CN114091300A - Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system - Google Patents

Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system Download PDF

Info

Publication number
CN114091300A
CN114091300A CN202111339195.9A CN202111339195A CN114091300A CN 114091300 A CN114091300 A CN 114091300A CN 202111339195 A CN202111339195 A CN 202111339195A CN 114091300 A CN114091300 A CN 114091300A
Authority
CN
China
Prior art keywords
ball screw
feeding system
screw feeding
model
twin
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111339195.9A
Other languages
Chinese (zh)
Inventor
张玮
朱坚民
黄之文
朱迪
刘绪荣
李孝茹
赵展
田丰庆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Shanghai for Science and Technology
Original Assignee
University of Shanghai for Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Shanghai for Science and Technology filed Critical University of Shanghai for Science and Technology
Priority to CN202111339195.9A priority Critical patent/CN114091300A/en
Publication of CN114091300A publication Critical patent/CN114091300A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/17—Mechanical parametric or variational design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/10—Numerical modelling
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Computer Hardware Design (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Transmission Devices (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention discloses a method for identifying dynamic characteristic parameters of a rolling joint part of a ball screw feeding system, which comprises the following steps: establishing a twin mechanical model of the ball screw feeding system by taking the ball screw feeding system as an analysis object; selecting a series of elastic joint equivalent rolling joint parts, and establishing a twin finite element model of the ball screw feeding system, which can simultaneously consider a plurality of rolling joint parts and the rigidity and the damping of each direction of the rolling joint parts; taking a dynamic characteristic parameter to be identified as input and taking the first six-order natural frequency of the system as output, designing a Deep Neural Network (DNN) model, training, preferably optimizing the number of the optimal network layers, determining a twin data model and finally completing the establishment of a dynamic digital twin model of the ball screw feeding system; and constructing an optimized identification model by combining modal test data, and solving by adopting a particle swarm algorithm so as to identify each rigidity and damping parameter. According to the invention, the dynamic modeling precision of the ball screw feeding system is improved, and the dynamic characteristic parameters of each rolling joint part are efficiently and accurately identified.

