CN115167276B - Cutter life prediction method based on cutter path running track - Google Patents
Cutter life prediction method based on cutter path running track Download PDFInfo
- Publication number
- CN115167276B CN115167276B CN202210713491.9A CN202210713491A CN115167276B CN 115167276 B CN115167276 B CN 115167276B CN 202210713491 A CN202210713491 A CN 202210713491A CN 115167276 B CN115167276 B CN 115167276B
- Authority
- CN
- China
- Prior art keywords
- cutter
- processing
- tool
- cutting
- parameters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
- G05B19/19—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by positioning or contouring control systems, e.g. to control position from one programmed point to another or to control movement along a programmed continuous path
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/36—Nc in input of data, input key till input tape
- G05B2219/36291—Cutting, machining conditions by empirical equation, like tool life
Abstract
The invention discloses a tool life prediction method based on tool path running tracks, which extracts all tool processing paths and tool processing parameters according to NC programs, calculates a process model of workpiece geometry after each processing procedure based on a digital model before each processing procedure of a tool according to the tool processing paths of each processing procedure; dividing the cutter processing path according to unit time according to a process model of workpiece geometry and the cutter processing path, and calculating the cutting quantity of each time point of the cutter; obtaining historical processing parameter information of the scrapped tool, and establishing a tool life prediction model of the relation between the stress condition and the life of the tool based on deep learning; predicting the processed simulation parameters of the current cutter to obtain the residual service life expectation of the current cutter. The method and the device can improve the prediction precision of the service life of the cutter.
Description
Technical Field
The invention relates to the field of numerical control machining tool life management, in particular to a tool life prediction method based on a tool path running track.
Background
Most current tool life prediction methods are directly related to the type and process of a processed workpiece: setting an initial life according to the cutter performance parameters and the machining experience, dividing the machined workpieces into a plurality of large types, and setting and distributing different life loss values according to the rough machining type, the medium machining type and the finish machining type of each type of workpieces. The life of the tool is subtracted from the life of the current machining type after each use of the tool.
Under the condition of more types of processed workpieces, the processing loss values of the tools for different types of workpieces are difficult to be counted effectively. The workpieces are difficult to sort, and some workpieces are the same in type, but the actual sizes may have magnitude differences, so that the set loss parameters are unreasonable. In the course of processing, details are not considered enough, for example: both single cutting and back and forth cutting can process planes, but the processing track of the two can be obviously different from the resistance of the cutter in the cutting process. The accumulated factors often generate great deviation for predicting the service life of the cutter.
Disclosure of Invention
The invention mainly aims to provide a cutter life prediction method based on a cutter path running track, which improves prediction accuracy.
The technical scheme adopted by the invention is as follows: a cutter life prediction method based on a cutter path running track comprises the following steps:
s1, extracting processing information:
extracting all cutter processing paths and cutter processing parameters according to an NC program, wherein the cutter processing parameters comprise cutter type, cutting speed and feeding speed, and obtaining cutter numbers by butting with an MES system;
s2, reconstructing a workpiece machining process model:
calculating a process model of the geometry of the workpiece after each machining process according to the machining path of the cutter before each machining process based on the digital model of the cutter before each machining process;
s3, calculating cutting quantity:
dividing the cutter processing path according to unit time according to a process model of workpiece geometry and the cutter processing path, and calculating the cutting quantity of each time point of the cutter;
s4, storing the serialization parameters:
storing the cutting speed, the feeding speed and the cutting quantity of all the cutters in the machining process in an array mode;
s5, establishing a cutter life prediction model:
counting numbers of all scrapped tools in an MES system, combining S1-S4 to obtain historical processing parameter information of the scrapped tools, and establishing a tool life prediction model of the relation between the stress condition and the life of the tools based on deep learning according to the principle that the continuous work of the tools leads to deduction of the life;
s6, predicting the service life of the current cutter:
and predicting the processed simulation parameters of the current cutter by using a cutter life prediction model to obtain the residual service life of the current cutter.
According to the scheme, the S1 analyzes text data in the NC file by reading the NC file processed by the data to obtain a cutter processing path and cutter processing parameters, and the keyword rule is extracted based on the ISO standard.
