CN115685881A - Low-stress high-precision electric arc additive process control method based on computational intelligence - Google Patents
Low-stress high-precision electric arc additive process control method based on computational intelligence Download PDFInfo
- Publication number
- CN115685881A CN115685881A CN202211395464.8A CN202211395464A CN115685881A CN 115685881 A CN115685881 A CN 115685881A CN 202211395464 A CN202211395464 A CN 202211395464A CN 115685881 A CN115685881 A CN 115685881A
- Authority
- CN
- China
- Prior art keywords
- arc additive
- additive manufacturing
- workpiece
- precision
- electric arc
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P10/00—Technologies related to metal processing
- Y02P10/25—Process efficiency
Landscapes
- Arc Welding In General (AREA)
Abstract
The invention provides a low-stress high-precision electric arc additive process control method based on computational intelligence. The method comprises the steps of quantitatively predicting residual stress, surface flatness, weld bead fusion width and weld bead residual height of an arc additive manufacturing workpiece under the synergistic effect of laser-electromagnetism-ultrasound based on an improved wavelet neural network algorithm, taking the closeness degree of each index predicted value and a target value and the uniformity weighted combination of the predicted values as a unified control optimization target, and finally scientifically, efficiently and accurately seeking the optimal process parameter combination by utilizing the improved Hui wolf optimization algorithm, so that a key process cooperative optimization method is provided for accurate, stable and rapid automatic control of a low-stress high-precision metal arc additive manufacturing process, and the method has important revelation and reference significance for developing an arc additive manufacturing technology of a large-size metal complex component with low stress, high precision, high reliability and high efficiency.
Description
Technical Field
The patent relates to the technical field of electric arc additive manufacturing and automation, in particular to a low-stress high-precision electric arc additive process control method based on computational intelligence.
Background
Additive Manufacturing (AM), as a new generation of manufacturing technology with revolutionary significance, is considered as a driving engine that drives the manufacturing industry to upgrade. The additive manufacturing adopts a layer-by-layer accumulation mode to produce the required parts, and compared with the traditional additive reduction or equal material manufacturing technology, the additive manufacturing has the advantages of rapid near-net forming technology, high material utilization rate, low manufacturing cost, short production period and good material performance, and particularly can realize the die-free, high-freedom and customized forming of precise and complex parts.
Metal additive manufacturing is one of the most important branches of additive manufacturing technology. Metal additive manufacturing is mainly classified into laser additive manufacturing, electron beam additive manufacturing, and arc additive manufacturing according to a heat source. The electric arc additive manufacturing technology has the advantages of high cladding efficiency, large forming size, mature equipment, high expandability and the like, and has wide application prospects in the field of metal structural materials.
The electric arc additive manufacturing technology is developed from the traditional argon arc welding, the process foundation is deep, and the technology has great technical popularization and application potential. The electric arc additive manufacturing is insensitive to environment and metal materials, high in deposition efficiency, high in wire material utilization rate, short in whole manufacturing period and low in production cost, can form aluminum alloy and copper alloy with high laser reflectivity, and is suitable for the technology of quickly forming large metal components with complex structures. In addition, the parts manufactured by the electric arc additive manufacturing are made of all-welded seam metal, have uniform chemical components and high density, have mechanical properties superior to those of cast parts, can reach the level of forged parts by proper tempering means, and have the advantages of high strength and good toughness compared with integrally forged parts. Compared with the traditional material reduction manufacturing, the electric arc material increase manufacturing can shorten the forming time by 40-60%, the material utilization rate is high, and the subsequent machining time can also be shortened by 15-20%. The continuous development of the electric arc additive manufacturing technology enables the additive manufacturing of large-scale complex structural components, and compared with the traditional material reducing processing method, the electric arc additive manufacturing can save 78% of the raw material cost.
However, the electric arc additive manufacturing is not perfect, and has many disadvantages, the traditional electric arc additive has low energy utilization rate, high heat input, serious workpiece surface oxidation, the melting bath and the nearby part thereof are heated sharply at a speed far higher than the surrounding area to generate local melting, the material expands due to heating, the thermal expansion is restricted by the surrounding cooler area, the material generates thermal stress, after cooling, the material is relatively shortened, narrowed or reduced compared with the surrounding area, tensile stress is easily formed at the top, compression stress is formed at the bottom, the whole workpiece deforms, the dimensional accuracy is seriously influenced, the processing quality and the manufacturing accuracy of the workpiece cannot reach the manufacturing level of the traditional processing method, the process is difficult to realize standardization, and the quality consistency of the workpiece is difficult to ensure; furthermore, in the process of electric arc additive manufacturing, residual stress is generated by periodicity, acuteness, unsteadiness, cyclic heating and cooling, short-time unbalanced cyclic solid-state phase change and rapid solidification shrinkage of a moving molten pool under strong constraint, so that the workpiece is easily damaged by premature fracture due to load instability in the using process, and the service life of the formed workpiece is shortened.
Based on the problems of the conventional electric arc additive manufacturing process, the forming quality and the residual stress of the conventional electric arc additive manufacturing can be controlled and improved by adopting a composite forming technology and process parameter optimization. The method has the advantages that the laser assistance effect is increased, the arc combustion stability is improved, the cladding depth and the cladding rate are improved, the residual stress and deformation of a cladding layer are smaller, the electromagnetic assistance effect is increased, the support is provided for a welding channel molten pool, the shape and the characteristics of the molten pool are regulated and controlled, the internal structure of a forming layer is improved, the forming precision of a workpiece is improved, the ultrasonic impact treatment is increased, the fusing behavior of the molten pool is interfered, the structural structure form is regulated and controlled, the residual stress is reduced, the mechanical property is improved, an arc material adding process with the effects of concentrating an arc, improving the energy utilization rate and the stability of the arc is developed under the synergistic effect of laser-electromagnetism-ultrasound, the deposition layer precision is improved, the surface quality and the process performance of the arc material adding workpiece are improved, and the low-stress high-precision rapid forming technology of the arc material adding manufacturing process is realized.
