CN115576267B - Wheel hub machining dimension error correction method based on digital twin - Google Patents
Wheel hub machining dimension error correction method based on digital twin Download PDFInfo
- Publication number
- CN115576267B CN115576267B CN202211385897.5A CN202211385897A CN115576267B CN 115576267 B CN115576267 B CN 115576267B CN 202211385897 A CN202211385897 A CN 202211385897A CN 115576267 B CN115576267 B CN 115576267B
- Authority
- CN
- China
- Prior art keywords
- model
- error correction
- processing
- constructing
- dimensional
- 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
Images
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
- G05B2219/35—Nc in input of data, input till input file format
- G05B2219/35408—Calculate new position data from actual data to compensate for contour error
Abstract
The invention discloses a hub machining dimension error correction method based on digital twin, which comprises the steps of firstly constructing a hub machining digital twin system to realize real-time monitoring of production state and remote control of a machine tool; secondly, constructing a processing rule knowledge graph, and realizing precipitation and standardization of error correction processing experience knowledge; and then, a size error correction model based on differential evolution is provided, and an error correction problem optimal solution is calculated. The invention relates to the technical field of dimensional error correction and digital twinning. According to the hub machining dimension error correction method based on digital twin, the hub machining dimension error correction method based on digital twin is practically applied to a certain hub machining production line, and experimental verification is carried out according to data accumulated in production activities, so that the method is proved to effectively improve the accuracy of machining dimension error correction.
Description
Technical Field
The invention relates to the technical field of dimensional error correction and digital twin, in particular to a method for correcting machining dimensional error of a hub based on digital twin.
Background
The manufacturing industry rapidly develops to digitization and intellectualization. In recent years, the sales volume of Chinese aluminum alloy hubs is more than 60% of the world, and the Chinese aluminum alloy hubs become the major countries for manufacturing aluminum alloy hubs, and how to accelerate the transition from the major countries for manufacturing aluminum alloy hubs to the major countries for manufacturing aluminum alloy hubs by utilizing new generation information technology becomes key. The aluminum alloy hub production mainly comprises the working procedures of smelting, die casting, heat treatment, machining, coating and the like. Currently, the hub production line is basically popular with automatic equipment such as numerically-controlled machine tools, industrial robots and the like. However, some critical processes still rely on manual methods, such as calculation and distribution of machining dimensional error correction schemes, which make the production process difficult to standardize and result in lower accuracy. Therefore, how to improve the standardization and the intelligentization degree of the machining dimension error correction becomes a key link for realizing the intelligent manufacturing of the hub industry.
The digital twin technology integrates multiple physical, multi-scale and multidisciplinary attributes, has the characteristics of real-time synchronization, faithful mapping and high fidelity, and is an effective means for realizing information physical fusion. The digital twin technology provides a brand new solution for constructing a digital and intelligent production line, and has been paid more attention to the industrial field. Article [ Zhang Lei et al ] method for suppressing contour error of a numerically controlled machine tool with multiple axes based on digital twinning [ J ]. Computer integrated manufacturing system, 2021, 27 (12): 3391-3402. Aiming at the problem of difficult control of contour errors in the high-speed and high-precision machining process of a numerical control machine tool, a contour error suppression method based on digital twinning is provided, and interpolation control of a multi-axis feeding system under time-varying motion control parameters is realized; article [ Chen Xuan et al, circuit breaker flexible assembly shop digital twin system design [ J ]. Computer engineering and applications, 2022, 58 (14): the flexible automatic workshop assembly scheme of the circuit breaker based on the multi-robot motion control is provided by combining the digital twin technology with 245-257, and the production line structure and the operation method of the assembly workshop of the circuit breaker are optimized. Digital twinning technology is one of the reliable ways to solve the practical problems in various fields of industry.
Disclosure of Invention
(one) solving the technical problems
Aiming at the defects of the prior art, the invention provides a hub machining dimension error correction method based on digital twinning, which constructs a hub machining digital twinning system to realize real-time monitoring and remote control of equipment states; meanwhile, a size error correction model based on differential evolution is established, and accuracy of machining size error correction is effectively improved.
(II) technical scheme
In order to achieve the above purpose, the invention is realized by the following technical scheme: a hub machining dimension error correction method based on digital twinning specifically comprises the following steps:
a1, machining a digital twin system total frame: constructing a hub machining digital twin system frame by using a reference numeral twin five-dimensional model, wherein the hub machining digital twin system frame comprises a physical entity, a twin model, twin data, functional services and transmission connection among all parts;
a2, constructing a twin model: firstly, constructing a virtual device model by using 3dsMAX and SolidWorks, constructing a twin virtual scene by using unite 3D, defining a behavior model of the virtual model by using a finite state machine, and writing a corresponding script;
a3, constructing a transmission communication framework: establishing a data transmission communication foundation between entity equipment and a twin model based on an OPCUA protocol, and constructing a twin database by using MySQL;
a4, processing rule knowledge graph construction: the processing rule knowledge graph is used for defining the influence relation between a processing tool and the size characteristics and guiding the scheme calculation of error correction, and the construction process comprises the steps of constructing rule triples and fitting relation coefficients based on a ridge regression model;
a5, building a size error correction model based on differential evolution: firstly, establishing a mathematical model according to the dimensional error correction problem, then solving an optimal solution by using a differential evolution algorithm, simultaneously processing model constraint by using a penalty term, and finally generating a correction scheme according to the optimal solution.
