CN111859566B - Digital twinning-based surface roughness stabilization method - Google Patents
Digital twinning-based surface roughness stabilization method Download PDFInfo
- Publication number
- CN111859566B CN111859566B CN202010692248.4A CN202010692248A CN111859566B CN 111859566 B CN111859566 B CN 111859566B CN 202010692248 A CN202010692248 A CN 202010692248A CN 111859566 B CN111859566 B CN 111859566B
- Authority
- CN
- China
- Prior art keywords
- surface roughness
- parameters
- machining
- prediction model
- processing
- 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.)
- Active
Links
- 230000003746 surface roughness Effects 0.000 title claims abstract description 183
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000006641 stabilisation Effects 0.000 title claims abstract description 15
- 238000011105 stabilization Methods 0.000 title claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 76
- 238000003754 machining Methods 0.000 claims abstract description 66
- 238000011478 gradient descent method Methods 0.000 claims abstract description 8
- 238000013507 mapping Methods 0.000 claims abstract description 6
- 230000001351 cycling effect Effects 0.000 claims abstract description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 238000005520 cutting process Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 11
- 241000272201 Columbiformes Species 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 238000012706 support-vector machine Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 9
- 230000026676 system process Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 30
- 238000002474 experimental method Methods 0.000 description 9
- 239000000463 material Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000000087 stabilizing effect Effects 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000007423 decrease Effects 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000003050 experimental design method Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000000227 grinding Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 101100483039 Arabidopsis thaliana TT8 gene Proteins 0.000 description 1
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 208000013586 Complex regional pain syndrome type 1 Diseases 0.000 description 1
- 101100136092 Drosophila melanogaster peng gene Proteins 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 229910001069 Ti alloy Inorganic materials 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 244000144992 flock Species 0.000 description 1
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000000875 high-speed ball milling Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000005461 lubrication Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention discloses a digital twinning-based surface roughness stabilization method, which comprises the following steps: 1) A digital twin system of a virtual world is established based on a mechanical processing system of a physical world, and a surface roughness prediction model is established in the digital twin system; 2) Mapping a machining system by using a digital twin system, collecting machining parameters influencing surface roughness in the machining system in real time, and inputting the machining parameters into the digital twin system; 3) Predicting the surface roughness under the current processing condition by using a surface roughness prediction model; if the surface roughness is predictedIf the machining parameters are within the set threshold range, the surface roughness is stable, the current machining parameters are proved to meet the requirements, and the step 5) is executed; otherwise, executing the step 4); 4) Solving by gradient descent methodProcessing parameters within the set threshold range are fed back to the machining system, and step 5) is executed; 5) And (3) cycling the step 2) and the step 3) until the workpiece processing is completed.
Description
Technical Field
The invention belongs to the technical field of machining surface quality control, and particularly relates to a digital twinning-based surface roughness stabilizing method.
Background
Surface quality generally directly affects the physical, chemical, and mechanical properties of the workpiece, such as friction properties, fatigue properties, wear resistance, lubrication properties, and the like. The surface roughness is used as the most important index for evaluating the surface quality, and is therefore selected as a key technical requirement for workpiece production. In actual machining, surface roughness is often unstable and often tends to increase due to factors such as tool vibration, tool wear, and plastic deformation of the workpiece material. Therefore, in order to obtain better surface properties, an effective surface roughness stabilization method is required. In recent years, scholars at home and abroad have conducted a great deal of research on the establishment of surface roughness prediction models, and can be roughly divided into three categories: theoretical methods, experimental design methods, and artificial intelligence methods.
In the theoretical method, a surface roughness prediction model (generally, a mathematical equation of a machined surface) is established based on a machining theory, taking into consideration factors such as a tool shape, workpiece material characteristics, installation errors, machining dynamics and the like. For any workpiece and tool combination, munoz-Escalona and Maropoulos propose surface roughness prediction models based on tool path geometry analysis. Also, lu et al built a tool flexible deformation model based on cutting force, and used the model to build a surface topography simulation model that predicts surface roughness. To reduce the non-uniformity of surface roughness, sun et al propose a surface roughness Relative Standard Deviation (RSDS) method based on relative tool sharpness to predict surface non-uniformity. Considering tool pose, chatter, runout, cutting force, and material deformation, peng et al build a theoretical model describing the trajectory of the cutting edge and predict surface roughness based thereon.