Moreover, the laser-electromagnetic-ultrasonic assisted conventional arc additive manufacturing process is mainly embodied in how laser, electromagnetic, ultrasonic and arc additive manufacturing process parameters are matched, a high-quality forming layer can be obtained only by effectively and cooperatively controlling a plurality of process parameters with coupling effects, and the premise of ensuring forming precision and proper residual stress is also provided.
With the progress of technology and the increasing complexity of engineering practice problems, optimization problems to be actually solved are increasing complex, the scale of design variables is increasing, and constraint conditions are increasing. For this reason, the scholars try to develop new optimization methods, and the computational intelligence algorithm comes as it goes, which is a branch of artificial intelligence, and is a typical representative of the joint sense, also called bionic school or physiological school, and has the following functions: (1) The computational intelligence method adopts a heuristic random search strategy to search and optimize in a global space of the problem, and can find a global optimal solution or an acceptable solution within an acceptable time. (2) When the computational intelligent algorithm is used for processing the optimization problem, strict mathematical derivation is not needed for solving the problem, and the computational intelligent algorithm has good global search capability, universal adaptability and solving robustness.
In conclusion, according to the technical characteristics of the conventional electric arc additive manufacturing process assisted by laser-electromagnetism-ultrasound, an evaluation standard and process parameter determination method under a novel structure form is established on the basis of an advanced calculation intelligent algorithm, and is a key basis for realizing low-stress high-precision electric arc additive manufacturing.
Disclosure of Invention
The invention aims to provide a low-stress high-precision electric arc additive process control method based on computational intelligence. The method comprises the steps of preparing a workpiece by using an electric arc additive manufacturing process under the synergistic effect of laser-electromagnetism-ultrasound, quantitatively predicting residual stress, surface flatness, weld bead fusion width and weld bead residual height of the workpiece by using an improved wavelet neural network algorithm, taking the closeness degree of each index predicted value and a target value and the uniformity weighted combination of the predicted values as a unified control optimization target, and scientifically, efficiently and accurately seeking the optimal process combination by using an improved wolf optimization algorithm to finally determine a parameter control method of the low-stress high-precision electric arc additive manufacturing process.
In order to realize the purpose of the invention, the invention adopts the following technical scheme:
the low-stress high-precision electric arc additive process control method based on computational intelligence comprises the following steps:
(a) Preparing a workpiece by utilizing an electric arc additive manufacturing process under the synergistic action of laser, electromagnetism and ultrasound, and analyzing and acquiring residual stress and a welding bead section form function of a specific area of the electric arc additive manufacturing workpiece by combining various detection means and data, wherein the welding bead section form function comprises information such as surface flatness, welding bead fusion width and welding bead extra height of the workpiece;
further, the electric arc additive manufacturing process under the laser-electromagnetic-ultrasonic synergistic effect is that on the basis of a conventional electric arc additive manufacturing process, the laser auxiliary effect is increased to improve the cladding depth and the cladding rate, the electromagnetic auxiliary effect is increased to support a welding channel molten pool, the shape and the characteristics of the molten pool are regulated and controlled, the workpiece forming precision is improved, the ultrasonic impact treatment is increased to reduce the residual stress, the mechanical property is improved, and the low-stress high-precision rapid forming technology of the electric arc additive manufacturing process is realized under the laser-electromagnetic-ultrasonic synergistic effect, wherein a laser beam is coaxial with an electric arc of a welding gun, an excitation coil is fixed on the welding gun and is coaxial with a nozzle of the welding gun, ultrasonic waves can be applied to ultrasonic vibration by adopting an action mode of a fixed ultrasonic device at the bottom of a forming substrate or a follow-up ultrasonic impact system of the welding gun, and when the formed piece is large in size or complicated in shape, the follow-up ultrasonic impact system of the welding gun is preferably used.
Furthermore, the weld bead section form function is a digitalized image of the weld bead section form obtained through a scanning electron microscope or three-dimensional laser scanning and the like, then weld bead data are simplified and smoothed, then fitting and fitting error and mean square error analysis are carried out by adopting different functions based on a least square method, and finally a quantitative function relation between the weld bead fusion width (independent variable) and the weld bead residual height (dependent variable) with simple form and high precision is obtained.