Preferably, the step A4 of constructing the knowledge graph model of the processing rule includes the following steps:
t1, constructing a rule triplet: extracting cutter and size characteristic information related to rules from a document to form cutter entity nodes and size characteristic entity nodes, establishing a rule chain table of entity node relations according to mechanism knowledge, and constructing a rule triplet < cutter C, influence and size characteristic F > according to the rule chain table;
t2, fitting relation coefficient of ridge regression model: calculating a relation coefficient between the cutter and the size characteristic in the rule triplet by using a ridge regression model;
and T3, constructing a knowledge graph: combining the rule triples and the relation coefficients obtained in the step T1 and the step T2 into a processing rule knowledge graph and storing the processing rule knowledge graph into a database, and performing graph visualization operation by using Neo4 j.
Preferably, the process of fitting the relationship coefficient by the ridge regression model in the step T2 is as follows:
ith dimensional characteristic measurement y i Change value deltay of (a) i By adjusting the value k by the tool j And corresponding relation coefficient e ji Determining, wherein the calculation formula is as follows:
wherein h is the total number of tools involved in the adjustment, k j Represents the adjustment value of the j-th cutter, e ji Representing the coefficient of relationship between the jth tool and the ith dimensional feature, if the dimensional feature y i M copies of the relevant recorded data, an equation set is constructed:
wherein k is jt Representing the adjustment value of the jth cutter in the t-th recorded data, constructing a linear regression model by the equation set as follows:
y＝Kβ+ε；
where y= (Δy) i1 ，Δy i2 ，…，Δy im ) T ，β＝(e 1i ，e 2i ，…，e hi ) T ，ε＝(ε 1 ，ε 2 ，…，ε m ) T Representing a random error vector with a mean value of 0,
at K T When K is reversible, an estimated beta= (K) T K) -1 K T y；
In actual production, the adjustment of a particular set of tools needs to be consistent, i.e. there are multiple rows k jt The same situation leads to K T K is irreversible, and in order to solve the above-mentioned multiple collinearity problem, the coefficient is obtained by regression estimation of the following:
β＝(K T K+λI) -1 K T y；
wherein I is an identity matrix, lambda is a ridge parameter, and is determined by a cross-validation method.
Preferably, the step A5 of constructing the dimension error correction model based on differential evolution includes the following steps:
f1, establishing a mathematical model according to the dimensional error correction problem: solving error correction scheme belongs to single-target optimization problem, namely solving tool adjustmentInteger k vector, to make the expected value of the dimension feature
Upper limit u i And lower limit l i Between, simultaneously make->
Near standard value c i I=1, 2, …, n, the problem is known as the initial measurement m of the current n dimensional features i And the relation coefficient e of the cutter and each dimension characteristic ji ；
The planning model objective function is:
where n is the total number of dimensional features involved in the adjustment, c i Represents the i-th dimension characteristic standard value,
the expected value of the ith dimension characteristic is represented by the following calculation formula:
wherein m is i Representing the initial measurement of the ith dimensional feature, h is the total number of tools involved in the adjustment, k j Represents the adjustment value of the j-th cutter, e ji The relation coefficient of the jth cutter and the ith dimension characteristic is represented, meanwhile, according to the actual condition of the production line, the number of the adjusted cutters is reduced as much as possible, and the corrected objective function can be obtained by combining the two formulas:
wherein a is a constant value, and the expected size characteristic is assigned according to actual conditions
Needs to be limited within the upper and lower limits due toThis constraint is expressed as:
wherein l i Represents the ith dimension feature lower limit, u i Represents the upper limit of the ith dimensional feature;
e2, solving a differential evolution algorithm;
e3, processing constraint by using penalty term: when the planning model has no solution, constraint items in the model are converted into objective functions by adding penalty items, so that the purposes of reducing model constraint and further increasing the feasible region of the solution are achieved. The specific method is that the objective function of the planning problem model is rewritten as:
wherein f (x) represents the original objective function, p i (x) And q i (x) Respectively representing the constraint conditions of the original size lower limit and the size upper limit, and sigma represents the penalty factor sigma i A component vector representing the part of the current solution exceeding the constraint is calculated into the calculation of the objective function, thereby converting the constraint problem into an unconstrained problem;
according to the above formula, constructing a planning model with penalty terms, and combining the formula in the step E2 to obtain a new objective function formula as follows:
wherein sigma i For the penalty factor, its value represents the degree of influence of the constraint on the model solving process, each dimensional feature y i Corresponds to a penalty factor sigma i ，σ i The smaller the feature y i The less the relevant constraints have on the model solution.
Preferably, the idea source of the differential evolution algorithm in the step E2 is a genetic algorithm, which is a continuous variable optimization algorithm for solving the overall optimal solution in the multidimensional space.
