Machine tool structure thermal sensitive point identification and thermal deformation prediction method oriented to guideway accuracy

Abstract

The application discloses a machine tool structure thermal sensitive point identification and thermal deformation prediction method for guideway precision, belongs to the machining technical field of numerical control machine tools, and comprises the following steps: S1, thermal characteristic analysis of a machine tool is carried out by using thermal characteristic simulation analysis software, and the establishment and analysis of a steady temperature field and a thermal deformation field of structural components of the numerical control machine tool when the structural components are affected by a heat source; S2, thermal sensitive points are selected, and the number and positions of temperature monitoring points are optimized, so that the thermal deformation of the structural components can be accurately predicted by using fewer thermal sensitive points; and S3, a thermal error prediction model is established, and the relationship between the thermal sensitive points and key thermal errors of the structural components is specifically established, and the accuracy and effectiveness of the selected thermal sensitive points in predicting the deformation of the structural components are verified by computer simulation; and the machine tool structure thermal sensitive point identification and thermal deformation prediction method for guideway precision can accurately predict the thermal deformation of the structural components.

Description

Technical Field

[0001] The present invention relates to the technical field of numerical control machine tool processing, and in particular to a method for identifying thermally sensitive points of a machine tool structure and predicting thermal deformation in view of guide rail accuracy. Background Art

[0002] In the field of precision machining, machine tool accuracy, a key indicator of equipment manufacturing performance, directly impacts machining quality in industries such as aerospace. Current approaches to improving machine tool accuracy primarily focus on active temperature control and thermal error compensation. Both require real-time monitoring of the machine tool's thermal state through a network of temperature sensors to predict overall thermal deformation. However, existing research has yet to establish a system for identifying thermally sensitive points based on optimal guideway accuracy, resulting in systematic deviations in overall machine accuracy prediction.

[0003] The complexity of machine tool thermal deformation errors stems from the coupling of multiple physical fields. Guideways, as the reference element of machine tool motion, experience thermally induced deformations that are amplified step by step through the kinematic chain, ultimately leading to an exponential degradation of machining accuracy. Although scholars at home and abroad have constructed various thermal-structural coupling models using finite element analysis and conducted steady-state and transient thermal analyses with the help of simulation software, these studies generally suffer from limitations. Existing thermal deformation prediction models often employ global optimization objectives, and temperature sensor placement strategies lack guidance for maintaining guideway accuracy. This makes it difficult for existing thermal imaging-based temperature field reconstruction methods to accurately identify thermally sensitive points in guideway systems. Consequently, a dedicated evaluation function for maintaining guideway accuracy cannot be constructed, making it difficult to optimize the temperature measurement point layout to meet the requirements for optimal guideway accuracy. Therefore, it is urgent to develop a method for identifying thermally sensitive points guided by optimal guideway accuracy. By modeling the thermal-mechanical coupling transfer mechanism and innovating measurement point optimization strategies, accurate prediction of guideway thermal deformation using a minimum set of temperature measurement points can be achieved, providing a new theoretical paradigm for thermal error compensation in high-precision machine tools. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for identifying thermally sensitive points and predicting thermal deformation of a machine tool structure with respect to guide rail accuracy, so as to solve the problems existing in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention provides a method for identifying thermally sensitive points and predicting thermal deformation of a machine tool structure with respect to guide rail accuracy, comprising the following steps:

[0006] S1. Use thermal characteristics simulation analysis software to analyze the thermal characteristics of the machine tool. When the structural components of the CNC machine tool are affected by the heat source, the steady-state temperature field and thermal deformation field are established and analyzed, specifically including:

[0007] S11. Simplify the machine tool model, analyze the main heat source distribution, and calculate the boundary conditions required for thermal characteristics simulation analysis;

[0008] S12, obtain the temperature field and thermal deformation field results of the whole machine reaching steady state under a given working condition, and extract the temperature values of the preliminary arranged temperature measuring points and the thermal deformation simulation results of the path;

[0009] S2, select the heat-sensitive points, optimize the number and position of the temperature monitoring points, and use fewer heat-sensitive points to accurately predict the thermal deformation of the structural components, specifically including:

[0010] S21, list the original temperature data as a temperature matrix according to different working conditions, and perform data standardization;

[0011] S22, process the standardized data to establish a fuzzy homomorphism matrix;

[0012] S23, convert the fuzzy homomorphism matrix into a fuzzy equivalence matrix;

[0013] S24, perform clustering analysis based on the fuzzy equivalence matrix to determine the best classification result and obtain the heat-sensitive points;

[0014] S3, establish a thermal error prediction model, i.e., the relationship between the heat-sensitive point temperature and the key thermal error of the structural component, and verify the accuracy and effectiveness of predicting the deformation of the structural component using the selected heat-sensitive points through computer simulation, specifically including:

[0015] S31, based on the experimental data, construct the relationship between the heat-sensitive point temperature and the key thermal deformation of the structural component;

[0016] S32, compare and analyze the deviation between the thermal deformation predicted by the model and the actual simulation to verify the accuracy of the model.

[0017] Preferably, in step S11, some small features and small parts that have little effect on the simulation results are deleted, the main heat source distribution of the machine tool is analyzed, and the main heat sources are concentrated near the motor, motor seat, guide rail and bearing seat. Then calculate the boundary conditions, the calculation formula of the boundary conditions of different heat sources is as follows:

[0018] Motor boundary condition: Where, M T is the output torque of the servo motor; n m is the servo motor speed; η is the mechanical efficiency of the servo motor; Q m is the heat generated by the motor; V m is the volume of the motor;

[0019] Guide rail slider boundary condition: Q n = (μF + f)v, Where: μ is the dynamic friction coefficient; F is the load applied to the friction surface; f is the friction resistance of the slider sealing device; v is the relative motion speed of the guide rail slider; Qn is friction heat; A n is the contact area between the slider and the guide rail;

[0020] Bearing pair boundary conditions: Q j =0.1047n j M j , Among them, n j is the bearing speed; M j is the friction torque; Q j Generates heat for the bearing; V j is the bearing volume.

[0021] Preferably, step S12 is specifically as follows: the boundary condition calculation results of the given working condition are respectively substituted into the thermal simulation analysis software, wherein the heat generated by the motor, the heat generated by the motor seat and the heat generated by the bearing seat are loaded in the form of heat generation rate, the heat generated by the guide rail slider is loaded in the form of heat flux density, and the convection coefficient of the motor seat is 30W / (m 2 ·℃), the guide rail-slider convection coefficient is 20W / (m 2 ·℃), the convection coefficient of the spindle box is 22W / (m 2 ·℃), the convection coefficient at other locations is 30W / (m 2 ·℃), conduct whole machine simulation analysis, and extract temperature measurement point values ​​and structural component deformation data.

[0022] Preferably, step S21 is specifically as follows:

[0023] Assume T i (i=1,2,...,n) are n temperature measurement points to be classified, each of which has m measurement values. The original data matrix is ​​U, where

[0024] To perform standardization: The standardized data is Among them: Standardized data elements The data size interval is [0,1], max{T ij} is the largest element in the original data.

[0025] Preferably, in step S22, a data coupling analytical method is used to establish a fuzzy homomorphic matrix R. Among them, r ij =R(T i ,T j ) indicates T i With T j degree of similarity; k is the index variable; T ik is the kth measurement value of the i-th temperature measurement point in the original data matrix; T jkis the kth measurement value of the jth temperature measurement point in the original data matrix; is the average value of the temperature measurement point i; is the average value of the temperature at the jth measuring point.

[0026] Preferably, in step S23, the fuzzy homomorphic matrix is ​​converted into a fuzzy equivalent matrix, that is, the transitive closure t(R) of R is obtained by multiplying R by itself. 2 , and then repeat the multiplication to get R 4 , until a natural number k is found that satisfies the specific condition: R k =R 2k , let t(R)=R 2k , thus obtaining the fuzzy equivalence matrix t(R).