In the experimental design method, surface roughness models under different processing modes are established. Common experimental design methods include a field method, a full factor design method, a curved surface response method (RSM), and the like. Compared with other methods, the RSM method is generally applied to the prediction of surface roughness because only a small amount of experiments are required. Karkalos et Al studied the best processing parameters for Ti-6Al-4v titanium alloys at minimum surface roughness using RSM. Dikshit et al optimally select the minimum surface roughness parameters in high-speed ball milling by adopting a RSM-based center composite design method. The properties (mainly surface roughness and cutting forces) of cemented carbide tools were investigated by noondin et al by RSM. It was found that the feed rate was a major factor affecting the surface roughness. Mansource and Abdalla [13] also used RSM to build predictive models of surface roughness of EN32 materials.
In addition to the above two methods, the artificial intelligence method is also widely used for predicting surface roughness as a powerful prediction tool with self-learning and self-adaptation capabilities. In EN 24T steel turning under high pressure cooling conditions, mia et al propose an Artificial Neural Network (ANN) based average surface roughness prediction model. Ghosh et al propose a surface roughness prediction model based on an artificial neural network and search for the optimal cutting conditions by using a genetic algorithm and a particle swarm optimization algorithm.
The surface roughness predicted by the model is constant under the same processing conditions (including tools, workpieces, processing parameters and the like). However, in actual machining, the surface roughness of the workpiece is unstable due to the influence of factors such as tool wear, vibration, material property non-uniformity, stability of a process system, and the like. Therefore, a grinding surface roughness prediction method based on an improved support vector machine algorithm is proposed, which considers dynamic factors to accurately predict surface roughness, such as the one disclosed in chinese patent publication No. CN110348075a, however, the grinding surface roughness prediction method, although taking dynamic factors into consideration and predicting surface roughness under dynamic factors, cannot find one or more factors having the greatest influence on surface roughness under current machining conditions to achieve a technical purpose of stabilizing surface roughness by guiding a real machining scene, because many factors influencing surface roughness during machining are involved, when surface roughness in a real machining scene does not meet the requirements. In addition, the surface roughness generally fluctuates in a short time, which may make the workpiece to be processed unsatisfactory for practical use, thereby increasing manufacturing costs and processing time. Therefore, a real-time effective surface roughness control method is very important.
Disclosure of Invention
In view of the above, the present invention is directed to a digital twinning-based surface roughness stabilization method, which can adjust processing parameters affecting surface roughness on line when the surface roughness is unstable, thereby rapidly stabilizing the surface roughness and stabilizing the surface roughness within a set threshold range.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a digital twinning-based surface roughness stabilization method comprises the following steps:
1) A digital twin system of a virtual world is established based on a mechanical processing system of a physical world, and a surface roughness prediction model is established in the digital twin system;
2) Mapping a machining system by using a digital twin system, collecting machining parameters influencing surface roughness in the machining system in real time, and inputting the machining parameters into the digital twin system;
3) Predicting the surface roughness under the current processing condition by using a surface roughness prediction model; if the surface roughness is predictedIf the machining parameters are within the set threshold range, the surface roughness is stable, the current machining parameters are proved to meet the requirements, and the step 5) is executed; otherwise, executing the step 4);
4) Solving by gradient descent methodProcessing parameters within a set threshold range are fed back to a machining system, and the machining system processes a workpiece according to the processing parameters obtained by solving, and step 5) is executed;
5) And (3) cycling the step 2) and the step 3) until the workpiece processing is completed.
Further, in the step 1), the method for constructing the surface roughness prediction model is as follows:
21 Machining the workpiece by using a machining system to obtain the surface roughness under the conditions of different machining parameters;
22 Dividing the data obtained in the step 21) into a training set and a testing set, carrying out normalization treatment, and taking the normalized processing parameters and the surface roughness as an input sample and an output sample of a surface roughness prediction model respectively;
23 A surface roughness prediction model based on PIO-SVM is established.
Further, in the step 23), the method for establishing the surface roughness prediction model based on the PIO-SVM is as follows:
231 Initializing parameters of a pigeon swarm algorithm model, taking the mean square error corresponding to each group of punishment function C and kernel function parameter g as fitness function, and carrying out iterative search in the pigeon swarm algorithm to obtain C and g corresponding to the minimum mean square error, namely the optimal punishment function C and kernel function parameter g;
232 Substituting the optimal penalty function C and the kernel function parameter g obtained by solving into a support vector machine algorithm model, further constructing and obtaining a surface roughness prediction model, and checking the accuracy of the model by using a test set subjected to normalization processing.
