Method for Predicting Thermal Error of Spindle of Computer Numerical Control Machine Tool Based on Twin Feature Transferring of Virtual-Real Prototype

Abstract

A method for predicting a thermal error of a spindle of a computer numerical control (CNC) machine tool based on twin feature transferring of a virtual-real prototype, is provided, including: first, building a spindle physical prototype experiment table, and screening temperature sensitive points outside the physical prototype spindle, so as to establish an autoregressive distributed lag model; second, determining a temperature synchronization lag point on the physical prototype spindle corresponding to the temperature sensitive point outside the physical prototype spindle, so as to construct a thermal error analysis model; thereafter, establishing a virtual prototype and transferring the twin feature of the physical prototype, and by integrating a twin coupling relationship between the physical prototype spindle and the virtual prototype spindle, realizing the thermal error prediction. The present disclosure improves the accuracy of the thermal error prediction under the condition that it is difficult to arrange sensors on the spindle.

Claims

1. A method for predicting a thermal error of a spindle of a computer numerical control (CNC) machine tool based on twin feature transferring of a virtual-real prototype, comprising: S1: building a spindle physical prototype experiment table, uniformly arranging a plurality of measuring points outside a physical prototype spindle, screening and obtaining temperature sensitive points outside the physical prototype spindle from the plurality of measuring points, and establishing an autoregressive distributed lag model of the thermal error; S2: determining a temperature synchronization lag point on the physical prototype spindle corresponding to the temperature sensitive point outside the physical prototype spindle, establishing a temperature relationship between the temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle according to a grey system theory and denoting the temperature relationship as a first temperature relationship, and obtaining a thermal error analysis model of the physical prototype spindle with the temperature of the temperature synchronization lag point on the physical prototype spindle as a thermal error feature according to the first temperature relationship; S3: establishing a virtual prototype model of the machine tool, taking the temperature synchronization lag point on the physical prototype spindle as a twin feature and transferring and mapping the temperature synchronization lag point to the virtual prototype spindle of the machine tool to obtain a temperature transferring point of the virtual prototype spindle, and then determining a position of each temperature sensitive point outside the virtual prototype spindle, and then establishing a temperature relationship between the temperature transferring point on the virtual prototype spindle and the temperature sensitive point outside the virtual prototype spindle according to the grey system theory and denoting the temperature relationship as a second temperature relationship; and by integrating a twin coupling relationship between the physical prototype spindle and the virtual prototype spindle, the thermal error analysis model of the physical prototype spindle and the second temperature relationship, obtaining a thermal error prediction model of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype; S4: adjusting parameters of a physical prototype experiment table to keep running in the same working condition as the machine tool all the time, fusing the thermal error of the physical prototype spindle measured by a displacement sensor with the twin coupling relationship in S3 to obtain a real value of the thermal error of the spindle of the machine tool; thereafter, according to the temperature sensitive points outside the virtual prototype spindle, arranging a temperature sensor at the corresponding position outside the spindle of the machine tool, inputting acquired temperature data into the thermal error prediction model of the spindle of the CNC machine tool in real time to obtain a predicted value of the thermal error of the spindle of the machine tool; finally, calculating prediction accuracy according to the real value and the predicted value of the thermal error of the spindle of the machine tool, and optimizing the prediction model of the thermal error of the spindle of the CNC machine tool according to the prediction accuracy, so as to obtain a more accurate predicted value of the thermal error of the spindle of the machine tool.

2. The method for predicting the thermal error of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype according to claim 1, wherein S1 specifically comprises: S11: building the spindle physical prototype experiment table, uniformly arranging 10-20 measuring points outside the physical prototype spindle, and arranging a corresponding temperature sensor at each measuring point; S12: establishing a finite element model of the physical prototype experiment table, and preliminarily screening out leading measuring points outside the physical prototype spindle in which a temperature sudden change response is faster than a thermal error sudden change response at a sudden change of the working condition from the measuring points based on the finite element model of the physical prototype experiment table by simulating a sudden change process of the working condition; S13: using a thermal hysteresis clustering algorithm to select several leading measuring points from the leading measuring points outside the physical prototype spindle as the temperature sensitive points outside the physical prototype spindle, and by integrating a thermal hysteresis effect and a self-memory of the thermal error, establishing an autoregressive distributed lag model of the thermal error.