Preferably, the specific steps of the differential evolution algorithm are as follows:
u1, initializing a candidate solution population by using a real number coding mode;
u2, evaluating and analyzing the current population, including recording the optimal individuals and calculating individual fitness by using linear scale transformation;
u3, judging whether a stopping condition is met or not, including whether the maximum evolution algebra is exceeded or not, whether the population with the evolution stagnation exceeds the upper limit or not, if so, stopping the algorithm, and taking the optimal individual of the last generation as an optimal solution to be output;
u4, carrying out differential mutation on the current population, and obtaining variant individuals by using a DE/best/1 strategy, wherein a differential mutation scaling factor takes 0.5;
u5, merging the current population and the variant individuals, and generating new crossed individuals by using an exponential crossing method;
u6, adopting elite retention strategy between the current population and the iterative population to obtain a new generation population;
and U7, entering the next round of evolution, and returning to the step U2.
Preferably, in order to determine σ i To evaluate the importance of each dimensional feature in error correction using alpha e1, 10]The integer value of (a) represents the importance of the dimensional feature, when alpha i When not equal to 10, let sigma i ＝α i . When alpha is i When=10, the i-th dimension feature is represented as a critical dimension, let σ i =0 and maintain y in the newly constructed planning model i ≥l i And y i ≤u i Is a constraint on (c).
Preferably, the specific transformation mode of the behavior model of the virtual processing equipment model defined by using the finite state machine in the step A2 is as follows, wherein S i Representing model state, e i To trigger an event, a i In response to the action:
v1, when the system starts to operate, the initial state of the processing equipment is an idle standby state S 0 ；
V2, triggering e when the hub blank reaches the designated machining position 1 The processing equipment enters a normal operation state S 1 Milling operation a is started 1 ；
V3, if the processing is successfully completed, triggering e 2 The processing equipment enters a processing completion state S 3 Send a processing completion signal a 2 To inform the robot to perform the picking operation and receive the picking completion signal e 3 After that, the processing equipment returns to the idle waiting state S 0 Transmitting a waiting blank signal a 3 ；
V4, if abnormal conditions (such as equipment alarm, manual equipment suspension and the like) occur in the processing process, the processing equipment receives an equipment abnormal signal e 4 The processing is paused, the equipment enters into interrupt, and an exception handling signal a is sent 4 Informing the server side of processing according to the abnormal information, wherein the processing equipment enters a blocking waiting state S 2 ；
V5, if the exception is successfully processed, receiving an exception processing completion signal e 5 Continuing the machining to resume the state a before interruption 5 The processing equipment resumes the normal operation state S 1 Otherwise, when the abnormality cannot be handled and the current processing needs to be interrupted, sending a pickup signal e 2 And receives the signal e after the completion of the picking 3 After that, the processing equipment enters an idle standby state S 0 。
(III) beneficial effects
The invention provides a hub machining dimension error correction method based on digital twinning. Compared with the prior art, the method has the following beneficial effects:
(1) The method for correcting the machining dimension error of the hub based on the digital twin comprises the steps of firstly constructing a hub machining digital twin system to realize real-time monitoring of the production state and remote control of a machine tool; secondly, constructing a processing rule knowledge graph, and realizing precipitation and standardization of error correction processing experience knowledge; and then, a size error correction model based on differential evolution is provided, and an error correction problem optimal solution is calculated. The actual application of the production line data of a certain hub proves that the method effectively improves the accuracy of correction of the machining dimension errors.
(2) Compared with the traditional error correction method of the hub machining production line, the digital twin-based hub machining dimensional error correction method has the advantages that: the hub machining digital twin system is constructed, the system can realize real-time monitoring of the production state through real-time synchronization of a virtual model and production equipment by virtual mapping and virtual control, and the remote control is realized by completing data instruction issuing to the production equipment through a system communication framework, so that the production efficiency is improved; the processing rule knowledge graph is constructed, processing rule triples are constructed by combining experience knowledge and historical data, rule relation coefficients are fitted, and then the processing rule knowledge graph is constructed, so that knowledge experience related to error correction processing rules which cannot be effectively mastered by an enterprise originally is precipitated, a standardized method is provided for continuous accumulation of rule knowledge, and core competitiveness of the enterprise is improved; the dimensional error correction model based on differential evolution is provided, the problem is modeled through mathematical analysis, and the optimal solution of the problem is calculated by using a solution algorithm based on differential evolution, so that the accuracy of the error correction problem is effectively improved.
(3) According to the hub machining dimension error correction method based on the digital twin, the hub machining dimension error correction method based on the digital twin is practically applied to a certain hub machining production line, and experimental verification is carried out according to data accumulated in production activities, so that the method is proved to effectively improve the accuracy of machining dimension error correction.
Drawings
FIG. 1 is a frame diagram of a hub machining digital twin system of the present invention;
FIG. 2 is a diagram of a process plant behavior model according to the present invention;
FIG. 3 is a diagram of a data connection communication architecture of the present invention;
FIG. 4 is a diagram of the process of constructing the knowledge graph of the processing rule of the invention;
FIG. 5 is a flow chart of a solution of a dimensional error correction model based on differential evolution according to the present invention;
FIG. 6 is a comparison of the verification results of an example error correction method of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1-6, the embodiment of the present invention provides a technical solution: a hub machining dimension error correction method based on digital twinning specifically comprises the following steps:
a1: the reference numeral twin five-dimensional model is used for constructing a hub machining digital twin system frame as shown in fig. 1, wherein the system frame comprises a physical entity, a twin model, twin data, functional services and transmission connection among all parts; the main functions of the machine tool are two aspects, on one hand, real-time acquisition and monitoring of the production state of the production line are realized, instruction data are issued and written into the machine tool so as to control the machine tool, and virtual mapping and virtual control are realized, so that the production line efficiency is improved; on the other hand, the optimal solution is calculated and corrected in real time according to the product size error, so that virtual pre-compaction is realized, and the product size accuracy is improved.