In step 4), a point corresponding to the current processing condition is found in the surface roughness prediction model, a direction with the fastest surface roughness change is found in the surface roughness prediction model based on the point, a target point is taken in the direction with the fastest surface roughness change, the surface roughness of the target point is located within a set threshold range, and processing parameters corresponding to the target point are fed back to the machining system.
Further, the machining system is a five-axis machining system, and the machining parameters affecting the surface roughness include lead angle, tilt angle, cutting depth, spindle rotation speed, feed speed and average cutting force.
Further, processing parameters affecting the surface roughness are divided into on-line adjustable parameters and off-line adjustable parameters, wherein the on-line adjustable parameters comprise tool posture parameters and spindle rotating speed, and the tool posture parameters comprise lead angles and tilt angles; the non-on-line adjustable parameters include depth of cut, feed rate, and average cutting force.
In step 4), a point corresponding to the current processing condition is found in the surface roughness prediction model, a direction in which the surface roughness corresponding to the on-line adjustable parameter changes most rapidly is found in the surface roughness prediction model based on the point, a target point is taken in the direction in which the surface roughness changes most rapidly, the surface roughness at the target point is located within a set threshold range, and the processing parameter corresponding to the target point is fed back to the machining system.
Further, if the parameters can be adjusted on line for more than N times, the parameters still cannot be adjustedIf the processing parameters are within the set threshold range, the condition that only the gesture parameters of the tool and the rotating speed of the spindle cannot meet the current processing requirements is indicated, and at the moment, all the processing parameters affecting the surface roughness are taken as adjustment objects, wherein the adjustment method comprises the following steps: finding a point corresponding to the current processing condition in the surface roughness prediction model, finding the direction with the fastest surface roughness change in the surface roughness prediction model based on the point, taking a target point in the direction with the fastest surface roughness change, enabling the surface roughness of the target point to be located in a set threshold range, and feeding back processing parameters corresponding to the target point to a mechanical processing system, wherein N is a positive integer greater than or equal to 1.
The invention has the beneficial effects that:
according to the digital twin-based surface roughness stabilization method, the digital twin system is established to realize real-time mapping of the physical world machining system, so that the machining state of the machining system can be reflected in real time by the digital twin system; the surface roughness prediction model of the component in the digital twin system can accurately predict the surface roughness of the mechanical processing system of the physical world under the current processing condition based on dynamic factors, when the predicted surface roughness exceeds the range of a set threshold value, namely the surface roughness is unstable, the surface roughness prediction model can be solved based on a gradient descent method, so that new processing parameters can be rapidly acquired and fed back to the mechanical processing system of the physical world, and the surface roughness is kept in a stable state until the processing of a workpiece is completed; in summary, the surface roughness stabilization method based on digital twinning realizes the prediction of the surface roughness under the action of taking dynamic factors into consideration, and when the surface roughness is unstable, on-line adjustable parameters (such as tool posture parameters and spindle rotating speed) can be adjusted on line, so that the surface roughness is quickly stabilized and is stabilized within a set threshold range.
Drawings
In order to make the objects, technical solutions and advantageous effects of the present invention more clear, the present invention provides the following drawings for description:
FIG. 1 is a schematic diagram of the mapping and interaction of a machining system in the physical world and a digital twin system in the virtual world;
FIG. 2 is a flow chart of the construction of the PIO-SVM-based surface roughness prediction model in the present embodiment;
FIG. 3 is a flow chart for stabilizing surface roughness by adjusting process parameters using a gradient descent method;
FIG. 4 is a training set containing 40 sets of experimental data;
FIG. 5 is a test set containing 10 sets of experimental data;
fig. 6 is a simulation result of changing spindle rotation speed and tool pose to stabilize and control surface roughness.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and specific examples, which are not intended to limit the invention, so that those skilled in the art may better understand the invention and practice it.
The surface roughness stabilization method based on digital twinning in the embodiment comprises the following steps.
