CN113434971A - Multi-scale welding fatigue life prediction method, device and equipment - Google Patents
Multi-scale welding fatigue life prediction method, device and equipment Download PDFInfo
- Publication number
- CN113434971A CN113434971A CN202110633752.1A CN202110633752A CN113434971A CN 113434971 A CN113434971 A CN 113434971A CN 202110633752 A CN202110633752 A CN 202110633752A CN 113434971 A CN113434971 A CN 113434971A
- Authority
- CN
- China
- Prior art keywords
- fatigue life
- fatigue
- scale
- life prediction
- neural network
- 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 62
- 238000003466 welding Methods 0.000 title claims abstract description 54
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 80
- 238000013528 artificial neural network Methods 0.000 claims abstract description 45
- 239000002245 particle Substances 0.000 claims abstract description 39
- 238000012549 training Methods 0.000 claims abstract description 37
- 239000000463 material Substances 0.000 claims abstract description 23
- 238000005457 optimization Methods 0.000 claims abstract description 23
- 238000004458 analytical method Methods 0.000 claims description 33
- 238000007781 pre-processing Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 13
- 238000003062 neural network model Methods 0.000 claims description 9
- 230000009467 reduction Effects 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000013473 artificial intelligence Methods 0.000 abstract description 6
- 238000011160 research Methods 0.000 abstract description 4
- 238000009661 fatigue test Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- 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
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the specification discloses a multi-scale welding fatigue life prediction method, a multi-scale welding fatigue life prediction device and multi-scale welding fatigue life prediction equipment. The scheme comprises the following steps: and the fatigue life of the welding joint is accurately predicted by combining the characteristics of different algorithms. Analyzing a plurality of influence factors of fatigue life of materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges, service environments and the like by using a rough set algorithm, and obtaining weight coefficients of the influence factors; predicting the fatigue life through training based on a BP neural network algorithm; and (3) improving the problems of slow convergence and local optimum of the BP neural network by utilizing a particle swarm optimization algorithm, and optimizing the established BP neural network to obtain a multi-scale fatigue life prediction model. The multi-scale welding fatigue life prediction method based on the artificial intelligence technology can improve a large amount of manpower and material resources required by fatigue performance research, and provides a new idea while ensuring the fatigue life prediction precision.
Description
Technical Field
The application relates to the technical field of fatigue life prediction of welding structures, in particular to a multi-scale welding fatigue life prediction method, device and equipment.
Background
With the rapid development of world economy and industrial manufacturing technology, the current society puts higher demands on the use safety performance of key parts and large-scale engineering structures. At present, the factor influencing the safety performance is mainly the fatigue failure phenomenon, and the research on the fatigue performance usually needs to consume a large amount of manpower and material resources. Meanwhile, the fatigue performance of the welding part is comprehensively influenced by a plurality of factors such as materials, plate thickness, joint types, welding methods, stress ratios, stress ranges, service environments and the like, so that the comprehensive effect of different influencing factors is difficult to comprehensively consider and analyze in the fatigue failure test of the welding part.
Disclosure of Invention
In view of this, embodiments of the present application provide a method, an apparatus, and a device for predicting a multi-scale welding fatigue life, which are used to improve the prediction accuracy of the fatigue life, and simultaneously avoid the local optimization problem of a neural network and reduce the computation cost.
In order to solve the above technical problem, the embodiments of the present specification are implemented as follows:
the embodiment of the specification provides a multiscale welding fatigue life prediction method, which comprises the following steps:
establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
preprocessing the data in the multi-scale fatigue performance database to remove noise points;
selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
Optionally, the fatigue life influencing factors include: materials, sheet thickness, joint type, welding method, stress ratio, stress range, and service environment.
Optionally, determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm specifically includes:
discretizing the parameter values of the fatigue performance analysis decision system based on a rough set algorithm to realize reduction of fatigue life influence factors;
based on a rough set algorithm, the weight coefficients of different fatigue life influencing factors are obtained through reduction of the fatigue life influencing factors.