3. The method for predicting the thermal error of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype according to claim 2, wherein in S13, using the thermal hysteresis clustering algorithm to select several leading measuring points from the leading measuring points outside the physical prototype spindle as the temperature sensitive points outside the physical prototype spindle, specifically comprises: first, calculating average response time of each leading measuring point under each sudden change of the working condition in sequence; thereafter, according to the average response time, dividing the leading measuring points into k categories, taking an average value of average response time of all leading measuring points in each category of leading measuring points as a cluster center of this category of leading measuring points, and then calculating a sum of squares of errors under the current number of clusters according to the cluster centers of various categories of leading measuring points; changing the number of clusters k, calculating and obtaining a sum of squares of errors under different numbers of clusters, and then drawing a relationship diagram between the number of clusters k and the corresponding sum of squares of errors, and taking the number of clusters when a decline amplitude of the sum of squares of errors in the relationship diagram suddenly decreases as the number of target clusters; according to the number of target clusters, clustering the leading measuring points to obtain target clusters, and taking the most correlated leading measuring point in the target clusters as the temperature sensitive point outside the physical prototype spindle.

4. The method for predicting the thermal error of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype according to claim 3, wherein the correlation is specifically the correlation between the temperature of each leading measuring point and the real value of the thermal error of the physical prototype spindle.

5. The method for predicting the thermal error of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype according to claim 1, wherein S2 specifically comprises: S21: simulating the sudden change process of the working condition by using a finite element method, and finding the temperature synchronization lag point on the physical prototype spindle corresponding to each temperature sensitive point outside the physical prototype spindle; S22: intercepting a multi-condition limited temperature sequence of each temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle, and based on the multi-condition limited temperature sequence, determining the temperature relationship between the temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle by using the grey system theory and denoting the temperature relationship as the first temperature relationship; S23: fusing the first temperature relationship with the existing autoregressive distributed lag model of the thermal error, and obtaining the thermal error analysis model of the physical prototype spindle with the temperature of the temperature synchronization lag point on the physical prototype spindle as the thermal error feature.

6. The method for predicting the thermal error of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype according to claim 1, wherein in S3, the temperature sensitive points outside the virtual prototype spindle with the same thermal hysteresis effect as the temperature transferring points on the virtual prototype spindle are determined by a binary search method.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] FIG. 1 is an explanatory diagram of key parameters of the thermal error prediction according to the present disclosure.

[0015] FIG. 2 is a flow chart of a method according to the present disclosure.

[0016] FIG. 3 is a flow chart of selecting temperature sensitive points outside a physical prototype spindle.

[0017] FIG. 4 is a flow chart of establishing a thermal error analysis model of a physical prototype spindle.

[0018] FIG. 5 is a flow chart of performance calculation of a method for predicting a thermal error of a spindle of a CNC machine tool.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The present disclosure will be further explained with reference to the attached drawings and specific examples.

[0020] As shown in FIG. 1 and FIG. 2, the method includes the following steps S1-S4. [0021] S1: a spindle physical prototype experiment table is built, a plurality of measuring points are uniformly arranged outside a physical prototype spindle, temperature sensitive points outside the physical prototype spindle are screened from the measuring points, and an autoregressive distributed lag model of the thermal error is established. [0022] S1 specifically includes: [0023] S11: building the spindle physical prototype experiment table, uniformly arranging 10-20 measuring points outside the physical prototype spindle, and arranging a corresponding temperature sensor at each measuring point to record the temperature change of each measuring point; [0024] S12: establishing a finite element model of the physical prototype experiment table, and preliminarily screening out leading measuring points outside the physical prototype spindle in which a temperature sudden change response is faster than a thermal error sudden change response at a sudden change of the working condition from the measuring points based on the finite element model of the physical prototype experiment table by simulating a sudden change process of the working condition; and [0025] S13: using a thermal hysteresis clustering algorithm to select several leading measuring points from the leading measuring points outside the physical prototype spindle as the temperature sensitive points outside the physical prototype spindle, and by integrating a thermal hysteresis effect and a self-memory of the thermal error, establishing an autoregressive distributed lag model of the thermal error.

[0026] As shown in FIG. 3, using the thermal hysteresis clustering algorithm to select several leading measuring points from the leading measuring points outside the physical prototype spindle as the temperature sensitive points outside the physical prototype spindle, specifically includes the following steps.

[0027] First, average response time t.sub.r of each leading measuring point under each sudden change of the working condition is calculated in sequence, r=1, 2, . . . s, where s indicates the number of the leading measuring points. The calculation formula is as follows:

[00001]$t_r=\frac{1}{N}{.Math.}_{q=1}^N{t}_{\mathrm{rq}}$

[0028] where t.sub.rq indicates the response time of the r-th leading measuring point under the q-th working condition, q=1, 2, . . . N, where N indicates the number of sudden changes of the working condition.