Description

Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system
Technical Field
The invention relates to the technical field of machine tool joint dynamic characteristic parameter identification, in particular to a method for identifying dynamic characteristic parameters of a rolling joint of a ball screw feeding system.
Background
The ball screw feeding system is widely applied to the numerical control machine tool by the advantages of high positioning precision, high transmission efficiency, high reliability, relatively low cost and the like. The system is formed by connecting parts of different types, such as a lead screw, a nut, a rolling guide rail, a bearing, a workbench, a lathe bed and the like. The rolling joint part is used as a movable joint part, the rolling joint part is large in quantity and relatively dispersed in distribution, the rigidity value of the rolling joint part is relatively small, and the rolling joint part has a particularly obvious influence on the dynamic performance of the system. Therefore, the method has important significance for evaluating the dynamic performance of the feeding system of the numerical control machine tool and improving the design level of the machine tool by accurately establishing a dynamic model of the ball screw feeding system and accurately identifying the dynamic characteristic parameters of the rolling joint part of the ball screw feeding system. Establishing an accurate kinetic model is a prerequisite for identifying dynamic characteristic parameters of the joint.
From the present research, the following problems still exist with respect to ball screw feed system dynamics modeling and its rolling joint dynamic characteristic parameter identification. First, many documents assume complete decoupling of the dynamic performance in different directions, followed by independent identification of the dynamic characteristic parameters of the rolling joint in a single direction. Although a plurality of rolling joints can be considered simultaneously during dynamic modeling, the coupling effect which is not negligible in the dynamic characteristics of the feeding system in different directions is ignored, and the dynamic performance of the whole system is influenced by the combined action of the plurality of rolling joints distributed in the feeding system and the dynamic characteristic parameters in different directions, so that the dynamic modeling has limitations, and the modeling precision and the parameter identification accuracy are restricted. Secondly, a large number of documents consider the combined action of dynamic characteristic parameters in all directions of the space, perform dynamic modeling for a single joint, and identify all stiffness and damping parameters. Although the combined action of a plurality of directional dynamic characteristic parameters is considered during dynamic modeling, the studied rolling joint part is taken out of the ball screw feeding system, the assembly state of a study object is changed, the study object comes in and goes out of the actual situation, and the modeling precision and the parameter identification accuracy are also limited.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a method for identifying the dynamic characteristic parameters of the rolling joint part of the ball screw feeding system, which improves the dynamic modeling precision of the ball screw feeding system and efficiently and accurately identifies the dynamic characteristic parameters of each rolling joint part. To achieve the above objects and other advantages in accordance with the present invention, there is provided a ball screw feed system rolling joint dynamic characteristic parameter identification method, comprising:
s1, constructing a twin mechanical model of the ball screw feeding system by taking a physical entity of the ball screw feeding system as an analysis object;
s2, importing the twin mechanical model constructed in the step S1 into finite element software, and constructing a dynamic twin finite element model of the ball screw feeding system;
s3, generating a large amount of modal analysis data through a twin finite element model and a parametric analysis tool, and generating dynamic twin data of the ball screw feeding system;
s4, designing a DNN model by taking twin data as a data set, training, determining the optimal network depth, constructing a dynamic twin data model of the ball screw feeding system, and completing the establishment of a dynamic digital twin model of the ball screw feeding system;
s5, carrying out modal test on the physical entity of the ball screw feeding system to obtain experimental modal data;
s6, establishing an optimization model for identifying dynamic characteristic parameters of a rolling combination part of the ball screw feeding system, and identifying all rigidity and damping parameters at the same time;
and S7, substituting the recognition result in the step S6 into the twin finite element model to solve modal data, comparing the modal data with an experimental result, and checking the correctness and the precision.
Preferably, in step S1, attribute information such as mechanical structure, geometric dimension, etc. of the physical entity of the ball screw feeding system is digitally mirrored, and a twin mechanical model of the system is created by using three-dimensional software.
Preferably, in step S2, the twin mechanical model is imported into a multi-body dynamics module of a finite element software COMSOL Multiphysics, a series of matched elastic joints are selected according to actual conditions to be equivalent to each rolling joint, total stiffness and total damping in x, y and z directions of each joint are respectively defined, and material properties, constraint conditions and mesh division are further set, so as to establish a ball screw feeding system dynamics twin finite element model that can simultaneously consider a plurality of rolling joints and dynamic characteristic parameters in different directions thereof.