Optionally, constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and the BP neural network algorithm specifically includes:
initializing a neural network model, and setting training precision and maximum iteration times;
initializing parameters of a particle swarm optimization algorithm, setting the maximum iteration times and inertia factors, and taking the weight and the threshold of a neural network as position vectors of the particle swarm optimization algorithm;
obtaining fitness values of different particles based on a fitness function;
determining optimal positions of the particles and the population based on the fitness value;
updating the position and speed of the particles, inertia factors and learning factors in the particle swarm optimization algorithm based on the optimal position;
judging the training precision or the maximum iteration number, and if the training precision or the maximum iteration number meets the conditions, using the optimal position obtained at the algorithm stopping time as the optimal parameter of the neural network to realize the optimization of the network model;
training data based on the optimal parameters and the fatigue life prediction rule, and obtaining a multi-scale fatigue life prediction model when the training precision is met;
optionally, the method further includes:
preprocessing the S-N curve data, and verifying the accuracy of the S-N curve data by inputting the S-N curve data into a multi-scale fatigue life prediction model;
inputting test parameters to be predicted, and realizing intelligent prediction of the multi-scale welding fatigue life.
Optionally, the convergence of the neural network is enhanced and the neural network is prevented from falling into local optimum by using the global optimization function of the particle swarm optimization algorithm.
Optionally, the fitness function adopts the following formula:
wherein M represents the number of particles in the particle swarm, N represents the target spatial dimension, Yi,jRepresenting the actual output, Oi,jIndicating the desired output.
Optionally, the initializing the neural network model specifically includes:
setting parameters of an input layer, a hidden layer and an output layer of the neural network model, wherein the material, the plate thickness, the joint type, the welding method, the stress ratio and the service environment are parameters of the input layer, m and C are parameters of the output layer, and C is equal to emSN, S is the stress amplitude, and N is the fatigue life.
The embodiment of this description provides a multiscale welding fatigue life prediction device, includes:
the multi-scale fatigue performance database establishing module is used for establishing a multi-scale fatigue performance database based on documents and fatigue performance data, and the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
the preprocessing module is used for preprocessing the data in the multi-scale fatigue performance database and removing noise points;
the fatigue performance analysis decision system module is used for selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life to establish a fatigue performance analysis decision system;
the weight coefficient determining module is used for determining weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by utilizing a rough set algorithm;
the prediction rule determining module is used for determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
the multi-scale fatigue life prediction model building module is used for building a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and the service life prediction module is used for predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
The embodiment of the present specification provides a multiscale welding fatigue life prediction device, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
preprocessing the data in the multi-scale fatigue performance database to remove noise points;
selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
The embodiment of the specification adopts at least one technical scheme which can achieve the following beneficial effects:
the fatigue life prediction of the welding joint is realized by utilizing algorithms such as machine learning in the field of computer science, and the manpower and material resources consumed by the fatigue test are reduced. By utilizing various algorithm coupling technologies, the fatigue life can be predicted more accurately, the local optimal problem of a neural network is avoided, and the calculation cost is reduced.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a schematic flowchart of a first embodiment of a method for predicting a dimensional welding fatigue life according to an embodiment of the present disclosure;
FIG. 2 is a schematic flowchart of a second embodiment of a multi-scale welding fatigue life prediction method provided in an embodiment of the present disclosure;
FIG. 3 is a graph of the results of three predictions according to an embodiment of the present invention;
FIG. 4 is a graph of four predicted results based on an embodiment of the present invention;
FIG. 5 is a schematic structural diagram of a dimensional welding fatigue life prediction device corresponding to FIG. 1 provided in an embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of a dimensional welding fatigue life prediction device corresponding to fig. 1 provided in an embodiment of the present disclosure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the 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 application.
The artificial intelligence technology is an important branch of computer science, and complex work requiring human intelligence is completed by utilizing machine equipment in different industries through learning, simulating and expanding theoretical basic methods and scientific technology application of human brain. With the continuous development of computer machine learning algorithm, the artificial intelligence technology is applied to the aspect of fatigue life prediction, so that the labor and material resources consumed by a fatigue test can be greatly reduced while the working quality is ensured, and the method has important theoretical and engineering significance for the research of the fatigue performance of key parts and large-scale engineering structures.
