Thermal error modeling method for CNC machine tools based on MLR-AHP algorithm

Through the thermal error modeling method of CNC machine tools based on MLR-AHP algorithm, the temperature sensitive points are screened using K-modulation mean clustering and hierarchical analysis method, and a multivariate linear regression model is established, which solves the insufficient thermal error prediction accuracy and robustness in the existing technology. The problem of this method achieves higher thermal error prediction accuracy and robustness.

CN115729170BInactive Publication Date: 2024-11-05NORTHEAST DIANLI UNIVERSITY
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202211440640.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-11-17
Publication Date
2024-11-05
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing thermal error prediction model of CNC machine tools has shortcomings in the selection and modeling methods of temperature measurement point, resulting in insufficient thermal error prediction accuracy and robustness, affecting the accuracy of the machine tool.

Method used

Using the thermal error modeling method of CNC machine tools based on MLR-AHP algorithm, the linear axis positioning errors under different temperature states of the machine tool are measured, and the temperature sensitive points are screened using the K-modulation mean clustering algorithm and hierarchical analysis method to establish multivariate linear regression thermal errors. Model to improve prediction accuracy and robustness.

Benefits of technology

It significantly improves the accuracy and robustness of the thermal error prediction model, meets the error compensation requirements, and has the advantages of scientific and reasonableness, strong applicability and good robustness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115729170B_ABST
    Figure CN115729170B_ABST
Patent Text Reader

Abstract

The present invention is a method for modeling thermal errors of numerically controlled machine tools based on the MLR-AHP algorithm, which is characterized in that the method includes obtaining the temperature variables and positioning errors of the linear axis of the machine tool, extracting the temperature sensitive points required for thermal error modeling by using the K-harmonic mean clustering algorithm, further optimizing the temperature sensitive points by using the hierarchical analysis method, establishing the MLR-AHP model of thermal errors of the machine tool, and measuring the temperature at the positions of the temperature sensitive points of the machine tool in real time, bringing the temperature values ​​into the MLR-AHP model, and obtaining the prediction performance of the MLR-AHP model. The method can meet the requirements of error compensation and improve the prediction accuracy and robustness of the thermal error model. The method has the advantages of being scientific and reasonable, having strong applicability and good robustness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of precision machining machine tools and relates to a method for predicting thermal errors of numerically controlled machine tools based on an MLR-AHP model, which is suitable for predicting thermal errors of numerically controlled machine tools. Background Art

[0002] As the "working mother machine" of equipment manufacturing, machine tools are a symbol of the level of a country's manufacturing industry. In the actual processing process, changes in internal and external heat sources such as motor heating, friction heat, cutting heat and ambient temperature will cause thermal deformation of machine tool parts. As a result, the relative position between the tool and the workpiece deviates from the predetermined value, namely thermal error, which causes a decrease in machine tool accuracy. As the performance of CNC machine tools gradually improves, thermal error has gradually become the main factor affecting machine tool accuracy.

[0003] There are two problems in the research of machine tool thermal error compensation. One is the selection of temperature measurement points, and the other is the modeling method. At present, the most commonly used algorithms for selecting temperature measurement points include thermal modal analysis, fuzzy clustering and kernel principal component method to screen temperature sensitive points. These studies have achieved the optimization of temperature measurement points to a certain extent. On the other hand, for the construction of thermal error prediction models, after years of research by experts and scholars, a variety of machine tool thermal error models have been established, such as support vector machines, neural networks, finite element models, least squares support vector machines, etc. Among them, the least squares support vector machine model has a higher accuracy in predicting the regression problem of nonlinear small samples. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a method for establishing a thermal error model of CNC machine tools based on MLR-AHP, which is scientific, reasonable, highly applicable, and robust and can significantly improve the accuracy and robustness of the thermal error prediction model.

[0005] The technical solution adopted to achieve the purpose of the present invention is: a CNC machine tool thermal error modeling method based on MLR-AHP algorithm, characterized in that the method comprises the following steps:

[0006] 1) Measure the positioning error of the linear axis of the machine tool under different temperature conditions;

[0007] 2) Use K-harmonic means clustering algorithm to extract the temperature sensitive points required for thermal error modeling;

[0008] 3) Use the analytic hierarchy process to further screen temperature sensitive points to simplify the model;

[0009] 4) Establish the MLR-AHP model of machine tool thermal error;

[0010] 5) Obtaining a thermal error prediction value of the MLR-AHP model, and obtaining the prediction performance of the MLR-AHP model according to the difference between the thermal error prediction value and the thermal error measurement value.