[0027] Preferably, step S24 is specifically as follows:

[0028] Based on the fuzzy equivalence matrix t(R), multiple different λ values ​​are selected from the closed interval [0,1]. These λ values ​​determine different classification results, so as to achieve the goal of classifying the temperature variable; if r is satisfied ij ≥λ condition, let T ij =1, otherwise let T ij = 0, simplifying the fuzzy equivalence matrix into a Boolean matrix, thereby achieving a faster and more intuitive dynamic classification process; selecting the threshold λ∈[0,1], calculating the λ-truncation matrix R of the fuzzy equivalence matrix t(R) λ , remember R λ =(T ij ) n×n , when T ij =1, then T i and T j Classify them into one category and repeat this step until all temperature measurement points are grouped. Then evaluate each category using the compact separation quotient index and determine the optimal λ value and number of categories accordingly. The compact separation quotient index is defined as: Among them, V is the compact separation quotient index; c is the number of categories; p is the number of data; u ij is an element in the membership matrix, indicating the membership of the jth sample to the ith class; q is a weighted index, generally 2; Represents the sum of squares of distances between each point in the cluster and the cluster center, used to evaluate the compactness of the cluster within the cluster; min i≠j ||v i -v j || 2 Represents the minimum sum of squares between the class and the class center, used to evaluate the separation between classes; v i represents the center of the i-th class; x j represents the jth point in each class, v jV represents the center of the jth class, because good clustering division should make the intra-class compactness smaller and the inter-class separation larger, so the smaller V is, the better the classification is, and the better the classification effect is, and the group with the smallest V is selected as the final result of fuzzy classification; after fuzzy clustering and grouping of the temperature measuring points, a measuring point with strong correlation with thermal deformation is selected from each class as a thermal sensitive point, and the temperature measuring points of the same class are further screened and calculated by using the Pearson coefficient: Wherein, N is the number of temperature data and thermal deformation data; x i represents the temperature value of the ith point, represents the average value of all temperature data, y i represents the thermal deformation of the ith point, represents the average value of all thermal deformation data, and the temperature measuring point with a large Pearson coefficient calculation result in each class is selected as a thermal sensitive point.

[0029] Preferably, step S31 specifically comprises:

[0030] Lasso regression method is adopted to construct an independent mathematical model of deformation of machine tool structure parts, i.e. a mathematical model with multiple input and multiple output characteristics:

[0031]

[0032] Wherein, y is the dependent variable; β0 is the constant term; β1~β N is the regression coefficient; x1~x N is the independent variable; ε is the error term; δ 2 is the variance;

[0033] By introducing the L1 regularization term to constrain the parameter β, the objective function is defined as:

[0034]

[0035] Wherein, y i is the actual observation value; is the model prediction value; λ is the regularization coefficient; β j is the regression coefficient;

[0036] The parameter estimation needs to minimize the residual sum of squares with the regularization term, i.e.:

[0037]

[0038] Wherein, b0 is the constant term; b1~b N is the regression coefficient after Lasso regression; x i1 ~x iN is the independent variable, which is the temperature value of the thermal sensitive point;

[0039] The coordinate descent optimization algorithm is used to solve the parameters and obtain the sparse regression equation, which is the thermal deformation prediction model:

[0040]

[0041] Among them, x1~x N is the independent variable.

[0042] Preferably, step S32 is specifically as follows: using the Lasso regression method, the temperature data of the selected thermal sensitive point and the deformation δ of the machine tool structure are correlated. i Perform linear combination to obtain the corresponding expression: in, is the model prediction value of the deformation of the machine tool structure; b0, b i is the fitting coefficient of the error of each measuring point; T i is the temperature value of the thermally sensitive point; the temperature data of the thermally sensitive point is substituted into the above formula to obtain the predicted value of the deformation of the machine tool structure. The predicted value is compared with the actual value using drawing software, the deviation between the two is analyzed, and the accuracy of the model is verified.

[0043] Therefore, the present invention adopts the above-mentioned method for identifying thermal-sensitive points and predicting thermal deformation of machine tool structures for guide rail accuracy, imports the model into simulation software for thermal characteristic analysis, calculates and analyzes the required thermal loads and boundary conditions, obtains the temperature field and thermal deformation field of the machine tool, and extracts the thermal deformation simulation results of the critical path. With the goal of optimizing the prediction effect of the axis guide rail straightness error, the fuzzy clustering analysis method is used in combination with the compact separation quotient index and the data coupling analysis method to optimize the thermal-sensitive points. By installing fewer temperature sensors, most of the key information related to the deformation of the structural parts can be obtained. Then, the Lasso regression method is used to establish a thermal error prediction model, and then the accurate prediction of the thermal deformation of the structural parts is achieved based on the real-time temperature monitoring information of the thermal-sensitive points.