A2: constructing a twin model, namely firstly constructing a virtual device model by using 3dsMAX and SolidWorks and constructing a twin virtual scene by using unity 3D;
a finite state machine is then used to define a behavior model of the virtual model and write the corresponding script. The invention uses a finite state machine to build a twin behavior model, describes finite state sequences experienced by each twin model in the life cycle, and defines response external events, thereby completing the production behavior of the twin model mapping physical entities. The process equipment behavior model is shown in FIG. 2, wherein S i Representing model state, e i To trigger an event, a i In response to the action, the state transition process is as follows:
(1-1) when the system starts to operate, the initial state of the processing equipment is an idle standby state S 0 ；
(1-2) triggering e when the hub blank reaches the designated machining position 1 The processing equipment enters a normal operation state S 1 Milling operation a is started 1 ；
(1-3) if the processing is successfully completed, trigger e 2 The processing equipment enters a processing completion state S 3 Send a processing completion signal a 2 To inform the robot to take the workpiece; receiving the pick-up completion signal e 3 After that, the processing equipment returns to the idle waiting state S 0 Transmitting a waiting blank signal a 3 ；
(1-4) if abnormal conditions (such as equipment alarm, manual equipment suspension, etc.) occur in the processing process, the processing equipment receives an equipment abnormality signal e 4 The processing is paused, the equipment enters into interrupt, and an exception handling signal a is sent 4 Informing the server side of processing according to the abnormal information, wherein the processing equipment enters a blocking waiting state S 2 ；
(1-5) if the exception is successfully handled, receiving an exception handling completion signal e 5 Continuing the machining to resume the state a before interruption 5 The processing equipment resumes the normal operation state S 1 The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, when the abnormality cannot be handled and the current processing needs to be interrupted, sending a pickup signal e 2 And receives the signal e after the completion of the picking 3 After that, the processing equipment enters an idle standby state S 0 ；
A3: a data connection communication architecture is built, as shown in fig. 3. Based on the OPCUA protocol, a data transmission communication foundation between the entity equipment and the twin model is established, and the equipment such as a sensor collects real-time data and uploads the real-time data to the OPCUA server. The OPCUA server comprises a plurality of nodes, and the nodes are in one-to-one correspondence with the real-time data of the equipment and are accessed through unique addresses. The functional server reads, writes and monitors real-time data of the entity equipment through the point location address table. And constructing a MySQL database for storing production records, detection data and processing rule knowledge graph data. And a data preprocessing module is arranged in the connection interface and is responsible for checking the integrity of the read data from the OPCUA server and correcting error data. The preprocessed data is stored into a database through a MySQL interface, and the OPCUA server is not directly communicated with the database;
a4: the processing rule knowledge graph is constructed, and the construction process is shown in fig. 4. The processing rule knowledge graph defines the influence relation between the processing cutter and the size characteristics, and is used for guiding the scheme calculation of error correction, and the construction process is as follows;
(2-1) constructing a rule triplet: extracting cutter and size characteristic information related to a rule from a document to form cutter entity nodes and size characteristic entity nodes; establishing a rule chain table of entity node relations according to the mechanism knowledge; constructing a rule triplet < cutter C, influence and size characteristic F > according to the rule chain table;
(2-2) fitting relationship coefficients to a ridge regression model:
the relationship coefficients between the tool and the dimensional features in the rule triples were calculated using a ridge regression model, the sign definition of which is shown in table 1.
TABLE 1 regression model symbol definition table
Δy i From k j And corresponding relation coefficient e ji Determining, wherein the calculation formula is as follows:
if the dimension is characteristic y i M copies of the relevant recorded data, a system of equations can be constructed:
the linear regression model is constructed from the above equation set as:
y=kβ+ε, where y= (Δy) i1 ，Δy i2 ，…，Δy im ) T ，β＝(e 1i ，e 2i ，…，e hi ) T ，ε＝(ε 1 ，ε 2 ，…，ε m ) T Representing a random error vector with a mean value of 0,
at K T When K is reversible, an estimated beta= (K) T K) - 1 K T y。
In actual production, the adjustment of a particular set of tools needs to be consistent, i.e. there are multiple rows k jt The same situation leads to K T K is irreversible. To solve the above-described multiple collinearity problem, coefficients are found by regression estimation of the following terms:
β＝(K T K+λI) -1 K T y，
wherein, I is an identity matrix, lambda is a ridge parameter, and is determined by a cross validation method;
the specific steps for constructing the fitting relation coefficient of the ridge regression model are as follows:
step 4, obtaining the coefficient of the regression model of the current ridge by using the optimal lambda value, namely the dimension characteristic y i Relation coefficients of all relevant tools;
(2-3) constructing a knowledge graph: combining the rule triples and the relation coefficients obtained in (2-1) and (2-2) into a processing rule knowledge graph and storing the processing rule knowledge graph into a database, and performing graph visualization operation by using Neo4 j.