[0011] Further, the step 1) measuring the positioning error of the linear axis of the machine tool under different temperature states includes:

[0012] a. According to the temperature sensor layout strategy, 11 PT100 platinum resistance temperature sensors are initially installed on the machine tool, and the temperature distribution of the machine tool is collected in real time using the XSR90 recorder;

[0013] b. Continuously sample the temperature of the machine tool heat source and the surrounding environment;

[0014] c. Use Renishaw XL-80 laser interferometer to measure the positioning error of the linear axis Y of the machine tool under different temperature conditions;

[0015] d. Use the error separation method to separate the positioning error and obtain the initial geometric positioning error and the thermal positioning error under different temperature conditions.

[0016] Furthermore, the step 2) uses the K-harmonic mean clustering algorithm to extract the temperature sensitive points required for thermal error modeling, including:

[0017] i. Select the initial cluster center c j :

[0018] Randomly select j initial cluster centers c1, c2, ..., c j , where c j represents the cluster center of the jth class, and t is set as the number of iterations;

[0019] ii. Calculate the fitness function:

[0020]

[0021] Where: X={x1,x2,…,x n}, n is the number of objects to be clustered; p is an input parameter, usually p ≥ 2;

[0022] iii. For each temperature sensitive point x i , calculate its membership in each cluster center;

[0023]

[0024] iv. For each cluster center c j , according to its membership, weight and all temperature sensitive points x i , recalculate the cluster centers;

[0025]

[0026] v. Repeat steps ii to iv until the predetermined number of iterations is reached, and the optimal temperature sensitive points T1, T5 and T6 are selected;

[0027] Furthermore, the step 3) further screens the temperature sensitive points using the analytic hierarchy process, including:

[0028] i. Establish a hierarchical model:

[0029] Screen the temperature sensitive points T1, T5 and T6, take the thermal error as the target, take T1, T5 and T6 as the decision criteria, and build a hierarchical relationship based on the mutual relationship;

[0030] ii. Construct judgment matrix:

[0031]

[0032] Among them, a ij Indicates the temperature sensitive point T i and T j The relative importance of

[0033] iii. Hierarchical single sorting

[0034] Calculate the maximum eigenvalue λ of the judgment matrix max The eigenvector of is normalized, λ max =3.0649, the corresponding eigenvector is [0.2483, 0.9628, 0.1067], and its normalized w = [0.1884, 0.7306, 0.0810];

[0035] iv.Consistency check:

[0036] Calculate the consistency index CI

[0037]

[0038] Where n is the order of the judgment matrix;

[0039] Calculate the consistency ratio CR

[0040]

[0041] Among them, RI is the average random consistency index, when n=3, the value is 0.52,

[0042]

[0043] CR is less than 0.1, which means the consistency of the judgment matrix is ​​acceptable;

[0044] v. Reassign temperature sensitive points:

[0045] The redistribution weight coefficients of T1, T5 and T6 are 0.1184, 0.7306 and 0.0810, respectively. Therefore, according to the AHP analysis results, in order to further reduce the number of temperature sensitive points, T6 with the least influence was eliminated.

[0046] Further, the step 4) of establishing the MLR-AHP model of machine tool thermal error includes:

[0047] ① Geometric positioning error fitting: select the fifth-order polynomial for fitting

[0048] δ yy (Y) = -0.48119y - 0.01246y 2 +1.36575×10 -4 y 3 -4.99795×10 -7 y 4 +6.19125×10 -10 y 5

[0049] where δ yy (Y) represents the Y-axis positioning error, and y represents the Y-axis position;

[0050] ② Thermal positioning error modeling: With the thermal slope as output, AHP selected two temperature sensitive points as input and established a multivariate linear regression thermal error model:

[0051] K i =0.00429T1+0.04071T5-1.19986

[0052] Where K i is the thermal slope;

[0053] ③Construct a comprehensive positioning error model: Combine the geometric error term and the thermal error term, the formula is as follows:

[0054] δ yy (Y,T)=δ yy (Yy)+(K i -K0)y

[0055] =(0.00429T1+0.04071T5-1.28584)y-0.01246y 2 +1.36575×10 -4 y 3 -4.99795×10 -7 y 4 +6.19125×10 -10 y 5

[0056] Among them, δ yy (Y,T) represents the comprehensive positioning error of the Y axis.