[0029] Thereafter, according to the average response time, leading measuring points are divided into k categories, k[k.sub.min, k.sub.max], where k.sub.min indicates the minimum number of clusters, and k.sub.max indicates the maximum number of clusters, all of which are determined according to experience. It is uniformly divided between the maximum average response time t.sub.max and the minimum average response time t.sub.min, and the category of the leading measuring points is determined according to the following formula:

[00002]$l=.Math.\frac{t_r-t_{\min }}{t_{\max }-t_{\min }}.Math.$

[0030] Thereafter, an average value of the average response time of all leading measuring points in each category of leading measuring points is taken as a cluster center c.sub.l of this category of leading measuring points, l=1, 2, . . . , k, and then a sum of squares of errors under the current number of clusters is calculated according to the cluster centers of various categories of leading measuring points. The number of clusters k is changed. A sum of squares of errors under different numbers of clusters, that is, the sum of squares of errors of the response time of all measuring points under different values of k, is calculated and obtained. In the specific implementation, it is assumed that k=k.sub.min, and the sum of squares of errors of the response time of all measuring points when the measuring points are divided into k.sub.min categories is calculated. The value of k increases until k=k.sub.max, and the sums of squares of errors of the response time of all measuring points are obtained at different values of k in sequence. Thereafter, a relationship diagram between the number of clusters k and the corresponding sum of squares of errors is drawn, and the number of clusters when a decline amplitude of the sum of squares of errors in the relationship diagram suddenly decreases is taken as the number of target clusters; according to the number of target clusters, leading measuring points are clustered to obtain target clusters, and the most correlated leading measuring point in each target cluster is taken as the temperature sensitive point outside the physical prototype spindle. In the specific implementation, specifically, the correlation is the correlation between the temperature of each leading measuring point and the real value of the thermal error of the physical prototype spindle.

[0031] S2: a temperature synchronization lag point on the physical prototype spindle corresponding to the temperature sensitive point outside the physical prototype spindle is determined, a temperature relationship between the temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle is established according to a grey system theory and the temperature relationship is denoted as a first temperature relationship, and a thermal error analysis model of the physical prototype spindle, with the temperature of the temperature synchronization lag point on the physical prototype spindle as a thermal error feature, is constructed and obtained according to the first temperature relationship.

[0032] As shown in FIG. 4, S2 specifically includes the following steps S21-S23.

[0033] S21: the sudden change process of the working condition is simulated by using a finite element method, and the temperature synchronization lag point of each temperature sensitive point outside the physical prototype spindle is found on the physical prototype spindle. The temperature synchronization lag point refers to two points of which the ratio of average response time falls within the allowable range (usually, =0.05) in all sudden changes of the working condition. The temperature synchronization lag point on the physical prototype spindle corresponding to the temperature sensitive point outside the physical prototype spindle is determined according to the following formula:

[00003]$.Math.1-\frac{t_2}{t_1}.Math.$

[0034] where t.sub.1 and t.sub.2 are the average response time of the temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle, respectively.

[0035] S22: a multi-condition limited temperature sequence of each temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle is intercepted, and based on the multi-condition limited temperature sequence, the temperature relationship between the temperature sensitive point outside the physical prototype spindle and the temperature synchronization lag point on the physical prototype spindle is determined by using the grey system theory and the temperature relationship is denoted as the first temperature relationship.

[0036] S23: the first temperature relationship is fused with the existing autoregressive distributed lag model of the thermal error, and the thermal error analysis model of the physical prototype spindle, with the temperature of the temperature synchronization lag point on the physical prototype spindle as the thermal error feature, is constructed and obtained. The formula is as follows:

[00004]$Y_g=a_0+{.Math.}_{i=1}^m{a}_i{Y}_{g-i}+{.Math.}_{j=1}^p{.Math.}_{i=0}^n{}_{\mathrm{ji}}{\hat{x}}_{j1}^0\left(g-i\right)+{}_g$${\hat{x}}_{j1}^{\left(0\right)}\left(g-i\right)=\left[x_{j1}^{\left(0\right)}\left(1\right)-\frac{1}{a_j}.Math.b_{j2}{x}_{j2}^{\left(1\right)}\left(g-i\right)\right]e^{-a\left(g-i-1\right)}-\left[x_{j1}^{\left(0\right)}\left(1\right)-\frac{1}{a_j}.Math.b_{j2}{x}_{j2}^{\left(1\right)}\left(g-1-i\right)\right]e^{-a\left(g-i-2\right)}+\frac{1}{a_j}.Math.b_{j2}\left[x_{j2}^{\left(1\right)}\left(g-i\right)-x_{j2}^{\left(1\right)}\left(g-1-i\right)\right]$