The method realizes the prediction of the fatigue life of the welding joint by utilizing algorithms such as machine learning in the field of computer science, and reduces manpower and material resources consumed by fatigue tests. By utilizing various algorithm coupling technologies, the fatigue life can be predicted more accurately, the local optimal problem of a neural network is avoided, and the calculation cost is reduced.
Specifically, through comprehensive consideration of multiple factors such as materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges and service environments, influence weights of the materials and the joint types in the fatigue evolution process are obtained respectively, intelligent prediction of fatigue life is achieved through effective training, and an effective means is provided for research of fatigue performance of structural members.
The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Example one
Fig. 1 is a schematic flowchart of a first embodiment of a method for predicting a dimensional welding fatigue life provided in an embodiment of the present disclosure. From the viewpoint of a program, the execution subject of the flow may be a program installed in an application server or an application client.
As shown in fig. 1, the process may include the following steps:
step 102: establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
step 104: preprocessing the data in the multi-scale fatigue performance database to remove noise points;
step 106: selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
step 108: determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
step 110: determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
step 112: constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
step 114: and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
Based on the method of fig. 1, the embodiments of the present specification also provide some specific implementations of the method, which are described below.
Optionally, the fatigue life influencing factors include: materials, sheet thickness, joint type, welding method, stress ratio, stress range, and service environment.
Optionally, determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm specifically includes:
discretizing the parameter values of the fatigue performance analysis decision system based on a rough set algorithm to realize reduction of fatigue life influence factors;
based on a rough set algorithm, the weight coefficients of different fatigue life influencing factors are obtained through reduction of the fatigue life influencing factors.
Optionally, constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and the BP neural network algorithm specifically includes:
initializing a neural network model, and setting training precision and maximum iteration times;
initializing parameters of a particle swarm optimization algorithm, setting the maximum iteration times and inertia factors, and taking the weight and the threshold of a neural network as position vectors of the particle swarm optimization algorithm;
obtaining fitness values of different particles based on a fitness function;
determining optimal positions of the particles and the population based on the fitness value;
updating the position and speed of the particles, inertia factors and learning factors in the particle swarm optimization algorithm based on the optimal position;
judging the training precision or the maximum iteration number, and if the training precision or the maximum iteration number meets the conditions, using the optimal position obtained at the algorithm stopping time as the optimal parameter of the neural network to realize the optimization of the network model;
training data based on the optimal parameters and the fatigue life prediction rule, and obtaining a multi-scale fatigue life prediction model when the training precision is met;
optionally, the method further includes:
preprocessing the S-N curve data, and verifying the accuracy of the S-N curve data by inputting the S-N curve data into a multi-scale fatigue life prediction model;
inputting test parameters to be predicted, and realizing intelligent prediction of the multi-scale welding fatigue life.
Optionally, the convergence of the neural network is enhanced and the neural network is prevented from falling into local optimum by using the global optimization function of the particle swarm optimization algorithm.
Optionally, the fitness function adopts the following formula:
wherein M represents the number of particles in the particle swarm, N represents the target spatial dimension, Yi,jRepresenting the actual output, Oi,jIndicating the desired output.
Optionally, the initializing the neural network model specifically includes:
setting parameters of an input layer, a hidden layer and an output layer of the neural network model, wherein the material, the plate thickness, the joint type, the welding method, the stress ratio and the service environment are parameters of the input layer, m and C are parameters of the output layer, and C is equal to emSN, S is the stress amplitude, and N is the fatigue life.
Example two
The invention provides a multi-scale welding fatigue life prediction method based on an artificial intelligence technology, which combines the characteristics of different algorithms to realize accurate prediction of the fatigue life of a welding joint. Analyzing a plurality of influence factors of fatigue life of materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges, service environments and the like by using a rough set algorithm, and obtaining weight coefficients of the influence factors; predicting the fatigue life through training based on a BP neural network algorithm; and (3) improving the problems of slow convergence and local optimum of the BP neural network by using a PSO algorithm, and optimizing the established BP neural network to obtain a multi-scale fatigue life prediction model.