(b) On the premise of ensuring the reliability and effectiveness of the test result, repeating the step (a), determining the test scheme of the low-stress high-precision arc additive manufacturing process parameter control method according to the orthogonal test design principle based on the process parameter constraint range, and inputting data which are respectively welding voltage U e Welding current I e Frequency of current f e Height h of tungsten electrode w Scanning velocity v w Weld bead spacing d s Wire feed rate v f Substrate preheating temperature T b Protective gas flow Q g Laser power P l Diameter d of light spot l Defocus amount d d Magnetic induction B m Frequency f of magnetic field m The number n of the ultrasonic impact pins u Diameter d of ultrasonic impact needle u Ultrasonic impact amplitude A u Ultrasonic impact frequency f u (ii) a The output data is the residual stress of a certain specific area i of the electric arc additive manufacturing workpiece
Surface flatness χ i Weld bead fusion width W i Weld bead height H i (ii) a Reorganizing input data by utilizing class balance and inter-class intersection, performing primary and secondary factor division by utilizing a display rule of the data or mining an internal rule implied by the data through a grey correlation degree analysis method to obtain an input data weight, integrating the input data into an input quantity, and determining a hidden layer stimulusAn excitation function and an output layer excitation function are used for finally establishing a prediction model of residual stress, surface flatness, weld bead fusion width and weld bead residual height of an electric arc additive manufacturing workpiece by using a conjugate gradient wavelet neural network based on local learning;
further, the process parameter constraint range refers to the maximum and minimum values of the following parameters: welding voltage U e Welding current I e Frequency f of current e Height h of tungsten electrode w Scanning velocity v w Weld bead spacing d s Wire feed rate v f Substrate preheating temperature T b And protective gas flow Q g Laser power P l Spot diameter d l Defocus amount d d Magnetic induction B m Frequency f of magnetic field m N number of ultrasonic impact pins u Diameter d of ultrasonic impact needle u Ultrasonic impact amplitude A u Ultrasonic impact frequency f u In particular, if the ultrasonic vibration is applied by the apparatus of the fixed bottom of the shaped substrate, the number n of ultrasonic impact pins u ＝1。
Further, the surface flatness χ i The calculation formula of (c) is:
in the formula (I), the compound is shown in the specification,
-the height of the s peak at the surface of the workpiece cladding layer;
n is the number of the wave crests and the wave troughs within a certain length range;
note that the surface flatness χ i The value can fully reflect the change characteristic of the height of the geometric parameter of the surface of the workpiece, and the smaller the value is, the flatter the cladding forming plane is, and the more smooth the workpiece isThe better the surface quality of the part.
Furthermore, the hidden layer excitation function selects a Morlet wavelet function, and the output layer excitation function selects a Sigomid function.
Further, the selection mode of the kernel function mainly comprises prior knowledge, cross validation and a mixed kernel function.
(c) Recording the maximum value sigma of the residual stress in each selected area of the workpiece manufactured by the electric arc additive manufacturing in all the test results in the step (b) RSmax And minimum value σ RSmin Surface flatness maximum χ max And minimum value χ min Maximum value of weld bead fusion width W max And minimum value W min Maximum value of weld bead height H max And a minimum value H min ；
(d) An improved Hui wolf optimization algorithm is utilized to seek an optimal process scheme combination, effective control over a low-stress high-precision electric arc additive manufacturing process is achieved, a uniform target evaluation function F (X) of each index (namely output data) is established to serve as a fitness function to evaluate the quality of a solution corresponding to a variable, the smaller the value of the uniform target evaluation function F (X), the better the solution corresponding to the variable is explained, and the expression of the fitness function is as follows:
wherein X is a design variable, X = [ U ] e ,I e ,f e ,h w ,v w ,d s ,v f ,T b ,Q g ,P l ,d l ,d d ,B m ,f m ,n u ,d u ,A u ,f u ]；
λ 1 ,λ 2 ,λ 3 ,λ 4 ,η 1 ,η 2 ,η 3 ,η 4 -weighting factors, the value range of which is 0 to 1, the value of each value can be adjusted in the range of (0, 1) according to different requirements for each index parameter, wherein λ 1 +λ 2 +λ 3 +λ 4 ＝1；
n is the total number of the specific areas of the workpiece;
W i* ,H i* target values of weld bead fusion width and weld bead residual height of a specific area i of the workpiece;
in the formula (1), G is 1 (X)、G 2 (X)、G 3 (X)、G 4 The first term in (X) is used to assess the proximity of the indicator to a target value, where G 1 (X)、G 2 (X) the target values of residual stress and roughness are zero, and the second term is used for evaluating the uniformity of the index;
(e) Initializing, determining the number m of the grey wolf individuals in the search population, and randomly generating the position X of the m grey wolf in the search domain determined by the process parameter constraint conditions in the step (b) j (k) (j =1,2, \ 8230;, m), the maximum number of iterations k max As an optimization termination condition, and making the current iteration number k =0;
furthermore, the number m of the wolf individuals in the population ranges from 20 to 100.
Further, the maximum number of iterations k max The value range is 200-1000.
(f) Calculating the fitness value of each head of the population by using a formula (1), sorting according to the fitness value, comparing, and selecting the first three best grey wolf positions as an optimal position, a suboptimal position and a candidate position corresponding to alpha, beta and delta wolfs
(g) Updating the m-head wolf body position X by formula (2) j (k+1)(j＝1,2,…,m)：
In the formula, phi 1 ,φ 2 ,φ 3 -addingA weight coefficient;
w α ,w β ,w δ -an inertia weight;
μ α ,μ β ,μ δ -a non-linear modulation index, ranging from 0 to 3;
rand α (0,1),rand β (0,1),rand δ (0,1),rand′ α (0,1),rand′ β (0,1),rand′ δ (0,1)—[0,1]a random number in between;
(h) Calculating the fitness value of all the gray wolfs by using the formula (1) again to evaluate the whole gray wolf population;
(i) Judging whether the constraint condition in the formula (3) is satisfied, if so, directly switching to the step (j), and if not, switching to the step (g);
(j) Updating the fitness value and the position of the first three best global beta, delta wolfs,
(k) Judging whether an iteration condition k of the algorithm is satisfied, wherein k is more than k max If yes, enabling k = k +1, turning to the step (g), continuing to perform algorithm iteration, and otherwise, directly turning to the step (l);
(l) Outputting the current alpha wolf optimal position, stopping iteration, namely finding the global optimal solution by the algorithm
And finally, obtaining the optimal control method of the parameters of the low-stress high-precision electric arc additive manufacturing process.