Preferably, in step S3, according to the hertzian contact theory and the literature investigation, the value ranges of the stiffness and damping parameters in different directions of each rolling joint portion are determined, random, uniform and disordered sampling is performed in a value space, 6000 groups of dynamic characteristic parameter sample points to be identified are obtained, and the dynamic characteristic parameters are used as input items of finite element analysis; by utilizing a parametric scanning function of finite element software COMSOL Multiphysics, input parameters are automatically extracted group by group and substituted into a finite element model for modal analysis, 6000 groups of inherent frequencies of the first six orders of the ball screw feeding system under different corresponding input conditions are solved and used as output items of the finite element analysis, and therefore twin data of the ball screw feeding system digital twin kinetic model are generated.
Preferably, in step S4, the dynamic characteristic parameters of the rolling joint to be identified are used as input, the first six-order natural frequency of the ball screw feeding system is used as output, and DNN models with 2-10 hidden layers are respectively designed; training the constructed DNN model by taking the twin data generated in the step S3 as a data set, so as to determine the optimal number of hidden layers and further complete the establishment of a dynamic twin data model of the ball screw feeding system; and completing the establishment of a dynamic digital twinning model of the ball screw feeding system by fusing a twinning mechanical model, a twinning finite element model and a twinning data model.
Preferably, in step S5, the LMS test. lab vibration noise test system is used to perform a modal test on the ball screw feeding system by a single-point excitation and multi-point vibration pickup method, so as to obtain an experimental modal result of the system.
Preferably, in step S6, the stiffness and damping parameters to be identified are used as design variables, the value ranges of the dynamic characteristic parameters are used as constraint conditions, an optimization objective function is constructed by combining the natural frequency experimental values measured by modal testing and the natural frequency predicted values driven by the digital twin dynamics model, an optimization model for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system is established, a particle swarm optimization algorithm is used for solving, and the stiffness and damping parameters of all the rolling joints in different directions are identified at the same time.
Preferably, in step S7, the optimally identified stiffness and damping parameters are substituted into the twin finite element model for modal analysis, so as to obtain the first six-order natural frequency of the ball screw feeding system based on the identified parameters, and the first six-order natural frequency is compared with the corresponding order experiment results, so as to calculate the relative error, thereby checking the accuracy and precision of the method of the present invention.
Compared with the prior art, the invention has the beneficial effects that:
(1) a series of contact characteristics of equivalent rolling joint parts of elastic joints with different direction rigidity and damping can be considered at the same time, a finite element model of the ball screw feeding system in an assembly state is established, and the influence of multi-direction dynamic characteristic parameters of multiple joint parts on the dynamic characteristic of the feeding system is comprehensively considered.
(2) A digital twin dynamics model capable of faithfully mirroring the ball screw feeding system is established by means of three-dimensional mechanical modeling, parametric design, finite element simulation, DNN and the like, digital twin is applied to dynamic modeling of a machine tool joint and dynamic characteristic parameter identification, and dynamic modeling precision and parameter identification accuracy are improved.
(3) The method is not only suitable for the rolling joint part of the ball screw, but also suitable for solving the dynamics problem of a complex mechanical system with multiple joint parts and multi-directional dynamic characteristic parameters, and has high technical popularization value.
Drawings
FIG. 1 is a block diagram of a method for identifying dynamic characteristic parameters of a rolling joint of a ball screw feed system according to the present invention;
FIG. 2 is a twin mechanical model of a ball screw feed system rolling joint dynamic characteristic parameter identification method according to the present invention;
FIG. 3 is a diagram of a series of elastic joint equivalent rolling joints which can simultaneously consider stiffness and damping in different directions according to the dynamic characteristic parameter identification method of the rolling joints of the ball screw feeding system of the present invention;
FIG. 4 is a twin finite element model of a method for identifying dynamic characteristic parameters of a rolling contact portion of a ball screw feed system according to the present invention;
FIG. 5 is a performance diagram of DNN models of different depths according to a method for identifying dynamic characteristic parameters of a rolling joint of a ball screw feed system;
FIG. 6 is a modal test equivalent three-dimensional model and a test point distribution diagram of the method for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system according to the invention;