Fig. 2 is a schematic flowchart of a second embodiment of a multi-scale welding fatigue life prediction method provided in an embodiment of the present specification, where the method specifically includes the following steps:
the method comprises the following steps: obtaining fatigue performance data such as S-N curves and the like under the conditions of different materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges, service environments and the like through the fatigue test and the prior literature reference, and establishing a multi-scale fatigue performance database;
step two: preprocessing the multi-scale fatigue performance database in the step one, and removing noise points with larger fitting errors by combining different S-N curves;
step three: comprehensively considering different factors influencing the fatigue life, selecting S-N curve data in the fatigue performance database in the step one, and establishing a fatigue performance analysis decision system;
step four: discretizing the parameter values of the fatigue performance analysis decision system in the step three based on a rough set algorithm, and realizing reduction of fatigue life influence factors by reasonably setting a parameter lambda;
step five: based on a rough set algorithm, obtaining weight coefficients of different influence factors through reduction of fatigue life influence factors;
step six: according to the weight coefficient of the fatigue life influence factors obtained in the step five, a fatigue life prediction rule is formulated and embodied in a multi-scale fatigue life prediction model;
step seven: initializing a neural network through setting parameters of an input layer, a hidden layer, an output layer and the like, and setting training precision and times;
step eight: initializing parameters of a PSO algorithm, setting the maximum iteration times and an inertia factor omega, and taking the weight and the threshold of a neural network as a PSO algorithm position vector;
step ten: determining the optimal positions of the particles and the population based on the fitness value obtained in the step nine;
step eleven: updating parameters such as particle position and speed, inertia factor omega, learning factor and the like in the PSO algorithm based on the optimal position obtained in the step ten;
step twelve: judging the training precision or the maximum iteration number, if so, terminating the algorithm and carrying out a thirteen step, and if not, carrying out a ninth step;
step thirteen: the optimal position obtained at the stopping moment of the algorithm is used as the optimal initial weight and the threshold of the neural network to realize the optimization of the network model;
fourteen steps: training data based on the optimal parameters obtained in the step thirteen and the life prediction rules obtained in the step six, and obtaining a multi-scale fatigue life prediction model when the training precision is met;
step fifteen: and preprocessing an actual S-N curve obtained by an experiment, and verifying the accuracy of the actual S-N curve by inputting the actual S-N curve into a multi-scale fatigue life prediction model.
Sixthly, the steps are as follows: inputting test parameters to be predicted, and realizing intelligent prediction of the multi-scale welding fatigue life.
The method for predicting the multi-scale welding fatigue life based on the artificial intelligence technology of the invention is further described in detail with reference to the flow chart of fig. 2.
Example three:
a multi-scale welding fatigue life prediction method based on an artificial intelligence technology is disclosed, and the fatigue life of a welding joint is accurately predicted by combining the characteristics of different algorithms. Analyzing a plurality of influence factors of fatigue life of materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges, service environments and the like by using a rough set algorithm, and respectively obtaining weight coefficients of the influence factors; and (3) improving the problems of slow convergence and local optimum of the BP neural network by using a PSO algorithm, and optimizing the established BP neural network to obtain a multi-scale fatigue life prediction model.
The method comprises the following specific steps:
and obtaining fatigue performance data such as S-N curves and the like under the conditions of different materials, plate thicknesses, joint types, welding methods, stress ratios, stress ranges, service environments and the like through fatigue tests, finite element numerical simulation and the prior literature reference, establishing a multi-scale fatigue performance database, preprocessing the multi-scale fatigue performance database, and removing noise points with large fitting errors.
Selecting data points of the fatigue performance S-N curve library, establishing a fatigue performance analysis decision system, discretizing parameter values of the fatigue performance analysis decision system based on a rough set algorithm, and reasonably setting a parameter lambda to reduce fatigue life influence factors to obtain weight coefficients of different influence factors.