[0057] Furthermore, the step 5) obtains the thermal error prediction value of the MLR-AHP model, and obtains the prediction performance of the MLR-AHP model according to the difference between the thermal error prediction value and the thermal error measurement value, including:

[0058] (1) Real-time measurement of the positioning error of the machine tool linear axis and its corresponding temperature value in the moving state;

[0059] (2) The real-time temperature value is brought into the trained MLR-AHP model to obtain the real-time thermal error prediction value and compare the predicted value with the measured value;

[0060] (3) Through RMSE, R 2 The three indicators of and η are used to evaluate the prediction performance of the model overall;

[0061] The RMSE formula is:

[0062]

[0063] R 2 The formula is:

[0064]

[0065] The formula for η is:

[0066]

[0067] Among them, RMSE is the root mean square error, R 2 is the coefficient of determination, η precision, x i is the measured value; is the output value of the error model; is the mean of the measurements.

[0068] The invention discloses a method for modeling thermal errors of numerically controlled machine tools based on an MLR-AHP algorithm. The method includes the steps of obtaining the temperature variables and positioning errors of the linear axis of the machine tool, extracting the temperature sensitive points required for thermal error modeling by using a K-harmonic mean clustering algorithm, further optimizing the temperature sensitive points by using a hierarchical analysis method, establishing an MLR-AHP model of thermal errors of the machine tool, and measuring the temperature at the positions of the temperature sensitive points of the machine tool in real time, bringing the temperature values ​​into the MLR-AHP model, and obtaining the prediction performance of the MLR-AHP model. The method can meet the requirements of error compensation and improve the prediction accuracy and robustness of the thermal error model. The method has the advantages of being scientific and reasonable, having strong applicability, and having good robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flowchart of the method involved in the present invention;

[0070] Figure 2 is the sensor installation position of the embodiment;

[0071] Figure 3 is a positioning error curve graph measured in the embodiment;

[0072] Figure 4 is a temperature distribution curve diagram obtained in the embodiment;

[0073] Figure 5 is a schematic diagram of the MLR-AHP model fitting results of the embodiment;

[0074] Figure 6 is a positioning error curve diagram of the prediction experiment measurement in the embodiment;

[0075] Figure 7 is a temperature distribution curve diagram of the prediction experiment of the embodiment;

[0076] Figure 8 Schematic diagram of the prediction results of the MLR-AHP model of the embodiment. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The specific embodiments described here are only used to explain the present invention and are not used to limit the present invention.

[0078] This embodiment proposes a method for establishing a thermal error model of a CNC machine tool based on MLR-AHP, such as Figure 1 FIG. 1 is a flow chart of a method for establishing a thermal error model of a CNC machine tool based on MLR-AHP according to the present embodiment.

[0079] In order to verify the effectiveness of the present invention, a thermal error experiment of a CNC machine tool was conducted. On a three-axis experimental platform, a PT100 platinum resistance temperature sensor and an XSR90 recorder were used to collect temperature data in real time, the temperature values ​​of the machine tool heat source and the environment were sampled at intervals, and the positioning error of the linear axis Y of the machine tool under different temperature states was measured using a Renishaw XL-80 laser interferometer. The three-axis experimental platform is a prior art device.

[0080] The arrangement position of the sensor in this implementation is as follows: Figure 2 As shown, the specific distribution is T1 measuring the Y-axis motor, T2 measuring the column, T3 measuring the front end of the Y-axis guide, T4 measuring the bed, T5 measuring the rear end of the Y-axis guide, T6 measuring the surrounding environment of the machine tool, T7 measuring the front bearing of the Y-axis, T8 measuring the workbench, T9 measuring the rear bearing of the Y-axis, T 10Measuring ball screw nut, T 11 Measure the on-site environment.