Compared with the prior art, the invention has the following advantages and effects:
compared with the prior art, the invention realizes the control method of the low-stress high-precision electric arc additive manufacturing process parameters, and has the following advantages: (1) Determining an evaluation method of arc additive manufacturing process parameters under the synergistic effect of laser-electromagnetism-ultrasound on workpiece forming quality (evaluated by a welding channel section form function, which contains information such as surface flatness, welding channel molten width, welding channel residual height and the like of a workpiece) and residual stress effect, solving the problem of lack of the current evaluation standard, and providing an important reference basis for enriching and perfecting an additive manufacturing standard system; (2) The wavelet neural network is a product of combining a wavelet analysis theory and a neural network theory, has more degrees of freedom and stronger approximation ability, can achieve better approximation ability and faster convergence speed by properly selecting parameters, further improves approximation performance and convergence accuracy of the network after improving a learning strategy by using a conjugate gradient algorithm based on local learning, and improves model prediction efficiency and accuracy; (3) Compared with the conventional gray wolf optimization algorithm, the improved gray wolf optimization algorithm can update the position by utilizing the weighted synthesis of the optimal position, enhance the exploration capability of a search mechanism, avoid falling into the local optimum, balance the global search capability and the local development capability of the algorithm by adjusting the self-adaptive value of the convergence factor through the nonlinear modulation index, improve the convergence precision and accelerate the convergence speed, and enable the conventional gray wolf optimization to have more outstanding superiority; (4) Based on an improved intelligent optimization algorithm, the surface quality and the dimensional accuracy of an electric arc additive manufacturing workpiece are improved, the dimensional accuracy of the workpiece is ensured, the residual stress amplitude is reduced, the fatigue strength is improved, the service life is prolonged, a key process optimization method is provided for the accuracy and the stabilization control of a low-stress high-accuracy electric arc additive manufacturing process, the method has a reference value for developing an electric arc additive manufacturing technology of a large-size complex component which is low in stress, high in accuracy, high in reliability and high in efficiency and is suitable for wide application in the industrial field, and the method has important inspiration and reference significance for development of other additive manufacturing technologies, additive manufacturing frontiers and innovation technology exploration at present.
Drawings
In order to illustrate embodiments of the invention or prior art solutions more clearly, the drawings that are needed in the description of embodiments or prior art will be briefly described below, it being understood that the following drawings only show some embodiments of the invention and are therefore not to be considered limiting of scope, and that other relevant drawings may be derived from these drawings for a person skilled in the art without inventive effort.
FIG. 1 is a general flow chart of a low stress high accuracy electric arc additive process control method based on computational intelligence;
FIG. 2 is a schematic structural diagram of processing equipment for preparing a 2195 aluminum-lithium alloy workpiece by an arc additive manufacturing process under the synergistic effect of laser-electromagnetism-ultrasound in the embodiment;
wherein: 1-a laser beam; 2-a welding gun; 3-hollow tungsten electrode; 4-a field coil; 5-a workpiece; 6-a substrate; 7-forming the bottom of the substrate by a fixed ultrasonic device; 8-a welding gun follow-up type ultrasonic impact system; 21-a welding gun nozzle; 81-ultrasonic impact pin.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in detail and completely below with reference to the accompanying drawings of the embodiments of the present invention, however, the present invention can be implemented in many different forms, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. Thus, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely representative of selected embodiments of the invention, which, on the contrary, is provided for the purpose of providing a more thorough understanding of the present disclosure. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The parts not specifically described in the following embodiments are all known in the prior art and will not be described herein.
In the embodiment, a 2195 aluminum-lithium alloy electric arc additive manufacturing forming piece is selected as a test object, the wire is an ER4047Al-Si welding wire, the specification of the welding wire is 1.2mm, a forming substrate is 2195Al-Li alloy, the thickness of a plate is 3mm, and welding shielding gas is high-purity argon of 99.99%.
The low-stress high-precision electric arc additive process control method based on the computational intelligence comprises the following steps (a general flow chart is shown in figure 1):
(a) The method comprises the steps of preparing a 2195 aluminum lithium alloy workpiece by utilizing an electric arc additive manufacturing process under the synergistic action of laser-electromagnetism-ultrasound, wherein as shown in figure 2, a workpiece 5 is fixed on a substrate 6, a laser beam 1 is coaxial with an electric arc of a hollow tungsten electrode 3 of a welding gun 2, an excitation coil 4 is fixed on the welding gun 2 and is coaxial with a welding gun nozzle 21, ultrasonic vibration can be applied by adopting the action mode of a fixed ultrasonic device 7 at the bottom of a forming substrate or a welding gun follow-up ultrasonic impact system 8, in the embodiment, ultrasonic vibration is applied by adopting the welding gun follow-up ultrasonic impact system 8 with an ultrasonic impact needle 81, accurate and efficient intervention on a formed part is realized, and residual stress and a welding bead section form function of a specific area of the electric arc additive manufacturing workpiece are obtained by combining data analysis through various detection means, wherein the welding bead section form function comprises information of surface flatness, welding bead fusion width, welding bead height and the like of the workpiece; in this embodiment, the weld bead cross section shape function is a quantitative sine function relation between the weld bead fusion width (independent variable) and the weld bead residual height (dependent variable) with a simple form and high precision, which is obtained by obtaining a digitized image of the weld bead cross section shape through three-dimensional laser scanning and the like, then simplifying and smoothing the weld bead data, and then performing fitting and fitting error and mean square error analysis by using different functions based on a least square method.