And formulating a fatigue life prediction rule according to the weight coefficient of the fatigue life influencing factor, and embodying the fatigue life prediction rule in a multi-scale fatigue life prediction model.
The neural network is initialized through setting parameters of an input layer, a hidden layer, an output layer and the like, and training precision and times are set. Wherein, the material, the plate thickness, the joint type, the welding method, the stress ratio, the stress range and the service environment are input layers, and the fatigue life is an output layer.
Initializing parameters of a PSO algorithm, setting the maximum iteration times and an inertia factor omega, and taking the weight and the threshold of a neural network as a PSO algorithm position vector. And obtaining fitness values of different particles based on the fitness function, determining the optimal positions of the particles and the population, and updating the positions and the speeds of the particles, the inertia factor omega, the learning factor and other parameters in the PSO algorithm. And judging the training precision or the maximum iteration number, if the training precision or the maximum iteration number is not met, repeatedly obtaining the fitness values of different particles and the following steps, and if the training precision or the maximum iteration number is not met, terminating the algorithm and taking the obtained optimal position as the optimal initial weight and the threshold of the neural network to realize the optimization of the network model. Training the training data based on the obtained optimal parameters and the life prediction rule, and obtaining a multi-scale fatigue life prediction model when the training precision is met. And preprocessing an actual S-N curve obtained by an experiment, inputting the actual S-N curve into a multi-scale fatigue life prediction model, and verifying the accuracy of the life prediction model. And finally, inputting test parameters to be predicted to realize intelligent prediction of the multi-scale welding fatigue life, wherein part of prediction results are shown in figure 3.
Example four:
similar to the above case, firstly, a plurality of influence factors of fatigue life such as materials, plate thicknesses, joint types, welding methods, stress ratios, service environments and the like are analyzed by using a rough set algorithm, and weight coefficients of the influence factors are respectively obtained; and (3) improving the problems of slow convergence and local optimum of the BP neural network by using a PSO algorithm, and optimizing the established BP neural network to obtain a multi-scale fatigue life prediction model.
The method comprises the following specific steps:
obtaining fatigue performance data such as S-N curves of different materials, plate thicknesses, joint types, welding methods, stress ratios, service environments and the like through fatigue tests, finite element numerical simulation and prior literature reference, and obtaining an S-N curve formula (e) in an exponential function formmSN ═ C; wherein S is stress amplitude, N is fatigue life) and parameters m and C related to fatigue life influence factors are calculated, and a multi-scale fatigue performance database is obtained through processing and is preprocessed, and noise points with large fitting errors are removed.
Selecting data points of the fatigue performance S-N curve library, establishing a fatigue performance analysis decision system, discretizing parameter values of the fatigue performance analysis decision system based on a rough set algorithm, and reasonably setting a parameter lambda to reduce fatigue life influence factors to obtain weight coefficients of different influence factors.
And formulating a fatigue life prediction rule according to the weight coefficient of the fatigue life influencing factor, and embodying the fatigue life prediction rule in a multi-scale fatigue life prediction model. The neural network is initialized through setting parameters of an input layer, a hidden layer, an output layer and the like, and training precision and times are set. Wherein, the material, the plate thickness, the joint type, the welding method, the stress ratio and the service environment are input layers, and the parameters m and C are output layers. Initializing parameters of a PSO algorithm, setting the maximum iteration times and an inertia factor omega, and taking the weight and the threshold of a neural network as a PSO algorithm position vector. And obtaining fitness values of different particles based on the fitness function, determining the optimal positions of the particles and the population, and updating the positions and the speeds of the particles, the inertia factor omega, the learning factor and other parameters in the PSO algorithm. And judging the training precision or the maximum iteration number, if the training precision or the maximum iteration number is not met, repeatedly obtaining the fitness values of different particles and the following steps, and if the training precision or the maximum iteration number is not met, terminating the algorithm and taking the obtained optimal position as the optimal initial weight and the threshold of the neural network to realize the optimization of the network model. Training the training data based on the obtained optimal parameters and the life prediction rule, and obtaining a multi-scale fatigue life prediction model when the training precision is met. And (3) calculating parameters m and C of an actual S-N curve obtained by an experiment, inputting the parameters m and C into the multi-scale fatigue life prediction model, and verifying the accuracy of the life prediction model. Finally, test parameters needing to be predicted are input to obtain a corresponding S-N curve, intelligent prediction of the multi-scale welding fatigue life is achieved according to different stress ranges, and partial prediction results are shown in fig. 4.