FIG. 7 is a graph of the optimization convergence of particle swarm optimization algorithm for the dynamic characteristic parameter identification method of the rolling joint of the ball screw feeding system according to the present invention;
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 7, a method for identifying a dynamic characteristic parameter of a rolling contact portion of a ball screw feed system includes: s1, constructing a twin mechanical model of the ball screw feeding system by taking a physical entity of the ball screw feeding system as an analysis object;
s2, importing the twin mechanical model constructed in the step S1 into finite element software, and constructing a dynamic twin finite element model of the ball screw feeding system;
s3, generating a large amount of modal analysis data through a twin finite element model and a parametric analysis tool, and generating dynamic twin data of the ball screw feeding system;
s4, designing a DNN model by taking twin data as a data set, training, determining the optimal network depth, constructing a dynamic twin data model of the ball screw feeding system, and completing the establishment of a dynamic digital twin model of the ball screw feeding system;
s5, carrying out modal test on the physical entity of the ball screw feeding system to obtain experimental modal data;
s6, establishing an optimization model for identifying dynamic characteristic parameters of a rolling combination part of the ball screw feeding system, and identifying all rigidity and damping parameters at the same time;
and S7, substituting the recognition result in the step S6 into the twin finite element model to solve modal data, comparing the modal data with an experimental result, and checking the correctness and the precision.
Further, in step S1, attribute information such as mechanical structure, geometric dimension, and the like of the physical entity of the ball screw feeding system is digitally mirrored, and a twin mechanical model of the system is established by using three-dimensional software.
Further, in the step S2, the twin mechanical model is imported into a multi-body dynamics module of the finite element software COMSOL Multiphysics, a series of matched elastic joints are selected to be equivalent to each rolling joint according to actual conditions, total stiffness and total damping in x, y and z directions of each joint are respectively defined, and material properties, constraint conditions and mesh division are further set, so that a ball screw feeding system dynamics twin finite element model which can simultaneously consider a plurality of rolling joints and dynamic characteristic parameters in different directions thereof is established.
Further, in step S3, according to the hertzian contact theory and the literature investigation, determining the value ranges of the stiffness and damping parameters of each rolling joint in different directions, randomly, uniformly and disorderly sampling in a value space, obtaining 6000 groups of dynamic characteristic parameter sample points to be identified, and using the dynamic characteristic parameters as input items of finite element analysis; by utilizing a parametric scanning function of finite element software COMSOL Multiphysics, input parameters are automatically extracted group by group and substituted into a finite element model for modal analysis, 6000 groups of inherent frequencies of the first six orders of the ball screw feeding system under different corresponding input conditions are solved and used as output items of the finite element analysis, and therefore twin data of the ball screw feeding system digital twin kinetic model are generated.
Further, in the step S4, the dynamic characteristic parameters of the rolling joint to be identified are used as input, the first six-order natural frequency of the ball screw feeding system is used as output, and DNN models with 2-10 hidden layers are respectively designed; training the constructed DNN model by taking the twin data generated in the step S3 as a data set, so as to determine the optimal number of hidden layers and further complete the establishment of a dynamic twin data model of the ball screw feeding system; and completing the establishment of a dynamic digital twinning model of the ball screw feeding system by fusing a twinning mechanical model, a twinning finite element model and a twinning data model.
Furthermore, in step S5, a LMS test.lab vibration noise test system is used to perform a modal test on the ball screw feeding system by a single-point excitation and multi-point vibration pickup method, so as to obtain an experimental modal result of the system.
Further, in step S6, the stiffness and damping parameters to be identified are used as design variables, the value ranges of the dynamic characteristic parameters are used as constraint conditions, an optimization objective function is constructed by combining the natural frequency experimental values measured by modal testing and the natural frequency predicted values driven by the digital twin dynamic model, an optimization model for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system is established, a particle swarm optimization algorithm is used for solving, and the stiffness and damping parameters of all the rolling joints in different directions are identified at the same time.
Further, in the step S7, the optimally identified stiffness and damping parameters are substituted into the twin finite element model for modal analysis, the first six-order natural frequency of the ball screw feeding system based on the identified parameters is obtained, and is compared with the corresponding order experiment results, and the relative error is calculated, so as to check the accuracy and precision of the method of the present invention.
Example 1