[0044] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of the method for identifying thermal sensitive points and predicting thermal deformation of a machine tool structure with respect to guide rail accuracy according to the present invention;

[0046] Figure 2 Flowchart of step S1 in an embodiment of the present invention;

[0047] Figure 3 This is a flowchart of step S2 in an embodiment of the present invention;

[0048] Figure 4 This is a flow chart of step S3 in an embodiment of the present invention;

[0049] Figure 5Schematic diagram of thermal simulation of a machine tool in an embodiment of the present invention;

[0050] Figure 6 Schematic diagram of the distribution of temperature measurement points for measuring the temperature field of a machine tool in an embodiment of the present invention;

[0051] Figure 7 Schematic diagram comparing the predicted value and actual value of the machine tool guide rail deformation in an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort shall fall within the scope of protection of the present invention.

[0053] See also Figure 1-Figure 7 , a method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures for guide rail accuracy includes the following steps:

[0054] S1. Use thermal characteristics simulation analysis software to analyze the thermal characteristics of the machine tool. When the structural components of the CNC machine tool are affected by the heat source, the steady-state temperature field and thermal deformation field are established and analyzed, specifically including:

[0055] S11. Delete and simplify some small features and components that have little impact on the simulation results. Analyze the main heat source distribution of the machine tool, which is mainly concentrated near the motor, motor seat, guide rail, and bearing seat. Then calculate the boundary conditions. The calculation formulas for different heat source boundary conditions are as follows:

[0056] Motor boundary conditions: Among them, M T is the output torque of the servo motor; n m is the servo motor speed; η is the mechanical efficiency of the servo motor; Q m Generates heat for the motor; V m is the motor volume;

[0057] Guide rail slider boundary conditions: Q n =(μF+f)v, Where: μ is the dynamic friction coefficient; F is the load applied to the friction surface; f is the friction resistance of the slider seal; v is the relative motion speed of the guide rail slider; Q n is friction heat; A n is the contact area between the slider and the guide rail;

[0058] Boundary conditions of bearing pair: Q j =0.1047n j M j , wherein n j is the bearing speed; M j is the friction torque; Q j is the bearing heat generation; V j is the bearing volume.

[0059] S12, the boundary condition calculation results of the given working condition are respectively substituted into the thermal simulation analysis software, wherein the motor heat generation, the motor base heat generation and the bearing seat heat generation are loaded in the form of heat generation rate, the guide rail and the slider heat generation is loaded in the form of heat flow density, the motor base convection coefficient is 30 W / (m 2 ·℃), the guide rail and the slider convection coefficient is 20 W / (m 2 ·℃), the spindle box convection coefficient is 22 W / (m 2 ·℃), and the convection coefficient of the remaining positions is 30 W / (m 2 ·℃), the whole machine simulation analysis is carried out, and the temperature measuring point value and the structural component deformation data are extracted;

[0060] S2, the thermal sensitive points are selected, and the number and position of the temperature monitoring points are optimized, so as to realize accurate prediction of the thermal deformation of the structural component using fewer thermal sensitive points, and specifically including:

[0061] S21, the original temperature data is listed as a temperature matrix according to different working conditions, and data standardization is carried out; let T i (i=1, 2,..., n) be n temperature measuring points to be classified, wherein each temperature measuring point has m measurement values, the original data matrix is U, and wherein

[0062] standardization processing is carried out: the standardized data is obtained wherein the data size interval of the standardized data element is [0, 1], and max{T ij} is the maximum element in the original data.

[0063] S22, the standardized data is processed, a fuzzy homomorphism matrix R is established by using a data coupling analysis method, wherein r ij =R(T i ,T j ) represents the similarity degree of T i and T j ; k is an index variable; T ik is the kth measurement value of the ith temperature measuring point in the original data matrix; T jk is the kth measurement value of the jth temperature measuring point in the original data matrix; is the measurement average value of the ith temperature measuring point; is the average value of the temperature at the jth measuring point.

