN114117871A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111598711.XA CN114117871A (en) 2021-12-24 2021-12-24 Beam unit structure optimization method based on PSO-BP neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111598711.XA CN114117871A (en) 2021-12-24 2021-12-24 Beam unit structure optimization method based on PSO-BP neural network

Publications (1)

Publication Number Publication Date
CN114117871A true CN114117871A (en) 2022-03-01

Family

ID=80362243

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111598711.XA Pending CN114117871A (en) 2021-12-24 2021-12-24 Beam unit structure optimization method based on PSO-BP neural network

Country Status (1)

Country Link
CN (1) CN114117871A (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106202731A (en) * 2016-07-12 2016-12-07 南京理工大学 Bridge crane multi-flexibl e dynamics structural optimization method
CN113221415A (en) * 2021-05-13 2021-08-06 广东省科学院智能制造研究所 Truss girder structure optimization method and device based on ABAQUS

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106202731A (en) * 2016-07-12 2016-12-07 南京理工大学 Bridge crane multi-flexibl e dynamics structural optimization method
CN113221415A (en) * 2021-05-13 2021-08-06 广东省科学院智能制造研究所 Truss girder structure optimization method and device based on ABAQUS

Similar Documents

Publication Publication Date Title
CN109313670A (en) The method and system of lattice suggestion is generated in CAD application
JP7206419B2 (en) Artificial intelligence recommendation model feature processing method, device, electronic device, and computer program
US11281824B2 (en) Authoring loading and boundary conditions for simulation scenarios
CN107016155A (en) The convergence estimate of nonlinear PDEs and linear solution device
CN111488640A (en) Bridge engineering design file generation device, computer equipment and storage medium
US9965574B2 (en) CAD-based initial surface geometry correction
CN113158292A (en) Component matching method, engineering quantity calculation method, device and electronic equipment
CN115427968A (en) Robust artificial intelligence reasoning in edge computing devices
CN111859835A (en) Order reduction method, order reduction device and order reduction equipment of circuit interconnection network model
Paz et al. Lightweight parametric design optimization for 4D printed parts
CN111666393A (en) Verification method and device of intelligent question-answering system, computer equipment and storage medium
CN114398711A (en) Concrete gravity dam CAE integrated analysis method and device and electronic equipment
CN113221415A (en) Truss girder structure optimization method and device based on ABAQUS
CN114117871A (en) Beam unit structure optimization method based on PSO-BP neural network
JP5570228B2 (en) Method and apparatus for calculating simultaneous linear equations
CN115547431A (en) Short-cut fiber composite material fatigue performance prediction method based on machine learning and progressive damage
JP2019106187A (en) System for and method of repairing finite element mesh
US20220012378A1 (en) Method of performing design verification with automatic optimization and related design verification system
CN114757131A (en) Optimization method of proxy model suitable for CFD uncertainty quantification and related equipment
JPWO2021092366A5 (en)
CN110738318B (en) Network structure operation time evaluation and evaluation model generation method, system and device
KR102278284B1 (en) Apparatus and method for producing deep learning model using nested editable block
Hahn et al. A framework to efficiently smooth L 1 penalties for linear regression
US20210103821A1 (en) Dynamic management of weight update bit length
CN112766462A (en) Data processing method, device and computer readable storage medium

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