1) And a mechanical processing system based on the physical world establishes a digital twin system of the virtual world, and builds a surface roughness prediction model in the digital twin system. Specifically, the machining system of the present embodiment is a five-axis machining system, and the machining parameters affecting the surface roughness include lead angle (L), tilt angle (T), cutting depth (a p ) Spindle speed (n), feed speed (f) and average cutting force
21 Using a machining system to process the workpiece, and obtaining the surface roughness under different processing parameter conditions. The present example co-designed and conducted 50 sets of experiments including two sets of two-factor four-level full-factor experiments and two sets of three-factor three-level orthogonal experiments. For each set of experiments, the surface roughness R was measured by repeating twice a Substantially identical. Last measured surface roughness R a Is the average of the results of two experiments.
22 Dividing the data obtained in the step 21) into a training set and a testing set, carrying out normalization processing, and taking the normalized processing parameters and the surface roughness as an input sample and an output sample of the surface roughness prediction model respectively.
Specifically, in this embodiment, 40 groups of experiments are randomly selected as training sets, see fig. 4; the remaining 10 groups of experiments served as test sets, see fig. 5. After normalization processing, the training time can be shortened, the prediction precision of the regression model is improved, and normalization processing is required to be carried out on the training sample set. The training set is normalized to [0,1] herein. The calculation formula is as follows:
wherein x 'is' i And x i The sample data before and after normalization are respectively,and->Is the maximum and minimum of the surface roughness influence factor i.
23 A surface roughness prediction model based on PIO-SVM is established. Specifically, the method for establishing the PIO-SVM-based surface roughness prediction model comprises the following steps:
231 Initializing parameters of a pigeon swarm algorithm model, taking the mean square error corresponding to each group of punishment function C and kernel function parameter g as fitness function, and carrying out iterative search in the pigeon swarm algorithm to obtain C and g corresponding to the minimum mean square error, namely the optimal punishment function C and kernel function parameter g. The training set is used for training the surface roughness prediction model under different C, g, then the test set is used for checking the accuracy of the prediction model, and C, g corresponding to the minimum mean square error, namely the optimal C, g is found. The selection of the penalty function C and the kernel function parameter g has an important influence on the accuracy of the regression prediction model, and compared with a PSO algorithm and a GA algorithm, the PIO (pigeon-cluster algorithm model) has higher convergence speed, higher accuracy and higher stability. Thus, the present embodiment uses PIO to select the best C and g.
The PIO algorithm model consists of two parts, namely a map, a compass operator and a landmark operator. The mathematical model of each part is described as follows:
V m (k)＝V m (k-1)·e -fk +rand·(L b -L m (k-1))
L m (k)＝L m (k-1)+V m (k)
L m (k)＝L m (k-1)+rand·(L C (k)-L m (k-1))
wherein L is m And V m The position and speed of the mth pigeon are respectively, k is the iteration number, f is a map and compass operator, and rand is [0,1]]Random number in L b Represents the global optimal position of the iteration, L C (k) Representing the center position of the remaining pigeons, N t,1 Representing the number of iterations of map and compass operators, N t,2 Representing the number of iterations of landmark operators, fitness is the quality function of each solution, here the mean square error of the predictive model for each penalty function C and kernel parameter gDifference, N g Representing the number of pigeon flocks.
232 Substituting the optimal penalty function C and the kernel function parameter g obtained by solving into a support vector machine algorithm model, further constructing and obtaining a surface roughness prediction model, and checking the accuracy of the model by using a test set subjected to normalization processing.
The Support Vector Machine (SVM) is a statistical learning method based on the principle of structural risk minimization proposed by Vapnik. Compared with the traditional machine learning method represented by the neural network, the support vector machine has obvious advantages in the aspects of theoretical basis, training process, node number, weight vector, global optimal solution and the like. The basic idea of the support vector machine is to reduce the search for the optimal linear hyperplane to a convex programming problem. The sample space is non-linearly mapped to a high-dimensional or infinite-dimensional feature space. Thus, the linear learning machine can be used to solve the problem of non-linearity (including classification and regression) in the high-dimensional feature space.
The regression function of the vector machine algorithm model of this embodiment is:
wherein f (x) is a regression function, α i Andis the Lagrangian multiplier, C is the penalty function, K (x i ,x j ) Is a kernel function omega i As normal vector, x i ,x j Is at will->And->Epsilon is the insensitivity loss factor, b is the bias, y i The function value corresponds to the support vector.
The kernel function is an important component for establishing the regression prediction model of the support vector machine, and reasonable selection of the kernel function is helpful for improving the accuracy of the prediction model. The function of the kernel functions is to transform the linear inseparable problem in the low-dimensional space into the linear inseparable and linear regression problem in the high-dimensional space. The Radial Basis Function (RBF) has the advantages of relatively simple calculation form, few input parameters, strong learning capacity and the like. Accordingly, the selection of RBF as a kernel function of the regression prediction model can be described as follows:
K(x i ,x j )＝exp{-g|x i -x j | 2 },(g＞0)
where g is a kernel function parameter.
As can be seen from fig. 5, to verify the validity of the proposed model, the prediction error for each set of experiments was calculated using the following formula:
wherein,representing the surface roughness calculated by the predictive model, R a Indicating the measured surface roughness. It can be seen that predicted +.>And R in actual measurement a Substantially identical, the Average Prediction Error (APE) was only 8.69%.
2) And mapping the machining system by using a digital twin system, collecting machining parameters influencing the surface roughness in the machining system in real time, and inputting the machining parameters into the digital twin system.
3) Predicting the surface roughness under the current processing condition by using a surface roughness prediction model; if the surface roughness is predictedIf the machining parameters are within the set threshold range, the surface roughness is stable, the current machining parameters are proved to meet the requirements, and the step 5) is executed; otherwise, step 4) is performed.
4) Solving by gradient descent methodAnd (5) processing parameters within the set threshold range are fed back to a machining system, and the machining system processes the workpiece according to the processing parameters obtained by solving, so that step 5) is executed. Specifically, a general solution method of the gradient descent method is as follows: finding a point corresponding to the current processing condition in the surface roughness prediction model, finding the direction with the fastest surface roughness change in the surface roughness prediction model based on the point, taking a target point in the direction with the fastest surface roughness change, enabling the surface roughness of the target point to be located in a set threshold range, and feeding back processing parameters corresponding to the target point to a mechanical processing system.
The threshold range of the present embodiment is that the surface roughness is less than or equal toI.e. surface roughness of less than or equal to->Then is in a stable state, the surface roughness is greater than +.>Is in an unstable state. Of course, according to the actual processing requirement, the threshold range of the surface roughness can have a maximum value and a minimum value, and the surface roughness is in a stable state when being between the maximum value and the minimum value; the threshold range of the surface roughness may be a minimum value or more, and when the surface roughness is more than the minimum value, the surface roughness is in a stable state, and will not be described again.
The surface roughness of the workpiece is also unstable due to the instability of the cutting force. Thus, at a given processing parameter (resulting from the processing experience of a worker), the surface roughness may not meet the process requirements. According to the prediction model, the change trend of the surface roughness can be monitored on line, and the stability of the surface roughness can be judged. At the same time, considerThe feasibility of machine tool operation can online adjust the spindle rotation speed and tool posture in the machining process. Therefore, the surface roughness can be stabilized and controlled by adjusting the spindle rotation speed and the tool posture. To verify the effectiveness of this method, a set of simulation experiments were performed. The simulation results are shown in fig. 6. According to the experimental results, by adjusting the adjustable parameters (tool pose and spindle rotation speed), the surface roughness was significantly reduced from 0.4882 μm to 0.2995 μm. At the same time, a satisfactory surface roughness can be achieved for a variety of adjustable parameter combinations, which means that an efficient and simple selection of the optimal parameter combination is required. Therefore, the processing parameters affecting the surface roughness are divided into on-line adjustable parameters and off-line adjustable parameters, wherein the on-line adjustable parameters comprise tool posture parameters and spindle rotation speed, and the tool posture parameters comprise lead angles and tilt angles; the non-online adjustable parameters include depth of cut, feed rate, and average cutting force. At this time, the gradient descent method is solved as follows: finding a point corresponding to the current processing condition in the surface roughness prediction model, finding the direction of the fastest surface roughness decrease corresponding to the online adjustable parameter in the surface roughness prediction model based on the point, and taking a target point in the direction of the fastest surface roughness decrease to ensure that the surface roughness at the target point is less than or equal to the set maximum surface roughness valueAnd feeding back the processing parameters corresponding to the target point to the machining system. If the continuous adjustment can be carried out on the online adjustment parameters for more than N times, the parameters still cannot be met +.>The method for adjusting the tool attitude parameters and the spindle rotation speed only can not meet the current processing requirements, and all the processing parameters affecting the surface roughness are required to be used as adjustment objects, and the method for adjusting the tool attitude parameters and the spindle rotation speed comprises the following steps: finding a point corresponding to the current processing condition in the surface roughness prediction model, and finding the direction in which the surface roughness decreases most rapidly in the surface roughness prediction model based on the point, wherein the direction is the direction in which the surface roughness decreases most rapidlyTaking a target point to make the surface roughness of the target point less than or equal to the set maximum value of the surface roughnessAnd feeding back the processing parameter corresponding to the target point to the machining system, wherein N is a positive integer greater than or equal to 1, and n=4 in this embodiment.