(b) On the premise of ensuring the reliability and effectiveness of the test result, repeating the step (a), determining the test scheme of the low-stress high-precision 2195 aluminum-lithium alloy electric arc additive manufacturing forming process parameter control method according to the orthogonal test design principle based on the process parameter constraint range, and inputting data which are respectively welding voltage U e Welding current I e Frequency of current f e Height h of tungsten electrode w Scanning velocity v w Weld bead spacing d s Wire feed rate v f Substrate preheating temperature T b And safeguard forGuard gas flow Q g Laser power P l Spot diameter d l Defocus amount d d Magnetic induction B m Frequency f of magnetic field m N number of ultrasonic impact pins u Diameter d of ultrasonic impact needle u Ultrasonic impact amplitude A u Ultrasonic impact frequency f u (ii) a The output data is the residual stress of a certain specific area i of the 2195 aluminum lithium alloy electric arc additive manufacturing shaped piece
Surface flatness χ i Weld bead width W i Weld bead height H i (ii) a Reorganizing input data by utilizing class balance and inter-class intersection, performing primary and secondary factor division by utilizing a display rule of the data or an internal rule implied by the mined data through a grey correlation degree analysis method to obtain an input data weight, integrating the input data into an input quantity, determining a hidden layer excitation function according to a mixed kernel function, selecting a Morlet wavelet function, selecting a Sigomid function as an output layer excitation function, and finally establishing a prediction model of residual stress, surface flatness, weld bead fusion width and weld bead residual height of a 2195 aluminum lithium alloy electric arc additive manufacturing forming part by utilizing a conjugate gradient wavelet neural network based on local learning;
(c) Recording the maximum value sigma of the residual stress in each selected area of the 2195 aluminum lithium alloy arc additive manufactured part in all the test results in the step (b) RSmax And minimum value σ RSmin Surface flatness maximum χ max And minimum value χ min Maximum value of weld bead fusion width W max And minimum value W min Maximum value of weld bead height H max And a minimum value H min ；
(d) An optimal process scheme combination is sought by utilizing an improved gray wolf optimization algorithm, effective control on a low-stress high-precision 2195 aluminum lithium alloy electric arc additive manufacturing forming process is achieved, and a unified target evaluation function F (X) of each index (namely output data) is established as a fitness function according to a formula (1); wherein λ is 1 ＝λ 2 ＝λ 3 ＝λ 4 ＝0.25，η 1 ＝η 2 ＝η 3 ＝η 4 =0.5, and λ 1 +λ 2 +λ 3 +λ 4 ＝1；
(e) Initializing, determining the number m =90 of the wolf individuals in the search population, and randomly generating the position X of the 90 wolf in the search domain determined by the process parameter constraint condition in the step (b) j (k) (j =1,2, \8230;, 90), the maximum number of iterations k max =900 as an optimization termination condition, and let the current iteration number be k =0;
(f) Calculating the fitness value of each head of the population by using a formula (1), sorting according to the fitness value, comparing, and selecting the first three best grey wolf positions as an optimal position, a suboptimal position and a candidate position corresponding to alpha, beta and delta wolfs
(g) Updating the 90-head wolf body position X by using the formula (2) j (k+1)(j＝1,2,…,90)；
(h) Calculating the fitness value of all the grey wolves by using the formula (1) again to evaluate the whole grey wolves population;
(i) Judging whether the constraint condition in the formula (3) is satisfied, if so, directly switching to the step (j), and if not, switching to the step (g);
(j) Judging whether an iteration condition k of the algorithm is less than 900, if so, making k = k +1, turning to the step (g), continuing to perform algorithm iteration, and otherwise, directly turning to the step (l);
(k) Updating the fitness value and the position of the global first three best alpha, beta and delta wolfs;
(l) Outputting the current alpha wolf optimal position, and stopping iteration, namely finding the global optimal solution by the algorithm
And finally, obtaining the optimal control method of the low-stress high-precision 2195 aluminum-lithium alloy arc additive manufacturing forming process parameters, which mainly comprises the following parameters: welding voltage U ey =19V, welding current I ey =149A, current frequency f ey =180Hz, tungsten height h wy =2mm, scanning speedRate v wy =800mm/min and weld bead spacing d sy =0.8cm, wire feed rate v fy =1200mm/min, base material preheating temperature T by =200 ℃ protective gas flow Q gy =11L/min, laser power P ly =350W, spot diameter d ly =0.07cm, defocus amount d dy = -5cm, magnetic induction B my =90mT, magnetic field frequency f my =50Hz, number of ultrasonic impact pins n uy =7, ultrasonic impact pin diameter d uy =3mm, ultrasonic impact amplitude A uy =29 μm, ultrasonic frequency f uy ＝19.8kHz。
It should be noted that the optimized process parameters of the invention are welding voltage, welding current, current frequency, tungsten electrode height, scanning rate, welding bead spacing, wire feeding rate, substrate preheating temperature, shielding gas flow, laser power, spot diameter, defocusing amount, magnetic induction intensity, magnetic field frequency, ultrasonic impact needle number, ultrasonic impact needle diameter, ultrasonic impact amplitude and ultrasonic frequency; the output data is the residual stress, the surface flatness, the welding bead fusion width and the welding bead residual height of a certain specific area of the electric arc additive manufacturing workpiece, and in order to quickly and effectively realize the parameter control method of the low-stress high-precision electric arc additive manufacturing process, certain process parameters or certain process parameters can be properly added or deleted according to actual requirements, and only needs to be slightly improved on the basis of the optimization model provided by the invention.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (8)