[0081] The positioning error of the Y axis is located in the 280mm stroke, and bidirectional measurement is performed at intervals of 10mm. When the machine tool is in the state from shutdown to just starting up, the positioning error measured is the Y axis geometric positioning error. Then, at a speed of 20mm / s and 10mm / s 2 The Y axis is moved within the entire stroke at an acceleration of 1000 rpm to heat the machine. The positioning error of the Y axis is measured after 20, 50, 110, 190, 240, 340 and 390 minutes of operation until the machine reaches thermal equilibrium. The measured positioning error curve is shown in Figure 1. Figure 3 As shown, the eight numbers 0-7 represent the eight positioning error curves in chronological order. At the same time, when the machine tool reaches thermal equilibrium, the temperature distribution is measured every 10 minutes. Figure 4 The obtained temperature distribution curve is shown in Figure 2.

[0082] The present embodiment proposes a method for establishing a thermal error model of a CNC machine tool based on MLR-AHP, comprising the following steps:

[0083] S1: Measure the positioning error of the linear axis of the machine tool under different temperature conditions;

[0084] S101: According to the temperature sensor layout strategy, 11 PT100 platinum resistance temperature sensors are initially installed on the machine tool, and the temperature distribution of the machine tool is collected in real time using the XSR90 recorder;

[0085] S102: performing interval sampling of the temperature values ​​of the heat source of the machine tool and the surrounding environment;

[0086] S103: Using Renishaw XL-80 sensor to measure the positioning error of the linear axis Y of the machine tool under different temperature conditions;

[0087] S104: Separating the positioning error by using an error separation method to obtain an original geometric positioning error and a thermal positioning error;

[0088] S2: Use K-harmonic mean clustering algorithm to extract temperature sensitive points required for thermal error modeling

[0089] S201: Select the initial cluster center c j :

[0090] Specifically, randomly select j initial cluster centers c1, c2, ..., c j , where c j represents the cluster center of the jth class, and t is set as the number of iterations;

[0091] S202: Calculate fitness function;

[0092]

[0093] Where: X={x1,x2,…,x n}, n is the number of objects to be clustered; p is an input parameter, usually p ≥ 2;

[0094] S203: For each temperature sensitive point x i , calculate its membership in each cluster center;

[0095]

[0096] S204: For each cluster center c j , according to its membership, weight and all temperature sensitive points x i , recalculate the cluster centers;

[0097]

[0098] S205: repeating steps 2-4 until a predetermined number of iterations is reached, and the best temperature sensitive points T1, T5 and T6 are screened out;

[0099] S3: Use the analytic hierarchy process to further screen temperature sensitive points:

[0100] S301: Establishing a hierarchical model:

[0101] Screen the temperature sensitive points T1, T5 and T6, take the thermal error as the target, take T1, T5 and T6 as the decision criteria, and build a hierarchical relationship based on the mutual relationship;

[0102] S302: Construct a judgment matrix:

[0103]

[0104] Among them, a ij Indicates the temperature sensitive point T i and T j The relative importance of

[0105] S303: Hierarchical Single Sorting

[0106] Calculate the maximum eigenvalue λ of the judgment matrix max The eigenvector of is normalized, λ max =3.0649, the corresponding eigenvector is [0.2483, 0.9628, 0.1067], and its normalized w = [0.1884, 0.7306, 0.0810];

[0107] S304: Consistency check:

[0108] Calculate the consistency index CI

[0109]

[0110] Where n is the order of the judgment matrix.

[0111] Calculate the consistency ratio CR

[0112]

[0113] Among them, RI is the average random consistency index, and when n=3, the value is 0.52.

[0114]

[0115] CR is less than 0.1, and the consistency of the judgment matrix is ​​acceptable;

[0116] S305: Reassign temperature sensitive points:

[0117] The redistribution weight coefficients of T1, T5 and T6 are 0.1184, 0.7306 and 0.0810 respectively. Therefore, according to the AHP analysis results, in order to further reduce the number of temperature sensitive points, T6 with the least influence was eliminated.