[0037] where Y.sub.g indicates a predicted value of the thermal error of the spindle of the machine tool at time g, a.sub.0 indicates an adjustment coefficient, a.sub.1 indicates an i-th lagged variable weight of the thermal error, m indicates an autoregressive order, p indicates the number of exogenous variables, n indicates a distributed lag order, .sub.9 indicates white noise, .sub.ji indicates an i-th lagged variable weight of the temperature under a j-th exogenous variable, {circumflex over (x)}.sub.j1.sup.0(gi) indicates an i-th-order lagged variable of the temperature, x.sub.j1.sup.(0)(1) indicates a 1-AGO sequence of system temperature characteristic data under a j-th exogenous variable, a.sub.j indicates a development coefficient of a grey system, b.sub.j2 indicates a driving coefficient, and x.sub.j2.sup.(1) indicates a 1-AGO sequence of related factor data under a j-th exogenous variable.

[0038] S3: a virtual prototype model of the machine tool is established, the area with no obvious temperature change is excluded by using the finite element method, the arrangement area of the temperature sensor outside the spindle of the twin machine tool is delineated according to the actual machining environment on this basis, the temperature synchronization lag point on the physical prototype spindle is taken as a twin feature and is transferred and mapped to the virtual prototype spindle of the machine tool one by one to obtain a temperature transferring point of the virtual prototype spindle, and then the position of each temperature sensitive point outside the virtual prototype spindle is determined. Specifically, the virtual prototype of the finite element model of the machine tool is subjected to the same simulation process of the working condition sudden change process as S12. The temperature sensitive points outside the virtual prototype spindle with the same thermal hysteresis effect as the temperature transferring points on the virtual prototype spindle are determined by a binary search method. Thereafter, a temperature relationship between the temperature transferring point on the virtual prototype spindle and each temperature sensitive point outside the virtual prototype spindle is established according to the grey system theory and the temperature relationship is denoted as a second temperature relationship; thereafter, the twin coupling relationship between the temperatures of the corresponding point of the physical prototype spindle and the virtual prototype spindle and the twin coupling relationship between the thermal errors of the spindle are constructed. In this embodiment, the physical prototype spindle and the machine tool spindle are two spindles of the same model. In S12 and S3, the finite element model of the physical prototype and the virtual prototype of the machine tool are subjected to the exactly same working condition sudden change simulation experiment, respectively. Therefore, it is considered that the temperature changes at the corresponding positions on the two spindles are the same. From the point of view of thermo-mechanical analysis, it is known that the thermal error of the spindle is closely related to the temperature change of each position on the spindle. Therefore, the thermal error of the physical prototype spindle and the thermal error of the virtual prototype spindle are always the same.

[0039] Finally, by integrating a twin coupling relationship between the physical prototype spindle and the virtual prototype spindle, the thermal error analysis model of the physical prototype spindle and the second temperature relationship, a thermal error prediction model of the spindle of the CNC machine tool based on twin feature transferring of the virtual-real prototype is constructed and obtained.

[0040] S4: as shown in FIG. 5, parameters of the physical prototype experiment table is adjusted to keep running in the same working condition as the machine tool all the time, the thermal error of the physical prototype spindle measured by a displacement sensor is fused with the twin coupling relationship in S3 to obtain a real value of the thermal error of the spindle of the machine tool; thereafter, according to the temperature sensitive points outside the virtual prototype spindle, a temperature sensor is arranged at the corresponding position outside the spindle of the machine tool, the acquired temperature data is input into the thermal error prediction model of the spindle of the CNC machine tool in real time to obtain a predicted value of the thermal error of the spindle of the machine tool; finally, prediction accuracy is calculated according to the real value and the predicted value of the thermal error of the spindle of the machine tool, and the prediction model of the thermal error of the spindle of the CNC machine tool is optimized according to the prediction accuracy, so as to obtain a more accurate predicted value of the thermal error of the spindle of the machine tool. According to the indexes such as a root mean square error, the accuracy of the method for predicting the thermal error of the spindle of the machine tool is calculated. According to the present disclosure, the accuracy of the method for predicting the thermal error of the spindle of the machine tool is calculated in a manner of not arranging a displacement sensor on the machine tool.

[0041] Based on the twin coupling relationship established in S3, the thermal error of the physical prototype spindle measured by the displacement sensor in this embodiment is the real value of the thermal error of the spindle of the machine tool.

[0042] The above embodiments are only used to explain the present disclosure, rather than limit the present disclosure. Any modification and change made to the present disclosure within the spirit and the scope of protection of claims of the present disclosure fall within the scope of protection of the present disclosure.