A5: and constructing a size error correction model based on differential evolution, wherein a model solving flow is shown in fig. 5. Firstly, establishing a mathematical model according to the dimensional error correction problem, then solving an optimal solution by using a differential evolution algorithm, simultaneously processing model constraint by using a penalty term, and finally generating a correction scheme according to the optimal solution.
3-1) establishing a mathematical model according to the dimensional error correction problem:
solving error correction scheme belongs to single-objective optimization problem, i.e. solving k vector of tool adjustment value to make expected value of dimension characteristic
Upper limit u i And lower limit l i Between, simultaneously make->
Near standard value c i I=1, 2, …, n. The problem is known as the initial measurement m of the current n dimensional features i And the relation coefficient e of the cutter and each dimension characteristic ji . The parameter symbol definitions used in the error correction problem modeling process are shown in table 2.
Table 2 error correction problem modeling symbol definition table
The planning model objective function is:
wherein the method comprises the steps of
The calculation formula of (2) is as follows:
meanwhile, according to the actual conditions of the production line, the number of the adjusted cutters is reduced as much as possible, and the corrected objective function can be obtained by combining the two formulas:
wherein a is a constant value and is assigned according to actual conditions. Expected dimensional characteristics
It is required to be limited within the upper and lower limits, so the constraint is expressed as:
3-2) solving a differential evolution algorithm: the differential evolution algorithm thought source is a genetic algorithm, and is a continuous variable optimization algorithm for solving the overall optimal solution in the multidimensional space. The differential evolution algorithm comprises the following specific steps:
step 4, carrying out differential mutation on the current population, and obtaining variant individuals by using a DE/best/1 strategy, wherein a differential mutation scaling factor is 0.5;
(3-3) processing the constraint using a penalty term:
when the planning model has no solution, constraint items in the model are converted into objective functions by adding penalty items, so that the purposes of reducing model constraint and further increasing the feasible region of the solution are achieved. The specific method is that the objective function of the planning problem model is rewritten as:
wherein f (x) represents the original objective function, p i (x) And q i (x) Respectively representing the constraint conditions of the original size lower limit and the size upper limit, and sigma represents the penalty factor sigma i A component vector. The formula represents that the part of the current solution exceeding the constraint is calculated into the calculation of the objective function, thereby converting the constraint problem into the unconstrained problem.
From the above equation, a planning model with penalty terms may be constructed. Combining the formulas in (3-2) to obtain a new target function formula as follows:
wherein sigma i The penalty factor may be a value that represents the extent to which the constraint affects the model solution process. Each dimension feature y i Corresponds to a penalty factor sigma i ，σ i The smaller the feature y i The less the relevant constraints have on the model solution.
To determine sigma i To evaluate the importance of each dimensional feature in error correction. Alpha epsilon [1, 10 ] is used herein]The integer value of (a) represents the importance of the dimensional feature, when alpha i When not equal to 10, let sigma i ＝α i . When alpha is i When=10, the i-th dimension feature is represented as a critical dimension, let σ i =0 and maintain y in the newly constructed planning model i ≥l i And y i ≤u i Is a constraint on (c).
The specific steps for handling constraints using penalty terms are as follows:
step 4, if the algorithm still cannot solve the feasible solution, describing that constraint needs to be processed continuously, and returning to the step 1; if a feasible solution is obtained, evaluating the feasible solution, and predicting each dimension characteristic after error correction by using the solution according to a processing rule knowledge graph;
The test verification of the hub machining dimension error correction method based on digital twinning provided by the invention:
1. description of data
The experimental data set is measurement data for correcting the dimensional error in the actual production of a certain hub production line, and comprises 80 pieces of data accumulated by using a traditional manual method and 70 pieces of data accumulated after using the digital twin system. Each piece of data consists of three parts: and correcting the detection results of the various dimensional characteristics of the sample before correcting, and correcting the adjustment values of all the machining tools and detecting the sample again after correction. The data is classified by hub type, each hub type containing different dimensional characteristics.
2. Experimental procedure
Experiment-relationship coefficient fitting verification based on ridge regression model
The mean square error (Mean Squared Error, MSE) and R are used herein 2 As a criterion for evaluating the regression model, the closer the MSE is to 0, R 2 The closer to 1 represents the higher the model fitness. The experimental data were randomly divided into training and test sets at a ratio of 3:1, and the evaluation index results were calculated as shown in Table 3.
TABLE 3 evaluation index of ridge regression model
As can be seen from the table, the average MSE of the ridge regression model fit on the test set and training set is 0.116, and the average R 2 0.9425, the fitted processing rule relation coefficient is proved to have higher accuracy, and a condition basis can be provided for the error correction model.