5) And (3) cycling the step 2) and the step 3) until the workpiece processing is completed.
The above-described embodiments are merely preferred embodiments for fully explaining the present invention, and the scope of the present invention is not limited thereto. Equivalent substitutions and modifications will occur to those skilled in the art based on the present invention, and are intended to be within the scope of the present invention. The protection scope of the invention is subject to the claims.
Claims (6)
1. A digital twinning-based surface roughness stabilization method is characterized in that: the method comprises the following steps:
1) A digital twin system of a virtual world is established based on a mechanical processing system of a physical world, and a surface roughness prediction model is established in the digital twin system;
the method for constructing the surface roughness prediction model comprises the following steps:
21 Machining the workpiece by using a machining system to obtain the surface roughness under the conditions of different machining parameters;
22 Dividing the data obtained in the step 21) into a training set and a testing set, carrying out normalization treatment, and taking the normalized processing parameters and the surface roughness as an input sample and an output sample of a surface roughness prediction model respectively;
23 Establishing a surface roughness prediction model based on a PIO-SVM;
the method for establishing the surface roughness prediction model based on the PIO-SVM comprises the following steps:
231 Initializing parameters of a pigeon swarm algorithm model, taking the mean square error corresponding to each group of punishment function C and kernel function parameter g as fitness function, and carrying out iterative search in the pigeon swarm algorithm to obtain punishment function C and kernel function parameter g corresponding to the minimum mean square error, thereby obtaining the optimal punishment function C and kernel function parameter g;
232 Substituting the optimal penalty function C and the kernel function parameter g obtained by solving into a support vector machine algorithm model, further constructing a surface roughness prediction model, and checking the accuracy of the model by using a test set subjected to normalization processing;
2) Mapping a machining system by using a digital twin system, collecting machining parameters influencing surface roughness in the machining system in real time, and inputting the machining parameters into the digital twin system;
3) Predicting the surface roughness under the current processing condition by using a surface roughness prediction model; if the surface roughness is predictedIf the machining parameters are within the set threshold range, the surface roughness is stable, the current machining parameters are proved to meet the requirements, and the step 5) is executed; otherwise, executing the step 4);
4) Solving by gradient descent methodProcessing parameters within a set threshold range are fed back to a machining system, and the machining system processes a workpiece according to the processing parameters obtained by solving, and step 5) is executed;
5) And (3) cycling the step 2) and the step 3) until the workpiece processing is completed.
2. The digital twinning-based surface roughness stabilization method of claim 1, wherein: in the step 4), a point corresponding to the current processing condition is found in the surface roughness prediction model, a direction with the fastest surface roughness change is found in the surface roughness prediction model based on the point, a target point is taken in the direction with the fastest surface roughness change, the surface roughness of the target point is located within a set threshold range, and processing parameters corresponding to the target point are fed back to the mechanical processing system.
3. The digital twinning-based surface roughness stabilization method of claim 1 or 2, wherein: the machining system is a five-axis machining system, and machining parameters affecting surface roughness include lead angle, tilt angle, cutting depth, spindle rotation speed, feed speed and average cutting force.
4. A digital twinning-based surface roughness stabilization method as claimed in claim 3, wherein: based on the feasibility of machine tool operation, processing parameters affecting surface roughness are divided into on-line adjustable parameters and non-on-line adjustable parameters, wherein the on-line adjustable parameters comprise tool posture parameters and spindle rotating speed, and the tool posture parameters comprise lead angles and tilt angles; the non-on-line adjustable parameters include depth of cut, feed rate, and average cutting force.