According to the scheme, the cutter processing path comprises an arc section and a straight line section, and is extracted according to the following steps:
inquiring the NC program using the cutter in an MES system according to the cutter number;
obtaining a linear interpolation motion starting point and a linear interpolation motion ending point by identifying a G01 instruction in an NC program;
obtaining a starting point, an ending point and a circular arc radius of circular arc interpolation motion by identifying a G02/G03 instruction in an NC program;
converting the circular arc and straight line parameters into three-dimensional curves and recording corresponding line numbers of each curve in an NC program;
obtaining tool movement radius correction parameters by identifying G41/G42 in an NC program, recording line numbers, and converting all correction parameters into a serial data format of [ { "line number": 12 }, { "line number": 42 }, "radius correction amount": -0.02} ];
obtaining cutter length correction parameters by identifying G43/G44 in an NC program, recording line numbers, and converting all the length correction parameters into a serial data format of [ { "line numbers": 20 ": length correction amount": 0.1}, { "line numbers": 50 ": radius correction amount": 0.0} ];
each row of correction parameters will influence the subsequent straight line interpolation and arc interpolation instructions until new correction parameters are replaced; and (3) offsetting the circular arc and the linear motion track according to the correction amount and the correction direction.
According to the scheme, the cutting speed is obtained through the following steps:
inquiring the diameter D of the cutter from a cutter database according to the cutter model;
obtaining a spindle rotation parameter by identifying an M03/M04 instruction in an NC program, and recording an M03/M04 instruction line number so as to obtain a spindle rotation number N;
according to the formula: cutting speed vc=pi×dn/1000, and converting the cutting speed parameter into a serial data format of [ { "line number": 10, "cutting speed": 0.05}, { "line number": 80, "cutting speed": 0.025} ].
According to the scheme, the feeding speed is obtained through the following steps:
the spindle rotation parameters are obtained by identifying M03/M04 instructions in an NC program, and the line numbers of the M03/M04 instructions are recorded, so that the spindle rotation number N is obtained;
obtaining a cutter feeding speed F by identifying an F instruction in an NC program, and recording an F instruction line number;
according to the formula: feed speed v=n×f, calculate feed speed V, insert one feed speed each time there is a code line modified by N and F values: the feed rate conversion format is [ { "line number": 10, "feed rate": 0.2}, { "line number": 100, "feed rate": 0.5} ].
According to the above scheme, the S2 specifically includes:
inquiring a digital model of the cutter from a cutter database according to the cutter number;
creating a section line of the movement direction of the cutter based on the digital model of the cutter;
using the offset cutter machining path as a guide line to sweep a section line in the cutter movement direction, and creating a sweep model;
and performing difference solving by using the digitized model and the sweep model before workpiece processing to obtain a process model of the workpiece geometry after the current processing is completed.
According to the above scheme, the step S3 specifically comprises:
3.1, performing triangle gridding on a workpiece and a digital model of a cutter before the current processing;
3.2, converting the workpiece triangular mesh and the cutter triangular mesh into octree meshes with the precision of S, wherein the conversion process is as follows:
3.2.1, taking the center of the cutter as a reference, shifting X to the motion direction, wherein X is the simulation precision, so that the cutter is completely enveloped by the octree initial square;
3.2.2 if the octree mesh comprises any triangle in the triangle mesh and the current octree mesh size is greater than the conversion accuracy S, continuously subdividing the octree;
3.2.3 if the octree grid does not contain any triangle in the triangle grid or the octree grid size is smaller than the conversion precision S, ending the current grid subdivision task;
3.3, calculating the fitting condition of each cell in the workpiece octree grid and the cutter octree grid, wherein the number of the cells in the workpiece octree grid, which are fitted with the cutter octree grid, is Tg;
3.4, the pressed area of the cutter is positively correlated with Tg, the cutting quantity of the cutter is Tg x V, the cutting force Ty=k of the cutter is Tg x V, k is the hardness coefficient of the workpiece material, and V is the feeding speed;
3.5, setting an initial cutting point as A0, repeating the steps of 3.1-3.4 at intervals of L cutting distances, and performing simulation calculation on the stress condition once; each simulation point on the whole cutter processing path is marked with A0, A1 and A2 … … An;
3.6 comparing the stress difference ratio of A1 to A0: if the difference ratio is smaller than the standard value, the stress conditions of the two points are considered to be close to each other, and the two points are combined; processing time parameters of each stress section are obtained by combining the feeding speed of the cutter, and the combined serialization parameters are expressed as follows: { "processing time": 0,30 }, "cutting force": 200}; the cutting force is the average value of the cutting forces at the points A0 and A1;
3.7 continuously comparing the stress difference ratio of the point A2 and the point A0/A1:
and if the difference ratio is smaller than the standard value, combining the difference ratio with A0 and A1, and expressing the combined serialization parameters as follows: { "processing time": 0,45 }, "cutting force": 210}, continuing to compare the ratio of the difference between the stress at point A3 and the combined stress at point A0/A1/A2 according to 3.7, and thus, until An; if the difference is larger than the standard value, taking A2 as a new initial cutting point A0, and taking each simulation point A3 … … An after the A2 as a new A1 … … An-2, and repeating for 3.6;
and 3.8, finally obtaining a processing procedure stress condition parameter array and submitting the array to an MES system database, wherein the serialization format is as follows: [ { "processing time": 0,45], "cutting force": 200}, { "processing time": 45,60], "cutting force": 210} ].