Experiment two error correction method example verification
In the machining production of hubs, the error correction effect is usually evaluated using the difference between the actual measured value and the standard value, i.e., the closer the difference is to 0, the higher the correction accuracy. F (F) c The sum of the differences is represented by the following formula:
where n is the total number of dimensional features of the currently evaluated hub, m i Representing the actual measurement of the ith dimensional feature, c i The standard value of the ith dimension feature is indicated. F after error correction by using an artificial method and a digital twin system is calculated respectively according to the method c Values, results are shown in FIG. 6. The vertical axis in FIG. 6 represents F after error correction c The smaller the value, i.e., the closer the correction result is to the standard value, the higher the accuracy of the representative error correction. The calculation formula of the precision lifting quantity is as follows:
wherein F is cM Representing F corrected by artificial means c Value of F cDT Representing F after digital twin system correction c Values. As can be seen from fig. 6, the corrected F of all types of hubs when the digital twin system is used for error correction c The values are all closer to 0, i.e. the dimensional correction accuracy is higher. F corrected by manual method c The total average value is 5.997, and F is corrected by a digital twin system c The total average value is 3.3798, and the accuracy is improvedThe calculation formula improves the accuracy by 43.64%. Therefore, the error correction method based on digital twinning has higher correction accuracy compared with the traditional manual method.
And all that is not described in detail in this specification is well known to those skilled in the art.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, 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.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (7)
1. A hub machining dimension error correction method based on digital twinning is characterized by comprising the following steps of: the method specifically comprises the following steps:
a1, machining a digital twin system total frame: constructing a hub machining digital twin system frame by using a reference numeral twin five-dimensional model, wherein the hub machining digital twin system frame comprises a physical entity, a twin model, twin data, functional services and transmission connection among all parts;
a2, constructing a twin model: firstly, constructing a virtual device model by using 3dsMAX and SolidWorks, constructing a twin virtual scene by using unite 3D, defining a behavior model of the virtual model by using a finite state machine, and writing a corresponding script;
a3, constructing a transmission communication framework: establishing a data transmission communication foundation between entity equipment and a twin model based on an OPCUA protocol, and constructing a twin database by using MySQL;
a4, processing rule knowledge graph construction: the processing rule knowledge graph is used for defining the influence relation between a processing tool and the size characteristics and guiding the scheme calculation of error correction, and the construction process comprises the steps of constructing rule triples and fitting relation coefficients based on a ridge regression model; the construction of the processing rule knowledge graph model comprises the following steps:
t1, constructing a rule triplet: extracting cutter and size characteristic information related to rules from a document to form cutter entity nodes and size characteristic entity nodes, establishing a rule chain table of entity node relations according to mechanism knowledge, and constructing a rule triplet < cutter C, influence and size characteristic F > according to the rule chain table;
t2, fitting relation coefficient of ridge regression model: calculating a relation coefficient between the cutter and the size characteristic in the rule triplet by using a ridge regression model;
and T3, constructing a knowledge graph: combining the rule triples and the relation coefficients obtained in the step T1 and the step T2 into a processing rule knowledge graph and storing the processing rule knowledge graph into a database, and performing graph visualization operation by using Neo4 j;
a5, building a size error correction model based on differential evolution: firstly, establishing a mathematical model according to the dimensional error correction problem, then solving an optimal solution by using a differential evolution algorithm, simultaneously processing model constraint by using a penalty term, and finally generating a correction scheme according to the optimal solution.
2. The method for correcting the machining dimensional error of the hub based on digital twinning according to claim 1, wherein the method comprises the following steps: the process of fitting the relationship coefficient by the ridge regression model in the step T2 is as follows:
ith dimensional characteristic measurement y i Change value deltay of (a) i By adjusting the value k by the tool j And corresponding relation coefficient e ji Determining, wherein the calculation formula is as follows:
wherein h is the total number of tools involved in the adjustment, k j Represents the adjustment value of the j-th cutter, e ji Representing the coefficient of relationship between the jth tool and the ith dimensional feature, if the dimensional feature y i M copies of the relevant recorded data, an equation set is constructed:
wherein k is jt Representing the adjustment value of the jth cutter in the t-th recorded data, constructing a linear regression model by the equation set as follows:
y＝Kβ+ε；
where y= (Δy) i1 ，Δy i2 ，…，Δy im ) T ，β＝(e 1i ，e 2i ，…，e hi ) T ，ε＝(ε 1 ，ε 2 ，…，ε m ) T Representing a random error vector with a mean value of 0,
at K T When K is reversible, an estimated beta= (K) T K) -1 K T y；
In actual production, the adjustment of a particular set of tools needs to be consistent, i.e. there are multiple rows k jt The same situation leads to K T K is irreversible, in order to solve the multiple collinearity problem, the coefficients are obtained by regression estimation of the following terms:
β＝(K T K+λI) -1 K T y；
wherein I is an identity matrix, lambda is a ridge parameter, and is determined by a cross-validation method.