Based on the same idea, the embodiment of the present specification further provides a device corresponding to the above method. Fig. 5 is a schematic structural diagram of a dimensional welding fatigue life prediction device corresponding to fig. 1 provided in an embodiment of the present disclosure. As shown in fig. 5, the apparatus may include:
a multi-scale fatigue performance database establishing module 502, configured to establish a multi-scale fatigue performance database based on literature and fatigue performance data, where the multi-scale fatigue performance database includes S-N curve data under different fatigue life influencing factors;
a preprocessing module 504, configured to preprocess data in the multi-scale fatigue performance database and remove noise points;
a fatigue performance analysis decision system module 506, configured to select, based on different factors affecting fatigue life, the S-N curve data in the preprocessed multi-scale fatigue performance database, and establish a fatigue performance analysis decision system;
a weight coefficient determining module 508, configured to determine, according to the fatigue performance analysis decision system, weight coefficients of different fatigue life influencing factors by using a rough set algorithm;
a prediction rule determining module 510, configured to determine a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
a multi-scale fatigue life prediction model construction module 512, configured to construct a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and the service life prediction module 514 is used for predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
Fig. 6 is a schematic structural diagram of a dimensional welding fatigue life prediction device corresponding to fig. 1 provided in an embodiment of the present disclosure. The apparatus 600, comprising:
at least one processor 610; and the number of the first and second groups,
a memory 630 communicatively coupled to the at least one processor 610; wherein,
the memory 630 stores instructions 620 executable by the at least one processor 610 to enable the at least one processor 610 to:
establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
preprocessing the data in the multi-scale fatigue performance database to remove noise points;
selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A multiscale welding fatigue life prediction method is characterized by comprising the following steps:
establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
preprocessing the data in the multi-scale fatigue performance database to remove noise points;
selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
2. The method of claim 1, wherein the fatigue life affecting factors comprise: materials, sheet thickness, joint type, welding method, stress ratio, stress range, and service environment.
3. The method of claim 1, wherein determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm specifically comprises:
discretizing the parameter values of the fatigue performance analysis decision system based on a rough set algorithm to realize reduction of fatigue life influence factors;
based on a rough set algorithm, the weight coefficients of different fatigue life influencing factors are obtained through reduction of the fatigue life influencing factors.
4. The method of claim 1, wherein constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm specifically comprises:
initializing a neural network model, and setting training precision and maximum iteration times;
initializing parameters of a particle swarm optimization algorithm, setting the maximum iteration times and inertia factors, and taking the weight and the threshold of a neural network as position vectors of the particle swarm optimization algorithm;
obtaining fitness values of different particles based on a fitness function;
determining optimal positions of the particles and the population based on the fitness value;
updating the position and speed of the particles, inertia factors and learning factors in the particle swarm optimization algorithm based on the optimal position;
judging the training precision or the maximum iteration number, and if the training precision or the maximum iteration number meets the conditions, using the optimal position obtained at the algorithm stopping time as the optimal parameter of the neural network to realize the optimization of the network model;
training the training data based on the optimal parameters and the fatigue life prediction rule, and obtaining a multi-scale fatigue life prediction model when the training precision is met.
5. The method of claim 4, wherein the method further comprises:
preprocessing the S-N curve data, and verifying the accuracy of the S-N curve data by inputting the S-N curve data into a multi-scale fatigue life prediction model;
inputting test parameters to be predicted, and realizing intelligent prediction of the multi-scale welding fatigue life.
6. The method of claim 4, wherein a global optimization function of a particle swarm optimization algorithm is utilized to enhance convergence of the neural network and avoid it from falling into local optima.