1. The low-stress high-precision electric arc additive process control method based on computational intelligence is characterized by comprising the following steps of: the method comprises the following steps:
(a) Preparing a workpiece by utilizing an electric arc additive manufacturing process under the synergistic action of laser, electromagnetism and ultrasound, and acquiring residual stress and a welding bead section shape function of a specific area of the electric arc additive manufacturing workpiece by combining various detection means and data analysis, wherein the welding bead section shape function comprises information such as surface flatness, welding bead fusion width and welding bead residual height of the workpiece;
(b) On the premise of ensuring the reliability and effectiveness of the test result, repeating the step (a), determining the test scheme of the low-stress high-precision arc additive manufacturing process parameter control method according to the orthogonal test design principle based on the process parameter constraint range, and inputting data which are respectively welding voltage U e Welding current I e Frequency f of current e Height h of tungsten electrode w Scanning speed v w Weld bead spacing d s Wire feed rate v f Substrate preheating temperature T b Protective gas flow Q g Laser power P l Diameter d of light spot l Defocus amount d d Magnetic induction B m Frequency f of magnetic field m The number n of the ultrasonic impact pins u Diameter d of ultrasonic impact needle u Ultrasonic amplitude A u Ultrasonic frequency f u (ii) a The output data is the residual stress of a certain specific area i of the electric arc additive manufacturing workpiece
Surface flatness χ i Weld bead width W i Weld bead height H i (ii) a Reorganizing input data by utilizing class balance and inter-class crossing, performing primary and secondary factor division by utilizing a display rule of the data or an inherent rule of mining data implication through a grey correlation degree analysis method to obtain an input data weight, integrating the input data into an input quantity, determining an implication layer excitation function and an output layer excitation function, and finally establishing a prediction model of residual stress, surface flatness, welding bead fusion width and welding bead residual height of an electric arc additive manufacturing workpiece by utilizing a conjugate gradient wavelet neural network based on local learning;
(c) Recording the maximum value sigma of the residual stress in each selected area of the workpiece manufactured by the electric arc additive manufacturing in all the test results in the step (b) RSmax And minimum value σ RSmin Surface flatness maximum χ max And minimum value χ min Maximum value of weld bead fusion width W max And a minimum value W min Maximum value of weld bead height H max And a minimum value H min ；
(d) An optimal process scheme combination is sought by utilizing an improved wolf optimization algorithm, effective control on a low-stress high-precision electric arc additive manufacturing process is realized, a unified target evaluation function F (X) of each index (namely output data) is established as a fitness function to evaluate the quality of a solution corresponding to a variable, the smaller the value of the unified target evaluation function F (X) is, the better the solution corresponding to the variable is, and the expression of the fitness function is as follows:
wherein X is a design variable, X = [ U ] e ,I e ,f e ,h w ,v w ,d s ,v f ,T b ,Q g ,P l ,d l ,d d ,B m ,f m ,n u ,d u ,A u ,f u ]；
λ 1 ,λ 2 ,λ 3 ,λ 4 ,η 1 ,η 2 ,η 3 ,η 4 -a weighting factor, which ranges from 0 to 1, the value of each value being adjustable within the range (0, 1) according to the different requirements for each index parameter, wherein λ 1 +λ 2 +λ 3 +λ 4 ＝1；
n is the total number of the specific areas of the workpiece;
W i* ,H i* target values of weld bead fusion width and weld bead residual height of a specific area i of the workpiece;
in the formula (1), G is 1 (X)、G 2 (X)、G 3 (X)、G 4 The first term in (X) is used for evaluationProximity of the index to a target value, wherein G 1 (X)、G 2 The target values of residual stress and roughness in (X) are zero, and the second term is used for evaluating the uniformity of the index;
(e) Initializing, determining the number m of the wolf individuals in the search population, and randomly generating the position X of the m wolfs in the search domain determined by the process parameter constraint conditions in the step (b) j (k) (j =1,2, \ 8230;, m), the maximum number of iterations k max As an optimization termination condition, and making the current iteration number k =0;
(f) Calculating the fitness value of each head of the wolf of the population by using a formula (1), sorting according to the fitness value, comparing, and selecting the first three best wolf positions as an optimal position, a suboptimal position and a candidate position corresponding to alpha, beta and delta wolfs
(g) Updating the m-head grey wolf body position X by using a formula (2) j (k+1)(j＝1,2,…,m)：
In the formula, phi 1 ,φ 2 ,φ 3 -a weighting factor;
w α ,w β ,w δ -an inertia weight;
μ α ,μ β ,μ δ -a non-linear modulation index, the value range being 0 to 3;
rand α (0,1),rand β (0,1),rand δ (0,1),rand′ α (0,1),rand′ β (0,1),rand′ δ (0,1)—[0,1]a random number in between;
(h) Calculating the fitness value of all the grey wolves by using the formula (1) again to evaluate the whole grey wolves population;
(i) Judging whether the constraint condition in the formula (3) is satisfied, if so, directly switching to the step (j), and if not, switching to the step (g);
(j) Updating the fitness values and positions of the first three global best alpha, beta and delta wolves;
(k) Judging whether an iteration condition k of the algorithm is satisfied, wherein k is more than k max If yes, enabling k = k +1, turning to the step (g), continuing to perform algorithm iteration, and otherwise, directly turning to the step (l);
(l) Outputting the current alpha wolf optimal position, stopping iteration, namely finding the global optimal solution by the algorithm
And finally, obtaining the optimal control method of the parameters of the low-stress high-precision arc additive manufacturing process.
2. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: the electric arc additive manufacturing process under the laser-electromagnetic-ultrasonic synergistic effect in the step (a) is that on the basis of a conventional electric arc additive manufacturing process, the laser assistance effect is increased to improve the cladding depth and the cladding speed, the electromagnetic assistance effect is increased to provide support for a molten pool of a welding channel, the shape and the characteristics of the molten pool are regulated and controlled, the workpiece forming precision is improved, the ultrasonic impact treatment is increased to reduce the residual stress and improve the mechanical property, and the low-stress high-precision rapid forming technology of the electric arc additive manufacturing process is realized under the laser-electromagnetic-ultrasonic synergistic effect.
3. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: and (b) obtaining a digital image of the cross section form of the welding bead by a scanning electron microscope or three-dimensional laser scanning and the like for the shape function of the cross section of the welding bead in the step (a), then simplifying and smoothing the data of the welding bead, then performing fitting and fitting error and mean square error analysis by adopting different functions based on a least square method, and finally obtaining a quantitative function relation between the weld bead fusion width (independent variable) and the weld bead residual height (dependent variable) with simple form and higher precision.
4. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: the process parameter constraints recited in step (b) refer to the maximum and minimum values of the following parameters: welding voltage U e Welding current I e Frequency of current f e Height h of tungsten electrode w Scanning velocity v w Weld bead spacing d s Wire feed rate v f Substrate preheating temperature T b And protective gas flow Q g Laser power P l Diameter d of light spot l Defocus amount d d Magnetic induction B m Frequency f of magnetic field m The number n of the ultrasonic impact pins u Diameter d of ultrasonic impact needle u Ultrasonic impact amplitude A u Ultrasonic impact frequency f u In particular, if the ultrasonic vibration is applied using a forming substrate bottom-fixed type apparatus, the number n of ultrasonic impact pins is u ＝1。
5. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: surface flatness χ of step (b) i The calculation formula of (2) is as follows:
in the formula (I), the compound is shown in the specification,
-the height of the s peak at the surface of the workpiece cladding layer;
n-the number of the wave crests and the wave troughs within a certain length range;
note that the surface flatness χ i The value can fully reflect the change characteristic of the height of the geometric parameter of the surface of the workpiece, and within a certain length range, the smaller the value is, the flatter the cladding forming plane is, and the better the surface quality of the workpiece is.
6. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: selecting a Morlet wavelet function as the hidden layer excitation function in the step (b), and selecting a Sigomiod function as the output layer excitation function.
7. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: the value range of the number m of the wolfsbane individuals in the search population in the step (e) is 20-100.
8. The method for controlling the low-stress high-precision arc additive process based on the computational intelligence of claim 1, wherein the method comprises the following steps: said maximum number of iterations k for step (e) max The value range is 200-1000.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211395464.8A CN115685881B (en) | 2022-11-07 | 2022-11-07 | Low-stress high-precision arc additive process control method based on computational intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211395464.8A CN115685881B (en) | 2022-11-07 | 2022-11-07 | Low-stress high-precision arc additive process control method based on computational intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115685881A true CN115685881A (en) | 2023-02-03 |
| CN115685881B CN115685881B (en) | 2023-06-02 |
Family
ID=85050250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211395464.8A Active CN115685881B (en) | 2022-11-07 | 2022-11-07 | Low-stress high-precision arc additive process control method based on computational intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115685881B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116702631A (en) * | 2023-08-08 | 2023-09-05 | 四川大学 | Electron beam additive manufacturing constitutive relation calculation method based on artificial neural network |
| CN117428291A (en) * | 2023-12-18 | 2024-01-23 | 南京理工大学 | Weld bead fusion width quantification method based on sonogram characteristic analysis |
| CN118492415A (en) * | 2024-07-17 | 2024-08-16 | 南通理工学院 | Multi-axis linkage control method and system for synchronous composite additive manufacturing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106903448A (en) * | 2016-12-26 | 2017-06-30 | 南京航空航天大学 | A kind of electric arc, laser, magnetic field multi-energy a coupling method of manufacturing technology |
| CN108856974A (en) * | 2018-07-26 | 2018-11-23 | 重庆科技学院 | A kind of ultrasonic field coupling congruent melting pond mariages CMT electric arc increasing material moulding technique |
| CN112379589A (en) * | 2020-10-13 | 2021-02-19 | 重庆大学 | Worm wheel machining shape controllable process |
| WO2021128510A1 (en) * | 2019-12-27 | 2021-07-01 | 江苏科技大学 | Bearing defect identification method based on sdae and improved gwo-svm |
| CN113569352A (en) * | 2021-07-13 | 2021-10-29 | 华中科技大学 | Additive manufacturing size prediction and process optimization method and system based on machine learning |
| CN114818435A (en) * | 2022-05-20 | 2022-07-29 | 中国矿业大学 | Multi-target cooperative regulation and control method for electric arc additive manufacturing |
| CN114799415A (en) * | 2022-03-11 | 2022-07-29 | 南京航空航天大学 | Arc additive remanufacturing welding parameter-welding bead size positive and negative neural network prediction model |