3. The method for correcting the machining dimensional error of the hub based on digital twinning according to claim 1, wherein the method comprises the following steps: the step A5 is characterized in that the step of constructing the size error correction model based on differential evolution comprises the following steps:
e1, establishing a mathematical model according to the dimensional error correction problem: solving error correction scheme belongs to single-objective optimization problem, i.e. solving k vector of tool adjustment value to make expected value of dimension characteristic
Upper limit u i And lower limit l i Between, simultaneously make->
Near standard value c i I=1, 2, …, n, the problem is known as the initial measurement m of the current n dimensional features i And the relation coefficient e of the cutter and each dimension characteristic ji ；
The planning model objective function is:
where n is the total number of dimensional features involved in the adjustment, c i Represents the i-th dimension characteristic standard value,
the expected value of the ith dimension characteristic is represented by the following calculation formula:
wherein m is i Representing the initial measurement of the ith dimensional feature, h is the total number of tools involved in the adjustment, k j Represents the adjustment value of the j-th cutter, e ji The relation coefficient of the jth cutter and the ith dimension characteristic is represented, meanwhile, according to the actual condition of the production line, the number of the adjusted cutters is reduced as much as possible, and the corrected objective function can be obtained by combining the two formulas:
wherein a is a constant value, and t is assigned according to actual conditions a Indicating the number of tools with an adjustment value other than 0, the desired dimensional characteristics
It is required to be limited within the upper and lower limits, so the constraint is expressed as:
wherein l i Represents the ith dimension feature lower limit, u i Represents the upper limit of the ith dimensional feature;
e2, solving a differential evolution algorithm;
e3, processing constraint by using penalty term: when the planning model has no solution, constraint terms in the model are converted into objective functions by adding penalty terms, and the specific method is that the objective functions of the planning problem model are rewritten as follows:
wherein f (x) represents the original objective function, p i (x) And q i (x) Respectively representing the constraint conditions of the original size lower limit and the size upper limit, and sigma represents the penalty factor sigma i A component vector representing the part of the current solution exceeding the constraint is calculated into the calculation of the objective function, thereby converting the constraint problem into an unconstrained problem;
according to the above formula, constructing a planning model with penalty terms, and combining the formula in the step E2 to obtain a new objective function formula as follows:
wherein sigma i For the penalty factor, its value represents the degree of influence of the constraint on the model solving process, each dimensional feature y i Corresponds to a penalty factor sigma i ，σ i The smaller the feature y i The less the relevant constraints have on the model solution.
4. A method for correcting machining dimensional errors of a hub based on digital twinning according to claim 3, wherein: the idea source of the differential evolution algorithm in the step E2 is a genetic algorithm, and the differential evolution algorithm is a continuous variable optimization algorithm used for solving the overall optimal solution in the multidimensional space.
5. The method for correcting the machining dimensional error of the hub based on digital twinning according to claim 4, wherein the method comprises the following steps: the differential evolution algorithm comprises the following specific steps:
u1, initializing a candidate solution population by using a real number coding mode;
u2, evaluating and analyzing the current population, including recording the optimal individuals and calculating individual fitness by using linear scale transformation;
u3, judging whether a stopping condition is met or not, including whether the maximum evolution algebra is exceeded or not, whether the population with the evolution stagnation exceeds the upper limit or not, if so, stopping the algorithm, and taking the optimal individual of the last generation as an optimal solution to be output;
u4, carrying out differential mutation on the current population, and obtaining variant individuals by using a DE/best/1 strategy, wherein a differential mutation scaling factor takes 0.5;
u5, merging the current population and the variant individuals, and generating new crossed individuals by using an exponential crossing method;
u6, adopting elite retention strategy between the current population and the iterative population to obtain a new generation population;
and U7, entering the next round of evolution, and returning to the step U2.
6. A method for correcting machining dimensional errors of a hub based on digital twinning according to claim 3, whereinThe method comprises the following steps: to determine sigma i To evaluate the importance of each dimensional feature in error correction using alpha e1, 10]The integer value of (a) represents the importance of the dimensional feature, alpha i Indicating the importance of the ith dimensional feature, when alpha i When not equal to 10, let sigma i ＝α i The method comprises the steps of carrying out a first treatment on the surface of the When alpha is i When=10, the i-th dimension feature is represented as a critical dimension, let σ i =0 and maintain y in the newly constructed planning model i ≥l i And y i ≤u i Is a constraint on (c).