8. The method of claim 4, wherein initializing the neural network model specifically comprises:
setting parameters of an input layer, a hidden layer and an output layer of the neural network model, wherein the material, the plate thickness, the joint type, the welding method, the stress ratio and the service environment are parameters of the input layer, m and C are parameters of the output layer, and C is equal to emSN, S is the stress amplitude, and N is the fatigue life.
9. A multiscale weld fatigue life prediction apparatus, comprising:
the multi-scale fatigue performance database establishing module is used for establishing a multi-scale fatigue performance database based on documents and fatigue performance data, and the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
the preprocessing module is used for preprocessing the data in the multi-scale fatigue performance database and removing noise points;
the fatigue performance analysis decision system module is used for selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life to establish a fatigue performance analysis decision system;
the weight coefficient determining module is used for determining weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by utilizing a rough set algorithm;
the prediction rule determining module is used for determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
the multi-scale fatigue life prediction model building module is used for building a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and the service life prediction module is used for predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
10. A multiscale weld fatigue life prediction device, comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
establishing a multi-scale fatigue performance database based on literature and fatigue performance data, wherein the multi-scale fatigue performance database comprises S-N curve data under different fatigue life influence factors;
preprocessing the data in the multi-scale fatigue performance database to remove noise points;
selecting the preprocessed S-N curve data in the multi-scale fatigue performance database based on different factors influencing the fatigue life, and establishing a fatigue performance analysis decision system;
determining the weight coefficients of different fatigue life influencing factors according to the fatigue performance analysis decision system by using a rough set algorithm;
determining a prediction rule of the fatigue life according to the weight coefficient of the fatigue life influencing factor;
constructing a multi-scale fatigue life prediction model based on the fatigue life prediction rule and a BP neural network algorithm;
and predicting the service life of the weldment to be tested according to the multi-scale fatigue life prediction model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110633752.1A CN113434971A (en) | 2021-06-07 | 2021-06-07 | Multi-scale welding fatigue life prediction method, device and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110633752.1A CN113434971A (en) | 2021-06-07 | 2021-06-07 | Multi-scale welding fatigue life prediction method, device and equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113434971A true CN113434971A (en) | 2021-09-24 |
Family
ID=77803981
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110633752.1A Pending CN113434971A (en) | 2021-06-07 | 2021-06-07 | Multi-scale welding fatigue life prediction method, device and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113434971A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114021468A (en) * | 2021-11-11 | 2022-02-08 | 天津大学 | Fatigue life prediction method, device and equipment for multi-scale welding structure |
CN114595625A (en) * | 2022-01-25 | 2022-06-07 | 天津大学 | Welding structure fatigue performance analysis method based on data driving method |
CN114781250A (en) * | 2022-04-11 | 2022-07-22 | 西南交通大学 | Multi-factor influence environment fatigue life prediction method based on machine learning |
CN116467939A (en) * | 2023-04-10 | 2023-07-21 | 中国电子产品可靠性与环境试验研究所((工业和信息化部电子第五研究所)(中国赛宝实验室)) | Electric energy meter service life prediction method, electric energy meter service life prediction device, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077320A (en) * | 2013-01-18 | 2013-05-01 | 大连交通大学 | Welding fatigue analyzing method based on a rough set theory |
CN108153999A (en) * | 2018-01-26 | 2018-06-12 | 大连交通大学 | A kind of welding point fatigue life intelligent Forecasting based on fatigue characteristic domain |
CN111369141A (en) * | 2020-03-04 | 2020-07-03 | 大连交通大学 | Method for evaluating fatigue life influence factors of welded structure |
US20200394348A1 (en) * | 2019-06-12 | 2020-12-17 | Sichuan University | Method for assessing fatigue damage and a fatigue life based on a crystal plastic welding process model |
-
2021
- 2021-06-07 CN CN202110633752.1A patent/CN113434971A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077320A (en) * | 2013-01-18 | 2013-05-01 | 大连交通大学 | Welding fatigue analyzing method based on a rough set theory |
CN108153999A (en) * | 2018-01-26 | 2018-06-12 | 大连交通大学 | A kind of welding point fatigue life intelligent Forecasting based on fatigue characteristic domain |
US20200394348A1 (en) * | 2019-06-12 | 2020-12-17 | Sichuan University | Method for assessing fatigue damage and a fatigue life based on a crystal plastic welding process model |
CN111369141A (en) * | 2020-03-04 | 2020-07-03 | 大连交通大学 | Method for evaluating fatigue life influence factors of welded structure |
Non-Patent Citations (1)
Title |
---|
邹丽: ""焊接接头疲劳寿命智能评价方法及其应用研究"", 《中国博士学位论文全文数据库(电子期刊)》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114021468A (en) * | 2021-11-11 | 2022-02-08 | 天津大学 | Fatigue life prediction method, device and equipment for multi-scale welding structure |
CN114021468B (en) * | 2021-11-11 | 2024-05-07 | 天津大学 | Fatigue life prediction method, device and equipment for multi-scale welding structure |
CN114595625A (en) * | 2022-01-25 | 2022-06-07 | 天津大学 | Welding structure fatigue performance analysis method based on data driving method |
CN114781250A (en) * | 2022-04-11 | 2022-07-22 | 西南交通大学 | Multi-factor influence environment fatigue life prediction method based on machine learning |
CN116467939A (en) * | 2023-04-10 | 2023-07-21 | 中国电子产品可靠性与环境试验研究所((工业和信息化部电子第五研究所)(中国赛宝实验室)) | Electric energy meter service life prediction method, electric energy meter service life prediction device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113434971A (en) | Multi-scale welding fatigue life prediction method, device and equipment | |
Gomes et al. | A multiobjective sensor placement optimization for SHM systems considering Fisher information matrix and mode shape interpolation | |
Zhang et al. | Collaborative multifidelity-based surrogate models for genetic programming in dynamic flexible job shop scheduling | |
Sathish | Prediction of springback effect by the hybridisation of ANN with PSO in wipe bending process of sheet metal | |
Fei et al. | Enhanced network learning model with intelligent operator for the motion reliability evaluation of flexible mechanism | |
Ewees et al. | Enhanced salp swarm algorithm based on firefly algorithm for unrelated parallel machine scheduling with setup times | |
Atkeson | Using locally weighted regression for robot learning | |
Su et al. | Algorithms for solving assembly sequence planning problems | |
CN111127246A (en) | Intelligent prediction method for transmission line engineering cost | |
Kedong et al. | Improved grey prediction model based on exponential grey action quantity | |
CN103065042A (en) | Multiple target comprehensive decision evaluation method based on scene | |
CN108460191B (en) | Optimization method of machine tool cross beam design scheme based on quality function expansion | |
Wang et al. | Digital production control of manufacturing workshop based on internet of things | |
CN117370766A (en) | Satellite mission planning scheme evaluation method based on deep learning | |
Ghosh et al. | A method for capturing the entire fuzzy non-dominated set of a fuzzy multi-criteria optimization problem | |
Aljahdali | Development of software reliability growth models for industrial applications using fuzzy logic | |
Lavrov et al. | Human-centered management in polyergatic information systems. Multi-criteria distribution of functions between operators | |
Rao et al. | Fuzzy logic-based expert system to predict the results of finite element analysis | |
Egrioglu et al. | Modified pi sigma artificial neural networks for forecasting | |
Jiménez Tovar et al. | An optimization algorithm for the multi-objective flexible fuzzy job shop environment with partial flexibility based on adaptive teaching–learning considering fuzzy processing times | |
CN117892615A (en) | Assembly equipment state evaluation and dynamic adjustment method based on digital twin model | |
Motta et al. | Development of a computational efficient tool for robust structural optimization | |
Mirzapourrezaei et al. | Simulation of a manufacturing assembly line based on witness | |
CN110059871A (en) | Photovoltaic power generation power prediction method | |
CN114021468B (en) | Fatigue life prediction method, device and equipment for multi-scale welding structure |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210924 |