S1: a self-designed and manufactured ball screw feeding system is used as an analysis object, attribute information such as a mechanical structure, a geometric dimension and the like of a physical entity of the ball screw feeding system is mirrored, and a twin mechanical model of the ball screw feeding system is established by utilizing SolidWorks and is shown in figure 2.
S2: and (3) importing the twin mechanical model into a multi-body dynamic module of finite element software COMSOL Multiphysics, adopting a series of predefined elastic joints to be equivalent to each rolling joint part, and defining the total rigidity and the total damping of each elastic joint in each direction under a Cartesian coordinate system. The transmission direction of the ball screw feeding system is defined as an x direction, the supporting direction is defined as a y direction, and the direction perpendicular to both the x direction and the y direction is defined as a z direction. As shown in FIG. 3, the screw-nut equivalent screw joint in the software is selected, and the dynamic characteristic parameters of each direction are defined as
Figure BDA0003351243200000071
And
Figure BDA0003351243200000072
selecting hinge pair joint equivalent fixationAn end bearing joint part, the dynamic characteristic parameters of each direction of which are defined as
Figure BDA0003351243200000073
And
Figure BDA0003351243200000074
selecting the equivalent simply supported end bearing joint part of the cylindrical auxiliary joint, and defining the dynamic characteristic parameters of each direction as
Figure BDA0003351243200000075
And
Figure BDA0003351243200000076
selecting a prism joint equivalent rolling guide rail joint part, and defining dynamic characteristic parameters of each direction as
Figure BDA0003351243200000077
And
Figure BDA0003351243200000078
the screw rod is hardly loaded in the x direction of the simple support end, the bearing can play a small amount along the direction, so that the deformation and thermal deformation of the screw rod caused by temperature change are released, and the rigidity and damping of the joint part in the x direction can be regarded as 0, namely the rigidity and damping of the joint part in the x direction are regarded as 0
Figure BDA0003351243200000079
The displacement of the rolling guide rail pair in the x direction is mainly that the nut drives the workbench to do linear feed motion, the joint part of the rolling guide rail is hardly loaded in the direction, and the rigidity and the damping of the joint part in the x direction can be regarded as 0, namely
Figure BDA0003351243200000081
And the lead screw is used as a revolving body, rotates around the x axis and is mainly acted by the pre-tightening force of the nut in the axial direction (x direction), and the lead screw-nut joint part, the fixed end joint part and the simple end bearing joint part are hardly influenced by the gravity of the workbench, so that the rolling joint of the three revolving kinematic pairs can be consideredThe contact characteristics of the joint in both radial directions (y-direction and z-direction) of the feed system are comparable, i.e.
Figure BDA0003351243200000082
Figure BDA0003351243200000083
And then defining material properties, setting constraint conditions and dividing proper grids so as to establish a twin finite element model of the ball screw feeding system, as shown in fig. 4.
S3: according to the Hertz contact theory and literature research, the value ranges of the stiffness and the damping parameters of different directions of each rolling joint part of the ball screw feeding system are determined as shown in the table 1.
TABLE 1 rolling combination part dynamic characteristic parameter value range in each direction
Figure BDA0003351243200000084
Figure BDA0003351243200000091
And randomly, uniformly and disorderly sampling in the range to generate 6000 groups of dynamic characteristic parameter sample points to be identified as shown in the formula (1), and taking the dynamic characteristic parameters as input items of finite element analysis.
Figure BDA0003351243200000092
By using the parameterized scanning function of finite element software COMSOLMUTIPhysics, input parameters are automatically extracted group by group and substituted into a finite element model for modal analysis, and 6000 groups of inherent frequencies of the ball screw feeding system under different input conditions shown in a formula (2) are solved to serve as output items of finite element analysis. Thereby generating twin data of the digital twin dynamics model of the ball screw feed system.
(f'1i,f'2i,f'3i,f'4i,f'5i,f'6i)i＝1,2,…,6000 (2)
S4: with the rolling joint dynamic characteristic parameter variable to be identified
Figure BDA0003351243200000093
For the first six natural frequencies (f) of the input, ball screw feed system1,f2,f3,f4,f5,f6) And respectively designing a DNN model with 2-10 hidden layers for output, wherein the input layer of the DNN model comprises 14 neurons, and the output layer comprises 6 neurons. Setting the number of the neurons of each layer to be 11, selecting 'Sigmod' as an excitation function of a hidden layer and 'Linear' as an excitation function of an output layer, setting the learning rate to be 0.01 and the minimum gradient to be 1 multiplied by 10-20. Taking global Mean Square Error (MSE) as a loss function, the calculation formula is shown as formula (3):
Figure BDA0003351243200000101
wherein ([ k, c)]) For input layer variables, N is the number of sample groups, fjThe actual value in the network training process, namely the j-th order natural frequency of the ball screw feeding system solved through finite element analysis,
Figure BDA0003351243200000102
the predicted value in the network training, namely the j-th order natural frequency of the ball screw feeding system predicted by the DNN model, is smaller, the smaller the MSE value is, the closer the predicted value is to the true value, and the more the model is converged.
Training the constructed DNN model by using the twin data generated in the step S3 as a data set, and setting a training cut-off condition to be 1 × 10 of the minimum global mean square error-8Or the maximum number of iterations is 2000, when one of the two conditions is met and the loss function converges, the DNN training is considered complete and valid.
Taking a data set of 500 groups of non-training sample points as a test set, and calculating the average absolute Percentage Error (MAPE) of the output result of the test set by using a trained DNN model with 2-10 hidden layers, wherein the calculation formula is shown as the formula (4):
Figure BDA0003351243200000103
the natural frequency is used as an evaluation index to reflect the closeness degree of a predicted value of the natural frequency and a finite element value in a test set, and the smaller the index value is, the predicted natural frequency
Figure BDA0003351243200000104
And finite element calculation result fijThe closer the model is, the better the model performance and the higher the prediction accuracy. The test set MAPE results obtained by the DNN models with different depths are shown in FIG. 5, and from the analysis results, the optimal number of hidden layer layers is 6, and finally the number of layer-by-layer neurons of the DNN model is determined to be 14-11-11-11-11-11-11-11-6.
S5: and performing modal test on the ball screw feeding system by using the LMS (least mean square) through a single-point excitation and multipoint vibration pickup method to obtain an experimental modal result of the system. Firstly, establishing an equivalent three-dimensional model as shown in fig. 6 in LMS modal analysis software, arranging 33 measuring points in one-to-one correspondence with positions in a real object, wherein the measuring point No. 33 at the center position of the surface of the workbench is set as an excitation point. This point was hammered with a Kistler model 9724a200 hammer to excite the system. And then picking up vibration signals of the point in the x, y and z directions respectively by using a BK4525B type three-way acceleration sensor arranged at the corresponding measuring point, transmitting the vibration signals to an LMS data acquisition device, and analyzing and processing experimental data by LMS modal analysis software. After the vibration signals of all the measuring points are collected, the LMS analyzes and calculates the natural frequency of the ball screw feeding system, as shown in Table 2.
TABLE 2 Experimental values of the first six-step natural frequency of a ball screw feed system
Order of the order Natural frequency experiment value (Hz)
1 59.63
2 88.47
3 178.11
4 237.17
5 279.95
6 383.59
S6: the stiffness and damping parameters to be identified are used as design variables, the value range of each dynamic characteristic parameter is used as a constraint condition, and the experimental value f of the first six-order natural frequency measured by combining modal testing is usedjNatural frequency prediction value driven by digital twin dynamic model
Figure BDA0003351243200000111
And (5) constructing an optimization objective function, and establishing an optimization model of the dynamic characteristic parameter identification of the rolling joint of the ball screw feeding system as shown in the formula (5).
Figure BDA0003351243200000121
The above equation is solved by a particle swarm optimization algorithm to identify the dynamic characteristic parameters of each rolling joint in each direction, the convergence curve is shown in fig. 7, and the identification result is shown in table 3.
TABLE 3 optimized recognition results
Figure BDA0003351243200000122
Figure BDA0003351243200000131
S7: and substituting the optimally identified rigidity and damping parameters into a twin finite element model in the digital twin dynamic model, carrying out modal analysis on the twin finite element model, obtaining the first six-order natural frequency of the ball screw feeding system based on the identified parameters, and comparing the six-order natural frequency with the experimental values step by step, wherein the comparison result is shown in table 4.
TABLE 4 comparison of natural frequency finite element calculated values and experimental values for ball screw feed systems based on identified dynamic characteristic parameters
Order of the order Experimental values (Hz) Finite element value (Hz) based on identified parameters Relative error (%)
1 59.63 59.55 -0.13
2 88.47 89.49 1.15
3 178.11 179.30 0.67
4 237.17 237.06 -0.05
5 279.95 279.46 -0.18
6 383.59 394.72 2.90
It can be seen from the above table that the relative error values of the natural frequency solved based on the identified dynamic characteristic parameters and the natural frequency measured by the modal test are not more than 3%, so that high precision is achieved, and the correctness and the effectiveness of the method are verified.
The number of devices and the scale of the processes described herein are intended to simplify the description of the invention, and applications, modifications and variations of the invention will be apparent to those skilled in the art.
While embodiments of the invention have been described above, it is not limited to the applications set forth in the description and the embodiments, which are fully applicable in various fields of endeavor to which the invention pertains, and further modifications may readily be made by those skilled in the art, it being understood that the invention is not limited to the details shown and described herein without departing from the general concept defined by the appended claims and their equivalents.