5. The digital twinning-based surface roughness stabilization method of claim 4, wherein: in the step 4), a point corresponding to the current processing condition is found in the surface roughness prediction model, a direction with the fastest surface roughness change corresponding to the on-line adjustable parameter is found in the surface roughness prediction model based on the point, a target point is taken in the direction with the fastest surface roughness change, the surface roughness of the target point is located in a set threshold range, and the processing parameter corresponding to the target point is fed back to the mechanical processing system.
6. The digital twinning-based surface roughness stabilization method of claim 5, wherein: if the continuous adjustment can adjust the parameters on line for more than N times, R can not be satisfied a pre If the processing parameters are within the set threshold range, the condition that only the gesture parameters of the tool and the rotating speed of the spindle cannot meet the current processing requirements is indicated, and at the moment, all the processing parameters affecting the surface roughness are taken as adjustment objects, wherein the adjustment method comprises the following steps: finding a point corresponding to the current processing condition in the surface roughness prediction model, and finding the surface roughness in the surface roughness prediction model based on the pointAnd taking a target point in the direction with the fastest change of the surface roughness to enable the surface roughness at the target point to be located in a set threshold range, and feeding back processing parameters corresponding to the target point to a machining system, wherein N is a positive integer greater than or equal to 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010692248.4A CN111859566B (en) | 2020-07-17 | 2020-07-17 | Digital twinning-based surface roughness stabilization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010692248.4A CN111859566B (en) | 2020-07-17 | 2020-07-17 | Digital twinning-based surface roughness stabilization method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111859566A CN111859566A (en) | 2020-10-30 |
| CN111859566B true CN111859566B (en) | 2023-11-17 |
Family
ID=73001836
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010692248.4A Active CN111859566B (en) | 2020-07-17 | 2020-07-17 | Digital twinning-based surface roughness stabilization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111859566B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI766489B (en) * | 2020-12-21 | 2022-06-01 | 財團法人工業技術研究院 | Monitoring method and system for machine tool |
| CN112859739B (en) * | 2021-01-15 | 2022-07-01 | 天津商业大学 | Digital twin-driven multi-axis numerical control machine tool contour error suppression method |
| CN113095195B (en) * | 2021-04-03 | 2023-04-07 | 西北工业大学 | Part unique identification method based on surface topography self-characteristics |
| CN114453630B (en) * | 2022-01-20 | 2023-04-04 | 湖北文理学院 | Method and device for controlling machine tool to mill non-stick tool, electronic equipment and storage medium |
| CN116089818B (en) * | 2023-01-10 | 2023-10-27 | 南京航空航天大学 | Workpiece surface roughness prediction method, system and product in machining process |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761429A (en) * | 2014-01-10 | 2014-04-30 | 大连理工大学 | Milling workpiece surface roughness predicting method |
| CN110045608A (en) * | 2019-04-02 | 2019-07-23 | 太原理工大学 | Based on the twin mechanical equipment component structural dynamic state of parameters optimization method of number |
| CN110348075A (en) * | 2019-06-20 | 2019-10-18 | 湖南科技大学 | A kind of grinding surface roughness prediction technique based on improvement algorithm of support vector machine |
| CN110705882A (en) * | 2019-09-30 | 2020-01-17 | 江苏科技大学 | Twin data driven ship assembly product quality control system and configuration method |
| CN110900307A (en) * | 2019-11-22 | 2020-03-24 | 北京航空航天大学 | Numerical control machine tool cutter monitoring system driven by digital twin |
| CN111008502A (en) * | 2019-11-25 | 2020-04-14 | 北京航空航天大学 | Fault prediction method for complex equipment driven by digital twin |
| CN111161410A (en) * | 2019-12-30 | 2020-05-15 | 中国矿业大学(北京) | Mine digital twinning model and construction method thereof |
| CN111210359A (en) * | 2019-12-30 | 2020-05-29 | 中国矿业大学(北京) | Intelligent mine scene oriented digital twin evolution mechanism and method |
| CN111365158A (en) * | 2020-03-02 | 2020-07-03 | 东方电气集团东方电机有限公司 | Real-time state evaluation and life cycle management prediction system for water turbine runner |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113168891A (en) * | 2018-09-14 | 2021-07-23 | 西北大学 | Data-driven representation and clustering discretization method and system for design optimization and/or performance prediction of material systems and application thereof |