[0064] S23. Convert the fuzzy homomorphic matrix into a fuzzy equivalent matrix, that is, find the transitive closure t(R) of R. Specifically, first obtain R by multiplying R by itself. 2 , and then repeat the multiplication to get R 4 , until a natural number k is found that satisfies the specific condition: R k =R 2k , let t(R)=R 2k , thus obtaining the fuzzy equivalent matrix t(R);

[0065] S24. Based on the fuzzy equivalence matrix t(R), multiple different λ values ​​are selected from the closed interval [0,1]. These λ values ​​determine different classification results, so as to achieve the goal of classifying the temperature variable; if r is satisfied ij ≥λ condition, let T ij =1, otherwise let T ij = 0, simplifying the fuzzy equivalence matrix into a Boolean matrix, thereby achieving a faster and more intuitive dynamic classification process; selecting the threshold λ∈[0,1], calculating the λ-truncation matrix R of the fuzzy equivalence matrix t(R) λ , remember R λ =(T ij ) n×n , when T ij =1, then T i and T j Classify them into one category and repeat this step until all temperature measurement points are grouped. Then evaluate each category using the compact separation quotient index and determine the optimal λ value and number of categories accordingly. The compact separation quotient index is defined as: Among them, V is the compact separation quotient index; c is the number of categories; p is the number of data; u ij is an element in the membership matrix, indicating the membership of the jth sample to the ith class; q is a weighted index, generally 2; Represents the sum of squares of distances between each point in the cluster and the cluster center, used to evaluate the compactness of the cluster within the cluster; min i≠j ||v i -v j || 2 Represents the minimum sum of squares between the class and the class center, used to evaluate the separation between classes; v i represents the center of the i-th class; x j represents the jth point in each class, v jV represents the center of the jth class, because good clustering division should make the intra-class compactness smaller and the inter-class separation larger, so the smaller V is, the better the classification is, and the better the classification effect is, and the group with the smallest V is selected as the final result of fuzzy classification; after fuzzy clustering and grouping of the temperature measuring points, a temperature measuring point with strong correlation with thermal deformation is selected from each class as a thermal sensitive point, and the temperature measuring points of the same class are further screened and calculated by using the Pearson coefficient: Wherein, N is the number of temperature data and thermal deformation data; x i represents the temperature value of the ith point, represents the average value of all temperature data, y i represents the thermal deformation of the ith point, y all represents the average value of all thermal deformation data, and the temperature measuring point with a large Pearson coefficient calculation result in each class is selected as a thermal sensitive point;

[0066] S3, a thermal error prediction model, i.e. the relationship between the temperature of the thermal sensitive point and the key thermal error of the structural part, is established, and the accuracy and effectiveness of the selected thermal sensitive point in predicting the deformation of the structural part are verified by computer simulation, specifically including:

[0067] S31, a Lasso regression method is used to construct an independent mathematical model of the deformation of the machine tool structural part, i.e. a mathematical model with multiple input and output characteristics:

[0068]

[0069] Wherein, y is the dependent variable; β0 is the constant term; β1~β N is the regression coefficient; x1~x N is the independent variable; ε is the error term; δ 2 is the variance;

[0070] By introducing the L1 regularization term to constrain the parameter β, the objective function is defined as:

[0071]

[0072] Wherein, y i is the actual observation value; is the model prediction value; λ is the regularization coefficient; β j is the regression coefficient;

[0073] The parameter estimation needs to minimize the residual sum of squares with the regularization term, i.e.:

[0074]

[0075] Wherein, b0 is the constant term; b1~b N is the regression coefficient after Lasso regression; xi1 ~x iN is the independent variable, which is the temperature value of the thermal sensitive point;

[0076] The coordinate descent optimization algorithm is used to solve the parameters and obtain the sparse regression equation, which is the thermal deformation prediction model:

[0077]

[0078] Among them, x1~x N is the independent variable;

[0079] S32, using Lasso regression method, the temperature data of the selected thermal sensitive points and the deformation of the machine tool structure δ i Perform linear combination to obtain the corresponding expression: in, is the model prediction value of the deformation of the machine tool structure; b0, b i is the fitting coefficient of the error of each measuring point; T i is the temperature value of the thermally sensitive point; the temperature data of the thermally sensitive point is substituted into the above formula to obtain the predicted value of the deformation of the machine tool structure. The predicted value is compared with the actual value using drawing software, the deviation between the two is analyzed, and the accuracy of the model is verified.