| WO2022166414A1 (en) * | 2021-02-08 | 2022-08-11 | 深圳先进技术研究院 | Shape-property integrated forming device system for plasma wire arc additive manufacturing |
-
2022
- 2022-11-07 CN CN202211395464.8A patent/CN115685881B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106903448A (en) * | 2016-12-26 | 2017-06-30 | 南京航空航天大学 | A kind of electric arc, laser, magnetic field multi-energy a coupling method of manufacturing technology |
| CN108856974A (en) * | 2018-07-26 | 2018-11-23 | 重庆科技学院 | A kind of ultrasonic field coupling congruent melting pond mariages CMT electric arc increasing material moulding technique |
| WO2021128510A1 (en) * | 2019-12-27 | 2021-07-01 | 江苏科技大学 | Bearing defect identification method based on sdae and improved gwo-svm |
| CN112379589A (en) * | 2020-10-13 | 2021-02-19 | 重庆大学 | Worm wheel machining shape controllable process |
| WO2022166414A1 (en) * | 2021-02-08 | 2022-08-11 | 深圳先进技术研究院 | Shape-property integrated forming device system for plasma wire arc additive manufacturing |
| CN113569352A (en) * | 2021-07-13 | 2021-10-29 | 华中科技大学 | Additive manufacturing size prediction and process optimization method and system based on machine learning |
| CN114799415A (en) * | 2022-03-11 | 2022-07-29 | 南京航空航天大学 | Arc additive remanufacturing welding parameter-welding bead size positive and negative neural network prediction model |
| CN114818435A (en) * | 2022-05-20 | 2022-07-29 | 中国矿业大学 | Multi-target cooperative regulation and control method for electric arc additive manufacturing |
Non-Patent Citations (2)
| Title |
|---|
| 傅蔚阳;刘以安;薛松;: "基于灰狼算法与小波神经网络的目标威胁评估", 浙江大学学报(工学版), no. 04 * |
| 范剑超等: "工艺参数对电弧增材制造成形质量的影响", 《山东建筑大学学报》, vol. 37, no. 5 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116702631A (en) * | 2023-08-08 | 2023-09-05 | 四川大学 | Electron beam additive manufacturing constitutive relation calculation method based on artificial neural network |
| CN116702631B (en) * | 2023-08-08 | 2023-10-27 | 四川大学 | Electron beam additive manufacturing constitutive relation calculation method based on artificial neural network |
| CN117428291A (en) * | 2023-12-18 | 2024-01-23 | 南京理工大学 | Weld bead fusion width quantification method based on sonogram characteristic analysis |
| CN118492415A (en) * | 2024-07-17 | 2024-08-16 | 南通理工学院 | Multi-axis linkage control method and system for synchronous composite additive manufacturing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115685881B (en) | 2023-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115685881A (en) | Low-stress high-precision electric arc additive process control method based on computational intelligence | |
| Shin et al. | Predictive modeling capabilities from incident powder and laser to mechanical properties for laser directed energy deposition | |
| Kaiming et al. | A study of laser cladding NiCrBSi/Mo composite coatings | |
| Caiazzo et al. | Additive manufacturing by means of laser-aided directed metal deposition of 2024 aluminium powder: Investigation and optimization | |
| Chatterjee et al. | Drilling of micro-holes on titanium alloy using pulsed Nd: YAG laser: parametric appraisal and prediction of performance characteristics | |
| Das et al. | Phenomenological model-based study on electron beam welding process, and input-output modeling using neural networks trained by back-propagation algorithm, genetic algorithms, particle swarm optimization algorithm and bat algorithm | |
| Malik et al. | Multi-response optimization of laser-assisted jet electrochemical machining parameters based on gray relational analysis | |
| Ai et al. | Process modeling and parameter optimization using radial basis function neural network and genetic algorithm for laser welding of dissimilar materials | |
| Zavdoveev et al. | Optimization of the pulsed arc welding parameters for wire arc additive manufacturing in austenitic steel applications | |
| Chen et al. | Investigation on the process and microstructure evolution during direct laser metal deposition of 18Ni300 | |
| Du et al. | Multi-parameter optimization of laser cladding 15-5PH Using TOPSIS-GRA based on combined weighting method | |
| Wahid et al. | A simulation-based study on the effect of underwater friction stir welding process parameters using different evolutionary optimization algorithms | |
| Hirohata et al. | An investigation on laser-arc hybrid welding of one-pass full-penetration butt-joints for steel bridge members | |
| Fatoba et al. | Computational dynamics of anti-corrosion performance of laser alloyed metallic materials | |
| Li et al. | Research on heat source model and weld profile for fiber laser welding of A304 stainless steel thin sheet | |
| Neelamegam et al. | Hybrid intelligent modeling for optimizing welding process parameters for reduced activation ferritic-martensitic (RAFM) steel | |
| Das et al. | Input–output modeling and multi-objective optimization of weld attributes in EBW | |
| Gao et al. | Multi-objective optimization of process parameters for laser metal deposition of NiTi shape memory alloy based on neural network and genetic algorithm | |
| Mahmood et al. | On the development of smart framework for printability maps in additive manufacturing of AISI 316L stainless steel | |
| Gu et al. | Microstructure, mechanical properties, and residual stress distribution of AISI 316L stainless steel part fabricated by laser metal deposition | |
| Rajulu et al. | An integrated evolutionary approach for simultaneous optimization of laser weld bead characteristics | |
| Qazani et al. | A machine learning method for cutting parameter selection in rotary ultrasonic-assisted end grinding | |
| Chandran et al. | Laser Beam Welded Aluminum-Titanium Dissimilar Sheet Metals: Neural Network Based Strength and Hardness Prediction Model | |
| Skariya et al. | Optimizing parameters of TIG welding process using grey wolf optimization concerning 15CDV6 steel | |
| Lim et al. | Process window prediction in stainless steel selective laser melting using various energy densities: laser power, scan speed, and defocusing distance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |