CN116224905B - Machine tool thermal error prediction method and system based on joint distribution self-adaption - Google Patents
Machine tool thermal error prediction method and system based on joint distribution self-adaption Download PDFInfo
- Publication number
- CN116224905B CN116224905B CN202310189184.XA CN202310189184A CN116224905B CN 116224905 B CN116224905 B CN 116224905B CN 202310189184 A CN202310189184 A CN 202310189184A CN 116224905 B CN116224905 B CN 116224905B
- Authority
- CN
- China
- Prior art keywords
- thermal error
- source domain
- domain
- data
- target domain
- 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
- 238000009826 distribution Methods 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 52
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000005457 optimization Methods 0.000 claims abstract description 20
- 238000010606 normalization Methods 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 12
- 230000003044 adaptive effect Effects 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 8
- 230000006978 adaptation Effects 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 6
- 230000014509 gene expression Effects 0.000 claims description 6
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 229910052739 hydrogen Inorganic materials 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000003754 machining Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 239000003054 catalyst Substances 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
- G05B19/404—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/33—Director till display
- G05B2219/33133—For each action define function for compensation, enter parameters
-
- 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
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Numerical Control (AREA)
Abstract
The invention relates to the technical field of machine tool application optimization, in particular to a machine tool thermal error prediction method and system based on joint distribution self-adaption. The invention builds a thermal error regression model under a working condition and equipment in a source domain, can be applied to thermal error prediction tasks under other working conditions and equipment, and simultaneously expands and applies the thermal error prediction model to regression prediction scenes from scenes which can only be used for classifying problems by modifying the traditional method of the kernel function of the combined distribution self-adaptive algorithm, and adapts the classification evaluation index of the combined distribution self-adaptive algorithm to regression problem evaluation index, thereby being capable of providing a universal prediction scheme for thermal error data in the variable working condition, equipment crossing, multi-direction and spindle cutter processing in the thermal error field in a machine tool system.
Description
Technical Field
The invention relates to the technical field of machine tool application optimization, in particular to a machine tool thermal error prediction method and system based on joint distribution self-adaption.
Background
Thermal errors are a major factor affecting the precision manufacturing of vertical numerically controlled machining centers. Related researches show that in the machining process of the machine tool spindle, the thermal error generated by thermal deformation accounts for 40% -70% of the total machining error of the machine tool, and the higher the precision of the machine tool is, the larger the thermal error accounts for. Therefore, machine tool thermal error prediction and compensation techniques are important and difficult to study at present.
The existing thermal error prediction modeling method is only used for arranging a data acquisition platform for a specific machine tool, setting a fixed rotating speed and processing and acquiring data under the same working condition, and establishing a prediction model, and the method can be used for acquiring higher prediction precision for specific prediction tasks, but is difficult to adapt to all occasions under a specific working condition and equipment under variable working conditions, equipment crossing, multiple types and main shaft band knife processing scenes, so that design scheme replacement is caused, and even modeling prediction is difficult because thermal error data cannot be acquired during band knife processing.
On the other hand, because the types of machine tool equipment are more, the distribution difference of thermal error data under different working conditions is larger, the thermal error types are different due to different main shaft arrangement modes, and thermal error data are difficult to collect during machine tool cutter-carrying processing, so that thermal error prediction models under different equipment, different working conditions and different types are mutually different, thermal error data are difficult to collect during cutter-carrying processing, and a traditional method for constructing a model by deep learning depending on characteristic data and labels is difficult to implement. Joint distribution adaptive algorithms such as nearest neighbor algorithm and KNN, which are often used in the field of error data processing, cannot be applied to regression prediction because the internal kernel function is a classification algorithm, and thus the regression problem cannot be solved by the conventional migration concept of joint distribution adaptive.
Disclosure of Invention
The invention aims to solve the technical problems that the prediction and compensation data of the machine tool thermal error are difficult to collect, the prediction is difficult and the algorithm is not universal in the prior art.
In order to solve the technical problems, in a first aspect, an embodiment of the present invention provides a machine tool thermal error prediction method based on joint distribution self-adaption, the machine tool thermal error prediction method includes the following steps:
s1, acquiring thermal error experimental data of a numerical control machine through a machine tool system with a multi-source sensor;
s2, screening abnormal values of the thermal error experimental data, and removing outliers;
s3, judging whether a system error exists according to the type of the thermal error experimental data, and compensating a correction value for the thermal error realization data with the system error through residual error analysis;
s4, dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And the heat of the main shaft in three directions of the space coordinate axisError value { y } i |e x ,e y ,e z }；
S5, for the source domain feature data X i Normalization processing is carried out, and the source domain characteristic data X is established i And the thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)；
S6, dividing a target domain thermal error prediction task D from the thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
S7, constructing a joint distribution self-adaptive algorithm for regression problem;
s8, the source domain characteristic data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
S9, pseudo tag of the target domainAs a thermal error prediction value, the machine tool system is input and compensated to a tool feed amount.
Further, in step S2, the method for screening the abnormal value of the thermal error experimental data includes at least one of 3σ criterion and box-line drawing.
Further, the source domain regression model f s (·) Is one of a support vector regression model, a random forest regression model, an artificial neural network model and a deep neural network model.
Further, in step S7, the joint distribution adaptive algorithm includes a regression algorithm kernel function.
Still further, the joint distributed adaptation algorithm comprises the sub-steps of:
s71, minimizing the joint probability distribution distance of the source domain and the target domain in the thermal error experimental data, wherein the step S71 satisfies the relation (1):
Dist(D s ,D t )＝||P(X s )-P(X t )||+||P(y s |X s )-P(y t |X s )|| (1)；
wherein ,P(Xs )、P(X t )、P(y s |X s )、P(y t |X s ) Respectively representing the conditional probability distribution of the source domain characteristic data, the conditional probability distribution of the target domain characteristic data, the edge probability distribution of the source domain thermal error relative to the source domain characteristic data and the edge probability distribution of the target domain thermal error relative to the target domain characteristic data;
s72, minimizing the maximum mean difference between the source domain and the target domain, wherein the step S72 satisfies the relation (2):
wherein m and n respectively represent the number of samples in the source domain and the target domain, and A, T, H respectively represent a transformation matrix, a transpose of the matrix, and a hilbert space;
s73, introducing a regression algorithm kernel function to simplify the relation (2) and obtaining a relation (3):
Dist(D s ,D t )＝tr(A T XM 0 X T A) (3)；
wherein ,M0 A kernel function for the regression algorithm;
s74, associating the regression algorithm kernel function in the relation (3) with the sample numbers of the source domain and the target domain, wherein the step S74 satisfies the relation (4):
s75, use (X) s ,y s ) Training a regression model to obtain pseudo tags
S76, determining an optimized value of the joint distribution adaptive algorithm, wherein the optimized value meets a relation (5):
wherein ：
s77, constructing an optimization target of the joint distribution self-adaptive algorithm, wherein the optimization target meets a relation (6):
wherein ,is a regular term;
and S78, unifying the optimization targets according to the relational expressions (1) to (5) to obtain a relational expression (7):
s79, performing stretching transformation on the relation (7) to obtain a relation (8):
s710, replacing a calculation item in an optimization target by using a Lagrangian multiplier phi, and solving the optimization target, wherein the step S710 satisfies a relation (9):
and S711, iterating the joint distribution self-adaptive algorithm until the optimization target meets a preset iteration precision value.
Further, in step S75, the regression model is one of support vector regression and random forest regression.
In a second aspect, an embodiment of the present invention further provides a machine tool thermal error prediction system based on joint distribution adaptation, where the machine tool thermal error prediction system includes:
the data acquisition module is used for acquiring thermal error experimental data of the numerical control machine through a machine tool system with a multi-source sensor;
the screening module is used for screening abnormal values of the thermal error experimental data and removing outliers;
the compensation module is used for judging whether a system error exists according to the type of the thermal error experimental data and compensating a correction value for the thermal error realization data with the system error through residual error analysis;
the source domain data processing module is used for dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And thermal error values { y } of the principal axis in three directions of the spatial coordinate axis i |e x ,e y ,e z }；
A source domain attribution module for attributing the source domain characteristic data X i Normalization processing is carried out, and the source domain characteristic data X is established i And said at least one ofThermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)；
The target domain attribution module is used for dividing a target domain thermal error prediction task D from the thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
The algorithm construction module is used for constructing a joint distribution self-adaptive algorithm for regression problems;
an error calculation module for calculating the source domain feature data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
An error feedback module for pseudo-labeling the target domainAs a thermal error prediction value, the machine tool system is input and compensated to a tool feed amount.
The invention has the beneficial technical effects that the invention provides a machine tool thermal error prediction method based on joint distribution self-adaption, which can be applied to thermal error prediction tasks under other working conditions and equipment by constructing a thermal error regression model under the working conditions and equipment in a source domain, and simultaneously, can be only applied to regression prediction scenes from classification problem scenes by modifying the traditional method of a joint distribution self-adaption algorithm kernel function, and can provide a universal prediction scheme for thermal error data in the field of thermal error in a machine tool system, in a cross-equipment mode, in a multi-direction mode and in spindle knife processing by adapting the classification evaluation index of the algorithm to the regression problem evaluation index.
Drawings
FIG. 1 is a schematic flow chart of a machine tool thermal error prediction method based on joint distribution adaptation provided by an embodiment of the invention;
FIG. 2 is a schematic diagram of a multi-source heterogeneous data collection platform designed for a machine tool system according to an embodiment of the present invention;
FIG. 3 is a graph of probability distribution of data provided by an embodiment of the present invention;
FIG. 4 is a schematic view of visualization of DNN model parameters according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of DNN model prediction accuracy provided by an embodiment of the present invention;
FIG. 6 is a schematic diagram of prediction accuracy of thermal error prediction values according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a local method of thermal error prediction value according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, fig. 1 is a flow chart of a machine tool thermal error prediction method based on joint distribution adaptation according to an embodiment of the present invention, where the machine tool thermal error prediction method includes the following steps:
s1, acquiring thermal error experimental data of a numerical control machine through a machine tool system with a multi-source sensor.
S2, performing outlier screening on the thermal error experimental data to remove outliers.
Further, in step S2, the method for screening the abnormal value of the thermal error experimental data includes at least one of 3σ criterion and box-line drawing.
S3, judging whether a system error exists according to the type of the thermal error experimental data, and compensating a correction value for the thermal error realization data with the system error through residual error analysis.
S4, dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And thermal error values { y } of the principal axis in three directions of the spatial coordinate axis i |e x ,e y ,e z }。
S5, for the source domain feature data X i Normalization processing is carried out, and the source domain characteristic data X is established i And the thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)。
Further, the source domain regression model f s (. Cndot.) is one of a support vector regression model, a random forest regression model, an artificial neural network model, and a deep neural network model.
S6, dividing a target domain thermal error prediction task D from the thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
S7, constructing a joint distribution self-adaptive algorithm for regression problem.
Further, in step S7, the joint distribution adaptive algorithm includes a regression algorithm kernel function.
Still further, the joint distributed adaptation algorithm comprises the sub-steps of:
s71, minimizing the joint probability distribution distance of the source domain and the target domain in the thermal error experimental data, wherein the step S71 satisfies the relation (1):
Dist(D s ,D t )＝||P(X s )-P(X t )||+||P(y s |X s )-P(y t |X s )|| (1)；
wherein ,P(Xs )、P(X t )、P(y s |X s )、P(y t |X s ) Respectively representing the conditional probability distribution of the source domain characteristic data, the conditional probability distribution of the target domain characteristic data, the edge probability distribution of the source domain thermal error relative to the source domain characteristic data and the edge probability distribution of the target domain thermal error relative to the target domain characteristic data;
specifically, the joint probability distribution distance between the source domain and the target domain in the thermal error experimental data is minimized by simultaneously reducing the conditional probability distribution between the source domain and the target domain and the edge probability distribution between the source domain and the target domain, and the final distance Dist (D s ,D t ) The minimum represents the minimum of the two-domain joint probability distribution.
S72, minimizing the maximum mean difference between the source domain and the target domain, wherein the step S72 satisfies the relation (2):
wherein m, n represent the number of samples in the source domain and the target domain, respectively, and A, T, H represent a transformation matrix, a transpose of the matrix, and a Hilbert space (Hilbert space), respectively; the transformation matrix is used for transforming the source domain data X and the target domain data X i ，X j Mapping to Hilbert space, A T For transpose of transform matrix A, solving Dist in the Hilbert space MMD 。
S73, introducing a regression algorithm kernel function to simplify the relation (2) and obtaining a relation (3):
Dist(D s ,D t )＝tr(A T XM 0 X T A) (3)；
wherein ,M0 A kernel function for the regression algorithm;
s74, associating the regression algorithm kernel function in the relation (3) with the sample numbers of the source domain and the target domain, wherein the step S74 satisfies the relation (4):
s75, use (X) s ,y s ) Training a regression model to obtain pseudo tagsThe regression model is one of support vector regression and random forest regression;
s76, determining an optimized value of the joint distribution adaptive algorithm, wherein the optimized value meets a relation (5):
wherein ：
s77, constructing an optimization target of the joint distribution self-adaptive algorithm, wherein the optimization target meets a relation (6):
wherein ,is a regular term;
and S78, unifying the optimization targets according to the relational expressions (1) to (5) to obtain a relational expression (7):
s79, performing stretching transformation on the relation (7) to obtain a relation (8):
s710, replacing a calculation item in an optimization target by using a Lagrangian multiplier phi, and solving the optimization target, wherein the step S710 satisfies a relation (9):
and S711, iterating the joint distribution self-adaptive algorithm until the optimization target meets a preset iteration precision value.
S8, the source domain characteristic data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
S9, pseudo tag of the target domainAs a thermal error prediction value, the machine tool system is input and compensated to a tool feed amount.
In order to facilitate understanding, the embodiment of the invention also provides a modeling example of a machine tool thermal error prediction method based on joint distribution self-adaption, a multi-source heterogeneous data acquisition platform designed for a machine tool system is shown in fig. 2, data at spindle rotation speed 9000rpm and 5000rpm are acquired on the basis of the acquisition platform of fig. 2, wherein the rotation speed 9000rpm is used as source domain data, the rotation speed 5000rpm is used as target domain data, a regression model established on the source domain data set is migrated to the target domain data set, and table 1 below is used as specific parameters of the source domain and target domain data sets.
TABLE 1 Source Domain and target Domain data set specific parameters
The difference of the data sets at two rotating speeds is shown more clearly, the embodiment of the invention draws the probability distribution diagram of the data based on the table 1 shown in fig. 3, and as can be seen from fig. 3, no obvious correlation exists between the data at two rotating speeds.
Source domain feature data X i After normalization, the source domain characteristic data X is established i And thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (. Cndot.) the use of a catalyst. The source domain regression model in the embodiment of the present invention is used for researching the mapping relationship between the independent variable and the dependent variable, and does not specifically refer to a specific algorithm, and exemplary, the embodiment of the present invention selects deep neural network (Deep Neural Network, DNN) for modeling, and the model parameter visualization is shown in fig. 4.
Subtracting the mean value of the training set of the source domain from the characteristic data of the target domain, dividing the mean value by the standard deviation of the training set, normalizing, and constructing a regression model f of the source domain s (. Cndot.) is directly applied to the feature number of the target domain to obtain the pseudo tag of the target domainThe prediction accuracy obtained by directly using the DNN model can be seen from fig. 5.
To source domain feature data X i Source domain thermal error value y i |e x ,e y ,e z Target domain feature data X j Pseudo tag for target domainJoint distributed adaptive algorithm adaptation constructed by embodiments of the present inventionMatching edge distribution and condition distribution to obtain target and pseudo tag->Taking the value as a thermal error predicted value y of a target domain in a migration task from a source domain to the target domain j The prediction accuracy is shown in fig. 6, and the partial method is shown in fig. 7.
The invention has the beneficial technical effects that the invention provides a machine tool thermal error prediction method based on joint distribution self-adaption, which can be applied to thermal error prediction tasks under other working conditions and equipment by constructing a thermal error regression model under the working conditions and equipment in a source domain, and simultaneously, can be only applied to regression prediction scenes from classification problem scenes by modifying the traditional method of a joint distribution self-adaption algorithm kernel function, and can provide a universal prediction scheme for thermal error data in the field of thermal error in a machine tool system, in a cross-equipment mode, in a multi-direction mode and in spindle knife processing by adapting the classification evaluation index of the algorithm to the regression problem evaluation index.
The embodiment of the invention also provides a machine tool thermal error prediction system based on joint distribution self-adaption, which comprises:
the data acquisition module is used for acquiring thermal error experimental data of the numerical control machine through a machine tool system with a multi-source sensor;
the screening module is used for screening abnormal values of the thermal error experimental data and removing outliers;
the compensation module is used for judging whether a system error exists according to the type of the thermal error experimental data and compensating a correction value for the thermal error realization data with the system error through residual error analysis;
the source domain data processing module is used for dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And thermal error values { y } of the principal axis in three directions of the spatial coordinate axis i |e x ,e y ,e z }；
A source domain attribution module for attributing the source domain characteristic data X i Normalization processing is carried out, and the source domain characteristic data X is established i And the thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)；
The target domain attribution module is used for dividing a target domain thermal error prediction task D from the thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
The algorithm construction module is used for constructing a joint distribution self-adaptive algorithm for regression problems;
an error calculation module for calculating the source domain feature data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
An error feedback module for pseudo-labeling the target domainAs a thermal error prediction value, the machine tool system is input and compensated to a tool feed amount.
The machine tool thermal error prediction system based on the joint distribution self-adaption provided by the embodiment of the invention can realize the steps in the machine tool thermal error prediction method based on the joint distribution self-adaption in the embodiment, can realize the same technical effects, and is not repeated herein with reference to the description in the embodiment.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM) or the like.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
While the embodiments of the present invention have been illustrated and described in connection with the drawings, what is presently considered to be the most practical and preferred embodiments of the invention, it is to be understood that the invention is not limited to the disclosed embodiments, but on the contrary, is intended to cover various equivalent modifications and equivalent arrangements included within the spirit and scope of the appended claims.
Claims (5)
1. The machine tool thermal error prediction method based on the joint distribution self-adaption is characterized by comprising the following steps of:
s1, acquiring thermal error experimental data of a numerical control machine through a machine tool system with a multi-source sensor;
s2, screening abnormal values of the thermal error experimental data, and removing outliers;
s3, judging whether a system error exists according to the type of the thermal error experimental data, and compensating a correction value for the thermal error realization data with the system error through residual error analysis;
s4, dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And thermal error values { y } of the principal axis in three directions of the spatial coordinate axis i |e x ,e y ,e z }；
S5, for the source domain feature data X i Normalization processing is carried out, and the source domain characteristic data X is established i And the thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)；
S6, dividing a target domain thermal error prediction task D from the thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
S7, constructing a joint distribution self-adaptive algorithm for regression problem;
s8, the source domain characteristic data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
S9, pseudo tag of the target domainInputting the machine tool system as a thermal error prediction value, and compensating to a cutter feeding amount;
wherein the joint distribution self-adaptive algorithm comprises a regression algorithm kernel function, and the step S7 comprises the following substeps:
s71, minimizing the joint probability distribution distance of the source domain and the target domain in the thermal error experimental data, wherein the step S71 satisfies the relation (1):
Dist(D s ,D t )＝‖P(X s )-P(X t )‖+‖P(y s |X s )-P(y t |X s )‖ (1)；
wherein ,P(Xs )、P(X t )、P(y s |X s )、P(y t |X s ) Respectively representing the conditional probability distribution of the source domain characteristic data, the conditional probability distribution of the target domain characteristic data, the edge probability distribution of the source domain thermal error relative to the source domain characteristic data and the edge probability distribution of the target domain thermal error relative to the target domain characteristic data;
s72, minimizing the maximum mean difference between the source domain and the target domain, wherein the step S72 satisfies the relation (2):
wherein m and n respectively represent the number of samples in the source domain and the target domain, and A, T, H respectively represent a transformation matrix, a transpose of the matrix, and a hilbert space;
s73, introducing a regression algorithm kernel function to simplify the relation (2) and obtaining a relation (3):
Dist(D s ,D t )＝tr(A T XM 0 X T A) (3)；
wherein ,M0 A kernel function for the regression algorithm;
s74, associating the regression algorithm kernel function in the relation (3) with the sample numbers of the source domain and the target domain, wherein the step S74 satisfies the relation (4):
s75, use (X) s ,y s ) Training a regression model to obtain pseudo tags
S76, determining an optimized value of the joint distribution adaptive algorithm, wherein the optimized value meets a relation (5):
wherein ：
s77, constructing an optimization target of the joint distribution self-adaptive algorithm, wherein the optimization target meets a relation (6):
wherein ,is a regular term;
and S78, unifying the optimization targets according to the relational expressions (1) to (5) to obtain a relational expression (7):
s79, performing stretching transformation on the relation (7) to obtain a relation (8):
s710, replacing a calculation item in an optimization target by using a Lagrangian multiplier phi, and solving the optimization target, wherein the step S710 satisfies a relation (9):
and S711, iterating the joint distribution self-adaptive algorithm until the optimization target meets a preset iteration precision value.
2. The machine tool thermal error prediction method based on joint distribution self-adaption according to claim 1, wherein in the step S2, the method for screening abnormal values of the thermal error experimental data comprises at least one of 3 sigma criteria and box line drawing.
3. The method as claimed in claim 1The machine tool thermal error prediction method based on joint distribution self-adaption is characterized in that the source domain regression model f s (. Cndot.) is one of a support vector regression model, a random forest regression model, an artificial neural network model, and a deep neural network model.
4. The method for predicting thermal errors of a machine tool based on joint distribution adaptation according to claim 1, wherein in step S75, the regression model is one of support vector regression and random forest regression.
5. A machine tool thermal error prediction system implemented based on a joint distribution adaptive machine tool thermal error prediction method according to any one of claims 1-4, characterized in that the machine tool thermal error prediction system comprises:
the data acquisition module is used for acquiring thermal error experimental data of the numerical control machine through a machine tool system with a multi-source sensor;
the screening module is used for screening abnormal values of the thermal error experimental data and removing outliers;
the compensation module is used for judging whether a system error exists according to the type of the thermal error experimental data and compensating a correction value for the thermal error realization data with the system error through residual error analysis;
the source domain data processing module is used for dividing a source domain thermal error prediction task D from the thermal error experimental data s Task D is predicted by the source domain thermal error s Comprising a plurality of source domain feature data X i And thermal error values { y } of the principal axis in three directions of the spatial coordinate axis i |e x ,e y ,e z }；
A source domain attribution module for attributing the source domain characteristic data X i Normalization processing is carried out, and the source domain characteristic data X is established i And the thermal error value { y } i |e x ,e y ,e z Source domain regression model f s (·)；
A target domain normalization module for normalizing the target domain from thePartitioning target domain thermal error prediction task D in thermal error experimental data t And obtain the target domain feature data X j The target domain characteristic data X after normalization processing j Input Source Domain regression model f s (. Cndot.) obtaining target Domain pseudo tag
The algorithm construction module is used for constructing a joint distribution self-adaptive algorithm for regression problems;
an error calculation module for calculating the source domain feature data X i The source domain thermal error value { y } i |e x ,e y ,e z -said target domain feature data X j Thermal error pseudo tag of the target domainAligning edge distribution and conditional distribution through the joint distribution self-adaptive algorithm to obtain a target domain pseudo tag +.>
An error feedback module for pseudo-labeling the target domainAs a thermal error prediction value, the machine tool system is input and compensated to a tool feed amount.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310189184.XA CN116224905B (en) | 2023-03-02 | 2023-03-02 | Machine tool thermal error prediction method and system based on joint distribution self-adaption |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310189184.XA CN116224905B (en) | 2023-03-02 | 2023-03-02 | Machine tool thermal error prediction method and system based on joint distribution self-adaption |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116224905A CN116224905A (en) | 2023-06-06 |
| CN116224905B true CN116224905B (en) | 2023-10-20 |
Family
ID=86584022
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310189184.XA Active CN116224905B (en) | 2023-03-02 | 2023-03-02 | Machine tool thermal error prediction method and system based on joint distribution self-adaption |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116224905B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101436057A (en) * | 2008-12-18 | 2009-05-20 | 浙江大学 | Numerical control machining tool heat error Bayes network compensation method |
| CN108803486A (en) * | 2018-08-16 | 2018-11-13 | 重庆理工大学 | Numerical control machining tool heat error prediction based on deep learning network in parallel and compensation method |
| CN109146209A (en) * | 2018-11-02 | 2019-01-04 | 清华大学 | Machine tool spindle thermal error prediction technique based on wavelet neural networks of genetic algorithm |
| CN109240204A (en) * | 2018-09-30 | 2019-01-18 | 山东大学 | A kind of numerical control machining tool heat error modeling method based on two-step method |
| CN115718466A (en) * | 2022-11-21 | 2023-02-28 | 广东工业大学 | Digital twin workshop fault prediction method based on random forest and analytic hierarchy process |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080256008A1 (en) * | 2007-03-31 | 2008-10-16 | Mitchell Kwok | Human Artificial Intelligence Machine |
| US11769011B2 (en) * | 2020-12-18 | 2023-09-26 | Google Llc | Universal language segment representations learning with conditional masked language model |
-
2023
- 2023-03-02 CN CN202310189184.XA patent/CN116224905B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101436057A (en) * | 2008-12-18 | 2009-05-20 | 浙江大学 | Numerical control machining tool heat error Bayes network compensation method |
| CN108803486A (en) * | 2018-08-16 | 2018-11-13 | 重庆理工大学 | Numerical control machining tool heat error prediction based on deep learning network in parallel and compensation method |
| CN109240204A (en) * | 2018-09-30 | 2019-01-18 | 山东大学 | A kind of numerical control machining tool heat error modeling method based on two-step method |
| CN109146209A (en) * | 2018-11-02 | 2019-01-04 | 清华大学 | Machine tool spindle thermal error prediction technique based on wavelet neural networks of genetic algorithm |
| CN115718466A (en) * | 2022-11-21 | 2023-02-28 | 广东工业大学 | Digital twin workshop fault prediction method based on random forest and analytic hierarchy process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116224905A (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210004369A1 (en) | Systems and methods for searching a machining knowledge database | |
| Zhang et al. | Build orientation optimization for multi-part production in additive manufacturing | |
| Zhang et al. | Deep learning-enabled intelligent process planning for digital twin manufacturing cell | |
| Zhang et al. | A statistical method for build orientation determination in additive manufacturing | |
| CN109472057B (en) | Product processing quality prediction device and method based on cross-process implicit parameter memory | |
| Zhu et al. | Convolutional neural network for geometric deviation prediction in additive manufacturing | |
| CN115146875B (en) | Historical data based process parameter recommendation method, device, system and medium | |
| Fountas et al. | An integrated framework for optimizing sculptured surface CNC tool paths based on direct software object evaluation and viral intelligence | |
| Zellinger et al. | Multi-source transfer learning of time series in cyclical manufacturing | |
| Lu et al. | Tool path generation via the multi-criteria optimisation for flat-end milling of sculptured surfaces | |
| CN115586749B (en) | Workpiece machining track control method based on machine vision and related device | |
| Fountas et al. | Globally optimal tool paths for sculptured surfaces with emphasis to machining error and cutting posture smoothness | |
| CN111258984B (en) | Product quality end-edge-cloud collaborative forecasting method under industrial big data environment | |
| JP4653547B2 (en) | Apparatus and method for analyzing relation between operation and quality in manufacturing process, computer program, and computer-readable recording medium | |
| Bui et al. | Analyzing nonparametric part-to-part variation in surface point cloud data | |
| CN116224905B (en) | Machine tool thermal error prediction method and system based on joint distribution self-adaption | |
| Abdikerimova et al. | Applying textural Law’s masks to images using machine learning. | |
| CN108537249B (en) | Industrial process data clustering method for density peak clustering | |
| Msakni et al. | Using machine learning prediction models for quality control: a case study from the automotive industry | |
| CN112465062A (en) | Clustering method based on manifold learning and rank constraint | |
| CN116700149A (en) | Self-adaptive filtering optimization method for numerical control machining tool path | |
| CN109344448B (en) | fuzzy-FQD-based helical bevel gear shape collaborative manufacturing optimization method | |
| Haq et al. | TransNAS-TSAD: Harnessing Transformers for Multi-Objective Neural Architecture Search in Time Series Anomaly Detection | |
| CN114936279A (en) | Unstructured chart data analysis method for collaborative manufacturing enterprise | |
| Kuchuganov et al. | Clustering algorithm for a set of machine parts on the basis of engineering drawings |
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 |