[0118] S4: Establishing the MLR-AHP model of machine tool thermal error:

[0119] S401: Geometric positioning error fitting: Select a fifth-order polynomial for fitting

[0120] δ yy (Y) = -0.48119y - 0.01246y 2 +1.36575×10 -4 y 3 -4.99795×10 -7 y 4 +6.19125×10 -10 y 5

[0121] where δ yy (Y) represents the Y-axis positioning error, and y represents the Y-axis position;

[0122] S402: Thermal positioning error modeling: With the thermal slope as output, AHP selected two temperature sensitive points as input and established a multivariate linear regression thermal error model:

[0123] K i =0.00429T1+0.04071T5-1.19986

[0124] Where K iis the thermal slope;

[0125] S403: Construct a comprehensive positioning error model: Combine the geometric error term and the thermal error term, and the formula is as follows:

[0126] δ yy (Y,T)=δ yy (Yy)+(K i -K0)y

[0127] =(0.00429T1+0.04071T5-1.28584)y-0.01246y 2 +1.36575×10 -4 y 3 -4.99795×10 -7 y 4 +6.19125×10 -10 y 5

[0128] Among them, δ yy (Y,T) represents the comprehensive positioning error of the Y axis. The fitting results are as follows: Figure 5 As shown (dots represent measured values).

[0129] S5: Obtaining the thermal error prediction value of the MLR-AHP model, and obtaining the prediction performance of the MLR-AHP model according to the difference between the thermal error prediction value and the thermal error measurement value, including:

[0130] S501: The positioning error of the linear axis of the machine tool in the moving state and its corresponding temperature value are measured in real time for prediction. The measured positioning error curve is as follows: Figure 6 As shown in the figure, there are four positioning error curves under different temperature conditions, and the temperature distribution collected synchronously is as follows Figure 7 As shown;

[0131] S502: Bring the real-time temperature value into the trained MLR-AHP model to obtain the real-time thermal error prediction value, and compare the predicted value with the measured value. The result is as follows: Figure 8 As shown (dots represent measured values);

[0132] S503: Root mean square error RMSE, determination coefficient R 2 The prediction performance of the model is evaluated overall using three indicators:

[0133] The RMSE formula is:

[0134]

[0135] R 2 The formula is:

[0136]

[0137] The formula for η is:

[0138]

[0139] Among them, x i is the measured value; is the output value of the error model; is the mean of the measurements.

[0140] The performance evaluation of the model is shown in Table 1.

[0141] Table 1 Performance evaluation of MLR-AHP model

[0142]

[0143] Figure 5 As shown in Table 1, the model has very good fitting performance. Figure 8 It can also be seen from Table 1 that the difference between the predicted value and the measured value of the model is also very close. Therefore, the MLR-AHP model has a high prediction accuracy and can meet the error compensation requirements, which proves the feasibility of the MLR-AHP model as an error modeling method.