7. The method for correcting the machining dimensional error of the hub based on digital twinning according to claim 1, wherein the method comprises the following steps: the specific transformation mode of the behavior model of the virtual processing equipment model defined by using the finite state machine in the step A2 is as follows, wherein S i Representing model state, e i To trigger an event, a i In response to the action:
v1, when the system starts to operate, the initial state of the processing equipment is an idle standby state S 0 ；
V2, triggering e when the hub blank reaches the designated machining position 1 The processing equipment enters a normal operation state S 1 Milling operation a is started 1 ；
V3, if the processing is successfully completed, triggering e 2 The processing equipment enters a processing completion state S 3 Send a processing completion signal a 2 To inform the robot to perform the picking operation and receive the picking completion signal e 3 After that, the processing equipment returns to the idle waiting state S 0 Transmitting a waiting blank signal a 3 ；
V4, if abnormal conditions occur in the processing process, processing equipment receives an equipment abnormal signal e 4 The processing is paused, the equipment enters into interrupt, and an exception handling signal a is sent 4 Informing the server side of processing according to the abnormal information, wherein the processing equipment enters a blocking waiting state S 2 ；
V5, if the exception is successfully processed, receiving an exception processing completion signal e 5 Continuing the processingResume pre-interrupt state a 5 The processing equipment resumes the normal operation state S 1 Otherwise, when the abnormality cannot be handled and the current processing needs to be interrupted, sending a pickup signal e 2 And receives the signal e after the completion of the picking 3 After that, the processing equipment enters an idle standby state S 0 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211385897.5A CN115576267B (en) | 2022-11-07 | 2022-11-07 | Wheel hub machining dimension error correction method based on digital twin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211385897.5A CN115576267B (en) | 2022-11-07 | 2022-11-07 | Wheel hub machining dimension error correction method based on digital twin |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115576267A CN115576267A (en) | 2023-01-06 |
| CN115576267B true CN115576267B (en) | 2023-07-07 |
Family
ID=84588151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211385897.5A Active CN115576267B (en) | 2022-11-07 | 2022-11-07 | Wheel hub machining dimension error correction method based on digital twin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115576267B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116705211B (en) * | 2023-08-04 | 2023-10-10 | 昆明理工大学 | Digital twin-based online prediction method and system for copper loss rate of oxygen-enriched copper molten pool |
| CN117047556B (en) * | 2023-10-13 | 2023-12-08 | 南通百盛精密机械有限责任公司 | Optimized machining control method and system of numerical control machine tool |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110704411B (en) * | 2019-09-27 | 2022-12-09 | 京东方科技集团股份有限公司 | Knowledge graph building method and device suitable for art field and electronic equipment |
| CN110865607A (en) * | 2019-11-07 | 2020-03-06 | 天津大学 | Five-axis numerical control machine tool control method based on digital twinning |
| CN111695734A (en) * | 2020-06-12 | 2020-09-22 | 中国科学院重庆绿色智能技术研究院 | Multi-process planning comprehensive evaluation system and method based on digital twin and deep learning |
| CN112859739B (en) * | 2021-01-15 | 2022-07-01 | 天津商业大学 | Digital twin-driven multi-axis numerical control machine tool contour error suppression method |
| CN113065276A (en) * | 2021-03-09 | 2021-07-02 | 北京工业大学 | Intelligent construction method based on digital twins |
| CN113379788B (en) * | 2021-06-29 | 2024-03-29 | 西安理工大学 | Target tracking stability method based on triplet network |
| CN113361139B (en) * | 2021-07-08 | 2023-01-31 | 广东省智能机器人研究院 | Production line simulation rolling optimization system and method based on digital twin |
| CN114091231A (en) * | 2021-10-18 | 2022-02-25 | 广西电网有限责任公司电力科学研究院 | Switch cabinet modeling method based on digital twin and error adaptive optimization |
-
2022
- 2022-11-07 CN CN202211385897.5A patent/CN115576267B/en active Active
Non-Patent Citations (1)
| Title |
|---|
| "基于多尺度残差子域适应的轴承故障诊断方法";刘晶 等;《郑州大学学报(理学版)》;第1-8页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115576267A (en) | 2023-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115576267B (en) | Wheel hub machining dimension error correction method based on digital twin | |
| CN113569903B (en) | Method, system, equipment, medium and terminal for predicting cutter abrasion of numerical control machine tool | |
| CN114237155B (en) | Error prediction and compensation method, system and medium for multi-axis numerical control machining | |
| CN115358281B (en) | Machine learning-based cold and hot all-in-one machine monitoring control method and system | |
| CN116501005B (en) | Digital twin linkage factory operation management method and system | |
| Chen | An evolutionary economic-statistical design for VSI X control charts under non-normality | |
| CN116307067A (en) | Legal holiday electric quantity comprehensive prediction method based on historical data correction | |
| CN116468536A (en) | Automatic risk control rule generation method | |
| CN116028887A (en) | Analysis method of continuous industrial production data | |
| CN117608241B (en) | Method, system, device and medium for updating digital twin model of numerical control machine tool | |
| CN110738363A (en) | photovoltaic power generation power prediction model and construction method and application thereof | |
| CN116522096B (en) | Three-dimensional digital twin content intelligent manufacturing method based on motion capture | |
| CN117592656A (en) | Carbon footprint monitoring method and system based on carbon data accounting | |
| CN117406844A (en) | Display card fan control method and related device based on neural network | |
| WO2021253689A1 (en) | Multiple regression model-based method and system for predicting price of product processing | |
| CN114192583A (en) | Scada platform-based strip steel rolling process quality monitoring method and system | |
| CN112988529A (en) | Method and system for predicting database system performance based on machine learning | |
| CN116681184B (en) | Method, device and equipment for predicting growth index of live pigs and readable storage medium | |
| CN116748352B (en) | Metal pipe bending machine processing parameter monitoring control method, system and storage medium | |
| CN117311295B (en) | Production quality improving method and system based on wireless network equipment | |
| CN115169617B (en) | Mold maintenance prediction model training method, mold maintenance prediction method and system | |
| CN117592789B (en) | Power grid environment fire risk assessment method and equipment based on time sequence analysis | |
| CN112800672B (en) | Evaluation method, system, medium and electronic equipment for boiler fouling coefficient | |
| CN117933803A (en) | Air conditioner hose product performance test data management system | |
| CN116706939A (en) | Oscillation suppression method with energy recovery function |
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 |