Claims (8)

1. A method for identifying dynamic characteristic parameters of a rolling joint of a ball screw feeding system is characterized by comprising the following steps:
s1, constructing a twin mechanical model of the ball screw feeding system by taking a physical entity of the ball screw feeding system as an analysis object;
s2, importing the twin mechanical model constructed in the step S1 into finite element software, and constructing a dynamic twin finite element model of the ball screw feeding system;
s3, generating a large amount of modal analysis data through a twin finite element model and a parametric analysis tool, and generating dynamic twin data of the ball screw feeding system;
s4, designing a DNN model by taking twin data as a data set, training, determining the optimal network depth, constructing a dynamic twin data model of the ball screw feeding system, and completing the establishment of a dynamic digital twin model of the ball screw feeding system;
s5, carrying out modal test on the physical entity of the ball screw feeding system to obtain experimental modal data;
s6, establishing an optimization model for identifying dynamic characteristic parameters of a rolling combination part of the ball screw feeding system, and identifying all rigidity and damping parameters at the same time;
and S7, substituting the recognition result in the step S6 into the twin finite element model to solve modal data, comparing the modal data with an experimental result, and checking the correctness and the precision.
2. The method as claimed in claim 1, wherein the step S1 is performed by digitally mirroring the physical entity of the ball screw feeding system with the mechanical structure and geometric dimension of the physical entity, and using three-dimensional software to create a twin mechanical model of the physical entity.
3. The method for identifying the dynamic characteristic parameters of the rolling joints of the ball screw feeding system as claimed in claim 1, wherein in step S2, the twin mechanical model is introduced into a multi-body dynamic module of finite element software COMSOL Multiphysics, a series of matched elastic joints are selected according to actual conditions to be equivalent to each rolling joint, the total stiffness and the total damping in x, y and z directions of each joint are respectively defined, and material properties, constraint conditions and meshes are further set, so as to establish the ball screw feeding system dynamic twin finite element model which can simultaneously consider a plurality of rolling joints and dynamic characteristic parameters in different directions thereof.
4. The method for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system according to claim 1, wherein in step S3, the value ranges of the stiffness and damping parameters of the rolling joints in different directions are determined according to the hertzian contact theory and the literature research condition, and random, uniform and disordered sampling is performed in the value space to obtain 6000 groups of sample points of the dynamic characteristic parameters to be identified, and the dynamic characteristic parameters are used as input items of finite element analysis; by utilizing a parametric scanning function of finite element software COMSOL Multiphysics, input parameters are automatically extracted group by group and substituted into a finite element model for modal analysis, 6000 groups of inherent frequencies of the first six orders of the ball screw feeding system under different corresponding input conditions are solved and used as output items of the finite element analysis, and therefore twin data of the ball screw feeding system digital twin kinetic model are generated.
5. The method for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system according to claim 1, wherein in step S4, the dynamic characteristic parameters of the rolling joint to be identified are used as input, the first six-order natural frequency of the ball screw feeding system is used as output, and DNN models with 2-10 hidden layers are respectively designed; training the constructed DNN model by taking the twin data generated in the step S3 as a data set, so as to determine the optimal number of hidden layers and further complete the establishment of a dynamic twin data model of the ball screw feeding system; and completing the establishment of a dynamic digital twinning model of the ball screw feeding system by fusing a twinning mechanical model, a twinning finite element model and a twinning data model.
6. The method for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system according to claim 1, wherein in step S5, the LMS test.lab vibration noise test system is used to perform a modal test on the ball screw feeding system by a single-point excitation and multi-point vibration pickup method, so as to obtain experimental modal results of the system.
7. The method for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system according to claim 1, wherein in step S6, stiffness and damping parameters to be identified are used as design variables, the value range of each dynamic characteristic parameter is used as a constraint condition, an optimization objective function is constructed by combining an experimental value of natural frequency measured by modal testing and a predicted value of natural frequency driven by a digital twin dynamics model, an optimization model for identifying the dynamic characteristic parameters of the rolling joint of the ball screw feeding system is established, a particle swarm optimization algorithm is used for solving, and the stiffness and damping parameters of all the rolling joints in different directions are identified at the same time.
8. The method as claimed in claim 1, wherein the stiffness and damping parameters are inputted into the twin finite element model for modal analysis in step S7, and the first six-order natural frequency of the ball screw feeding system based on the identified parameters is obtained and compared with the experimental results of the corresponding order to calculate the relative error, thereby checking the accuracy and precision of the method.
CN202111339195.9A 2021-11-12 2021-11-12 Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system Pending CN114091300A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111339195.9A CN114091300A (en) 2021-11-12 2021-11-12 Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111339195.9A CN114091300A (en) 2021-11-12 2021-11-12 Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system

Publications (1)

Publication Number Publication Date
CN114091300A true CN114091300A (en) 2022-02-25

Family

ID=80300240

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111339195.9A Pending CN114091300A (en) 2021-11-12 2021-11-12 Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system

Country Status (1)

Country Link
CN (1) CN114091300A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114707428A (en) * 2022-06-01 2022-07-05 中科航迈数控软件(深圳)有限公司 Method, device, terminal and storage medium for simulating unobservable links of numerical control machine tool
CN114722660A (en) * 2022-03-22 2022-07-08 西北工业大学 Method for analyzing rigid-flexible coupling dynamic contact characteristics of planetary roller screw
CN114999596A (en) * 2022-05-31 2022-09-02 上海理工大学 Sliding joint virtual material layer parameter identification method

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114722660A (en) * 2022-03-22 2022-07-08 西北工业大学 Method for analyzing rigid-flexible coupling dynamic contact characteristics of planetary roller screw
CN114999596A (en) * 2022-05-31 2022-09-02 上海理工大学 Sliding joint virtual material layer parameter identification method
CN114707428A (en) * 2022-06-01 2022-07-05 中科航迈数控软件(深圳)有限公司 Method, device, terminal and storage medium for simulating unobservable links of numerical control machine tool

Similar Documents

Publication Publication Date Title
CN114091300A (en) Dynamic characteristic parameter identification method for rolling joint part of ball screw feeding system
Liu et al. Thermally-induced error compensation of spindle system based on long short term memory neural networks
Zhu et al. Robust modeling method for thermal error of CNC machine tools based on random forest algorithm
Bahloul et al. A study on optimal design of process parameters in single point incremental forming of sheet metal by combining Box–Behnken design of experiments, response surface methods and genetic algorithms
Papadimitriou et al. Variability of updated finite element models and their predictions consistent with vibration measurements
CN106372278A (en) Sensitivity analysis method jointly considering input parameter uncertainty and proxy model uncertainty
CN102785129A (en) On-line detection method of curved surface machining precision of complex part
CN104597842A (en) BP neutral network heavy machine tool thermal error modeling method optimized through genetic algorithm
CN110210079B (en) Machine tool supporting part quality matching method oriented to complete machine dynamic characteristics
CN112668227A (en) Thin-wall part cutter relieving deformation error prediction model establishing method and application thereof
CN112926152B (en) Digital twin-driven thin-wall part clamping force precise control and optimization method
CN110007645A (en) A kind of feed system hybrid modeling method based on dynamics and deep neural network
CN106354955A (en) Sliding bearing rigidity recognition method based on mill vibration mode parameters
Wang et al. Crack parameters identification based on a kriging surrogate model for operating rotors
Wei et al. Digital twin-driven manufacturing equipment development
Zheng et al. Distributed Dynamic load identification on irregular planar structures using subregion interpolation
CN114970028A (en) Digital twin driven research and development method and system for manufacturing equipment
Jin et al. Wavelet neural network based on NARMA-L2 model for prediction of thermal characteristics in a feed system
CN112990601B (en) Worm wheel machining precision self-healing system and method based on data mining
CN111368478B (en) Modal parameter identification method based on movable combination part of rolling linear guide rail
CN116611314A (en) Machine tool machining process thermal error online evaluation method based on physical information neural network
CN114999596A (en) Sliding joint virtual material layer parameter identification method
CN113849901B (en) Improved self-adaptive optimization method and system for contact heat exchange coefficient identification
Abdelwahab Surface roughness modeling and prediction based on vibration signal analysis and machining parameters in milling of aluminum by artificial neural network
Kumar et al. Experimental validation of characteristic diagram-parameterization for environment-induced thermal interactions on machine tools in a climate chamber

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