[0144] 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 it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some or all of the technical features therein can be replaced by equivalents. Therefore, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. A CNC machine tool thermal error modeling method based on MLR-AHP algorithm, characterized in that: The method comprises the following steps: 1) Measure the positioning error of the linear axis of the machine tool under different temperature conditions; 2) Use K-harmonic mean clustering (KHM) algorithm to extract the temperature sensitive points required for thermal error modeling; 3) Use the analytic hierarchy process to further screen temperature sensitive points to simplify the model; 4) Establish the MLR-AHP model of machine tool thermal error; 5) obtaining a thermal error prediction value of the MLR-AHP model, and obtaining the prediction performance of the MLR-AHP model according to the difference between the thermal error prediction value and the thermal error measurement value; The step 2) uses the K-harmonic mean clustering algorithm to extract the temperature sensitive points required for thermal error modeling, including: i. Select the initial cluster center c j : Randomly select j initial cluster centers c1, c2, ..., c j , where c j represents the cluster center of the jth class, and t is set as the number of iterations; ii. Calculate the fitness function: Where: X={x1,x2,…,x n }, n is the number of objects to be clustered; p is an input parameter, p ≥ 2; iii. For each temperature sensitive point x i , calculate its membership in each cluster center; iv. For each cluster center c j , according to its membership, weight and all temperature sensitive points x i , recalculate the cluster centers; v. Repeat steps ii to iv until the predetermined number of iterations is reached, and the temperature sensitive points T1, T5 and T6 are screened out; The step 3) further screens the temperature sensitive points using the analytic hierarchy process, including: i. Establish a hierarchical model: Screen the temperature sensitive points T1, T5 and T6, take the thermal error as the target, take T1, T5 and T6 as the decision criteria, and build a hierarchical relationship based on the mutual relationship; ii. Construct judgment matrix: Among them, a ij Indicates the temperature sensitive point T i and T j The relative importance of iii. Hierarchical single sorting Calculate the maximum eigenvalue λ of the judgment matrix max The eigenvector of is normalized, λ max =3.0649, the corresponding eigenvector is [0.2483, 0.9628, 0.1067], and its normalized w = [0.1884, 0.7306, 0.0810]; iv.Consistency check: Calculate the consistency index CI Where n is the order of the judgment matrix; Calculate the consistency ratio CR Among them, RI is the average random consistency index, and when n=3, the value is 0.52; CR is less than 0.1, which means the consistency of the judgment matrix is ​​acceptable; v. Reassign temperature sensitive points: The redistribution weight coefficients of T1, T5 and T6 are 0.1184, 0.7306 and 0.0810, respectively. Therefore, according to the AHP analysis results, in order to further reduce the number of temperature sensitive points, T6 with the least influence was eliminated; The step 4) of establishing the MLR-AHP model of machine tool thermal error includes: ① Geometric positioning error fitting: select the fifth-order polynomial for fitting δ yy (Y)=-0.48119y-0.01246y 2 +1.36575×10 -4 and 3 -4.99795×10 -7 and 4 +6.19125×10 -10 and 5 where δ yy (Y) represents the Y-axis positioning error, and y represents the Y-axis position; ② Thermal positioning error modeling: With the thermal slope as output, AHP selected two temperature sensitive points as input and established a multivariate linear regression thermal error model: K i =0.00429T1+0.04071T5-1.19986 Where K i is the thermal slope; ③Construct a comprehensive positioning error model: Combine the geometric error term and the thermal error term, the formula is: δ yy (Y,T)=δ yy (Yy)+(K i -K0)y =(0.00429T1+0.04071T5-1.28584)y-0.01246y 2 +1.36575×10 -4 y 3 -4.99795×10 -7 and 4 +6.19125×10 -10 and 5 Among them, δ yy (Y,T) represents the comprehensive positioning error of the Y axis.

2. The method for modeling thermal errors of CNC machine tools based on the MLR-AHP algorithm according to claim 1, characterized in that: The step 1) measuring the positioning error of the linear axis of the machine tool under different temperature states includes: a. According to the temperature sensor layout strategy, 11 PT100 platinum resistance temperature sensors are initially installed on the machine tool, and the temperature distribution of the machine tool is collected in real time using the XSR90 recorder; b. Continuously sample the temperature of the machine tool heat source and the surrounding environment; c. Use Renishaw XL-80 laser interferometer to measure the positioning error of the linear axis Y of the machine tool under different temperature conditions; d. Use the error separation method to separate the positioning error and obtain the original geometric positioning error and thermal positioning error.

3. The CNC machine tool thermal error modeling method based on the MLR-AHP algorithm according to claim 1 is characterized in that: The step 5) obtains the thermal error prediction value of the MLR-AHP model, and obtains the prediction performance of the MLR-AHP model according to the difference between the thermal error prediction value and the thermal error measurement value, including: (1) Real-time measurement of the positioning error of the machine tool linear axis and its corresponding temperature value in the moving state; (2) The real-time temperature value is brought into the trained MLR-AHP model to obtain the real-time thermal error prediction value and compare the predicted value with the measured value; (3) Through RMSE, R 2 The three indicators of and η are used to evaluate the prediction performance of the model overall; The RMSE formula is: R 2 The formula is: The formula for η is: Where RMSE is the root mean square error, R 2 is the coefficient of determination, η precision, x i is the measured value; is the output value of the error model; is the mean of the measurements.

Citation Information

Cited By

  • Machining workpiece surface shape prediction model establishment method considering machine tool motion shafting rigidity deformation

    CN116822246A