[0080] Therefore, the present invention adopts the above-mentioned method for identifying thermal-sensitive points and predicting thermal deformation of machine tool structures for guide rail accuracy. First, the model is imported into the simulation software for thermal characteristic analysis, the thermal load and boundary conditions required for analysis are calculated, the temperature field and thermal deformation field of the machine tool are obtained, and the thermal deformation simulation results of the critical path are extracted. With the goal of optimizing the prediction effect of the axis guide rail straightness error, the fuzzy clustering analysis method is used in combination with the compact separation quotient index and the data coupling analysis method to optimize the thermal sensitive points. Finally, the Lasso regression method is used to establish a thermal error prediction model, which can realize the accurate prediction of the thermal deformation of structural components.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy, characterized in that: The following steps are involved: S1. Use thermal characteristics simulation analysis software to analyze the thermal characteristics of the machine tool. When the structural components of the CNC machine tool are affected by the heat source, the steady-state temperature field and thermal deformation field are established and analyzed, specifically including: S11. Simplify the machine tool model, analyze the heat source distribution, and calculate the boundary conditions required for thermal characteristics simulation analysis; S12. Obtain the temperature field and thermal deformation field results of the entire machine when it reaches a steady state under given working conditions, and extract the temperature values ​​of several initially arranged temperature measurement points and the thermal deformation simulation results of the paths; S2. Select thermally sensitive points and optimize the number and location of temperature monitoring points to accurately predict the thermal deformation of structural components using a minimum number of thermally sensitive points. Specifically, the following are included: S21. List the original temperature data into a temperature matrix according to different working conditions and perform data standardization; S22, processing the standardized data to establish a fuzzy homomorphic matrix; S23, converting the fuzzy homomorphic matrix into a fuzzy equivalent matrix; S24. Perform cluster analysis based on the fuzzy equivalence matrix to determine the classification results and obtain thermal sensitive points; S3. Establish a thermal error prediction model, specifically the relationship between the temperature of thermally sensitive points and the key thermal errors of structural components, and verify the accuracy and effectiveness of predicting the deformation of structural components using selected thermally sensitive points through computer simulation, including: S31. Based on the experimental data, the relationship between the temperature of the thermally sensitive point and the key thermal deformation of the structural parts is established; S32, comparing the thermal deformation predicted by the model with the thermal deformation obtained by actual simulation to analyze the deviation between the two and verify the accuracy of the model; In step S22, a data coupling analytical method is used to establish a fuzzy homomorphic matrix , ,in, express and degree of similarity; ; , is the index variable, is the first The first temperature measurement point measurements, is the first The first temperature measurement point measurements, For the The average value of the temperature measurement points, For the The average value of the temperature measurement points; In step S23, the fuzzy homomorphic matrix is ​​converted into a fuzzy equivalent matrix to obtain The transitive closure of , specifically: first pass The self-multiplication operation gives , and then repeat the multiplication to get , until a natural number k is found that satisfies the specific conditions: ,make , thus obtaining the fuzzy equivalence matrix ; Step S24 is specifically as follows: Based on fuzzy equivalence matrix , select multiple different Values, these The value determines different classification results, so as to achieve the goal of classifying the temperature variable; if Condition, order =1, otherwise let =0, simplify the fuzzy equivalence matrix into a Boolean matrix to realize the dynamic classification process; select the threshold [0,1], calculate the fuzzy equivalent matrix of Truncation Matrix ,remember ,when =1, then and The temperature measurement points are grouped into one category and this step is repeated until all the temperature measurement points are grouped. Then the categories are evaluated by the compact separation quotient index and the best one is determined accordingly. values ​​and the number of categories, the compact separation quotient index is defined as: ,in, is the compact separation quotient indicator; is the classification number; is the number of data; is an element in the membership matrix, indicating the The samples for Class membership; is a weighted index; It represents the sum of squares of the distances between each point in the cluster and the cluster center, and is used to evaluate the intra-cluster compactness of the cluster; Represents the minimum sum of squares between the class and the class center, used to evaluate the separation between classes; Indicates the the center of the class; Indicates the first Points, Indicates the The center of the class, select The smallest group is taken as the final result of fuzzy classification; after fuzzy clustering the temperature measurement points, a measurement point with a strong correlation with thermal deformation is selected from each category and defined as a thermal sensitive point. The temperature measurement points of the same type are further screened and the Pearson coefficient is used for calculation: ,in, N is the number of temperature data and thermal deformation data; Indicates the The temperature value of each point, represents the average value of all temperature data, Indicates the Thermal deformation of each point, It represents the average value of all thermal deformation data. The temperature measurement point with the largest Pearson coefficient calculation result in each category is selected as the thermal sensitive point.