-
2020
- 2020-07-17 CN CN202010692248.4A patent/CN111859566B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761429A (en) * | 2014-01-10 | 2014-04-30 | 大连理工大学 | Milling workpiece surface roughness predicting method |
| CN110045608A (en) * | 2019-04-02 | 2019-07-23 | 太原理工大学 | Based on the twin mechanical equipment component structural dynamic state of parameters optimization method of number |
| CN110348075A (en) * | 2019-06-20 | 2019-10-18 | 湖南科技大学 | A kind of grinding surface roughness prediction technique based on improvement algorithm of support vector machine |
| CN110705882A (en) * | 2019-09-30 | 2020-01-17 | 江苏科技大学 | Twin data driven ship assembly product quality control system and configuration method |
| CN110900307A (en) * | 2019-11-22 | 2020-03-24 | 北京航空航天大学 | Numerical control machine tool cutter monitoring system driven by digital twin |
| CN111008502A (en) * | 2019-11-25 | 2020-04-14 | 北京航空航天大学 | Fault prediction method for complex equipment driven by digital twin |
| CN111161410A (en) * | 2019-12-30 | 2020-05-15 | 中国矿业大学(北京) | Mine digital twinning model and construction method thereof |
| CN111210359A (en) * | 2019-12-30 | 2020-05-29 | 中国矿业大学(北京) | Intelligent mine scene oriented digital twin evolution mechanism and method |
| CN111365158A (en) * | 2020-03-02 | 2020-07-03 | 东方电气集团东方电机有限公司 | Real-time state evaluation and life cycle management prediction system for water turbine runner |
Non-Patent Citations (1)
| Title |
|---|
| 王时龙 ; 王彦凯 ; 杨波 ; 王四宝.基于层次化数字孪生的工业互联网制造新范式——雾制造.《计算机集成制造系统》.2019,3070-3080. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111859566A (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111859566B (en) | Digital twinning-based surface roughness stabilization method | |
| EP3213161B1 (en) | Method for optimizing the productivity of a machining process of a cnc machine | |
| Wang et al. | Predictive modeling of surface roughness in lenses precision turning using regression and support vector machines | |
| CN114925596A (en) | Method for optimizing casting platform finish milling parameters based on wafer test equipment | |
| CN111975015B (en) | Ultra-precise turning method with uniformly distributed tool contacts | |
| CN109597355B (en) | Design method of curved surface micro-texture numerical control machining cutter shaft vector | |
| CN114509991B (en) | Numerical control machine tool cutting stability prediction and optimization method considering parameter uncertainty | |
| CN109093447A (en) | A kind of knife rail design method based on cutter uniform wear | |
| US6999843B2 (en) | Numerical control apparatus and CAM system | |
| Palanikumar et al. | Mitigated cutting force and surface roughness in titanium Alloy-Multiple effective guided chaotic multi objective Teaching learning based optimization | |
| Lin et al. | Optimal parameter ranges of material removal depth of abrasive cloth wheel polishing based on sensitivity analysis | |
| Cai et al. | Cutter orientation planning in NC machining for surface similar to revolution body with considering kinematic characteristics | |
| Zhang et al. | Machining parameter optimisation for aviation aluminium-alloy thin-walled parts in high-speed milling | |
| CN114750146B (en) | Robot milling track precision compensation method | |
| Zhao et al. | Theoretical and experimental investigation of the effect of the machining process chain on surface generation in ultra-precision fly cutting | |
| CN113204852B (en) | Method and system for predicting milling surface appearance of ball-end milling cutter | |
| Vainer | Determining the parameters of the technological space in bilateral end grinding. | |
| CN108381331A (en) | A kind of planar part overall situation correction of the flank shape processing unit (plant) and method | |
| Zhu et al. | Research on cutting force of turn-milling based on thin-walled blade | |
| CN113626953A (en) | High-energy-efficiency milling error dynamic distribution characteristic identification method | |
| Koprowski et al. | Influence of tilt and lead angles on 5-axis grinding with spherical mounted points | |
| Chen et al. | Research on polishing parameters optimization for free curved surface | |
| CN106960080B (en) | Prediction method for ultrasonic milling undeformed cutting thickness of non-uniform helical angle milling cutter | |
| Bilek et al. | Prediction and modeling of roughness in ball end milling with tool-surface inclination | |
| Duan et al. | Study of the effect of tool orientation on surface roughness in five-axis milling of 300M steel |
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 |