2. The method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy according to claim 1, characterized in that: The calculation formulas for the boundary conditions of different heat sources in step S11 are: Motor boundary conditions: , ,in, is the output torque of the servo motor; is the servo motor speed; is the mechanical efficiency of the servo motor; Generates heat for the motor; is the motor volume; Guide rail slider boundary conditions: , ,in: is the kinetic friction factor; is the load applied to the friction surface; The friction resistance of the slider seal; is the relative motion speed of the guide rail slider; is friction heat; is the contact area between the slider and the guide rail; Boundary conditions of bearing pair: , ,in, is the bearing speed; is the friction torque; Generates heat for bearings; is the bearing volume.

3. The method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy according to claim 2, characterized in that: Step S12 is specifically as follows: the boundary condition calculation results of the given working condition are respectively substituted into the thermal simulation analysis software, wherein the heat generated by the motor, the heat generated by the motor seat and the heat generated by the bearing seat are loaded in the form of heat generation rate, the heat generated by the guide rail slider is loaded in the form of heat flux density, and the convection coefficient of the motor seat is 30W / m 2 ℃, the guide rail-slider convection coefficient is 20W / m 2 ℃, the spindle box convection coefficient is 22W / m 2 ℃, the convection coefficient at other locations is 30W / m 2 ℃, conduct whole machine simulation analysis, extract temperature measurement point values ​​and structural component deformation data.

4. The method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy according to claim 3, characterized in that: Step S21 is specifically as follows: set up To be classified temperature measuring points, each of which has The original data matrix is ,in ; To perform standardization: The standardized data is , where: Standardized data elements The data size interval is [0,1], is the largest element in the original data.

5. The method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy according to claim 1, characterized in that: Step S31 is specifically as follows: The Lasso regression method is used to construct an independent mathematical model of the deformation of machine tool structural components: ； in, is the dependent variable; is a constant term; is the regression coefficient; is the independent variable; is the error term; is the variance; By introducing the L1 regularization term to the parameters The objective function is defined as: ； in, is the actual observed value; is the model prediction value; is the regularization coefficient; is the regression coefficient; Parameter estimation needs to minimize the residual sum of squares with regularization terms, which is: ； in, is a constant term; is the regression coefficient after Lasso regression; is the independent variable, which is the temperature value of the thermal sensitive point; The coordinate descent optimization algorithm is used to solve the parameters and obtain the sparse regression equation, which is the thermal deformation prediction model: ； in, is the independent variable.

6. The method for identifying thermal sensitive points and predicting thermal deformation of machine tool structures based on guide rail accuracy according to claim 5, characterized in that: Step S32 is specifically: using the Lasso regression method, the temperature data of the selected thermal sensitive points and the deformation of the machine tool structure are compared. Perform linear combination to obtain the corresponding expression: ,in, is the model prediction value of the machine tool structural component deformation; 、 is the fitting coefficient of the error at each measuring point; is the temperature value of the thermally sensitive point; the temperature data of the thermally sensitive point is substituted into the above formula to obtain the predicted value of the deformation of the machine tool structure. The predicted value is compared with the actual value using drawing software, the deviation between the two is analyzed, and the accuracy of the model is verified.

Images