According to the above scheme, the step S5 specifically comprises:
5.1, counting numbers of all scrapped cutters in the MES system;
5.2, obtaining historical processing parameter information of the scrapped cutter according to the calculation steps in S1, S2 and S3;
5.3, continuously working the cutter to lead the deduction service life to be converted into a polynomial of degree 2 for fitting, and obtaining the formula: Σ (a1+a2+ty+a0) (b1+tn+b2+tn+b0); tn identifies the continuous cutting time; a1, a2, ty, a0, b1, b2 and b0 are respectively polynomial coefficients to be solved and are obtained by training a deep learning network;
5.4 deep learning network adopts a combination of a multi-layer full-connection layer network and an LSTM network, wherein the full-connection layer network is used for carrying out formula fitting, and the LSTM is used for recursively influencing the service life of the current cutter by the service condition of the cutter; the input of the deep learning network is the historical processing duration time and the processing process cutting force of a single cutter, and the output of the full-connection layer is single processing life loss; the input of the LSTM network is the accumulation of the life loss of the output of the full-connection layer, and the output of the LSTM network is the total life loss of the cutter after the current machining is executed;
and 5.5, setting a deep learning model training iteration termination condition, and directly outputting a cutter life prediction model by a full-connection layer in the training model when the iteration termination condition is met.
According to the above scheme, the step S6 specifically comprises:
inquiring a cutter number in a cutter library according to the model of a cutter used for a workpiece to be processed, and inquiring the accumulated loss service life U loss of the cutter according to the cutter number;
and (3) obtaining simulation parameters of the tool in the processing process of the workpiece through the S1, the S2 and the S3, substituting the simulation parameters into the tool life prediction model constructed in the S5 for calculation, and obtaining the residual service life expected Uin the current tool.
According to the above scheme, the step S6 further includes: recording the theoretical life of the cutter as U, and if (U loss+U pre)/U is smaller than a first preset proportion, considering that the secondary cutter can be safely used in the processing; if (U loss+u pre)/U theory > = first preset proportion and (U loss+u pre)/U theory < = second preset proportion, giving an early warning in the MES system; if the (U loss+Upre)/U principle is greater than a second preset proportion, notifying the site to perform tool changing treatment;
the first preset proportion is less than the second preset proportion < =100%.
The invention has the beneficial effects that: the method is based on NC programs, extracts accurate cutter processing paths and cutter processing parameters, trains a cutter life prediction model by analyzing historical cutter processing parameters and service lives, predicts the residual life of a cutter, and improves prediction accuracy.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method according to an embodiment of the present invention.
FIG. 2 is an ISO standard code table for NC programs.
Fig. 3 is a schematic view of the division of the stress state area of the cutter.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
As shown in fig. 1, the present invention provides a tool life prediction method based on a tool path movement track, which includes the following steps:
s1, extracting processing information: all tool processing paths and tool processing parameters including tool type, cutting speed and feeding speed are extracted according to the NC program, and tool numbers are obtained by interfacing with an MES system (workshop production management system). The step S1 is to read the NC file (digital control file Numerical Control) of the data processing, parse the text data to obtain the tool processing path and the tool processing parameters, and extract the keyword rule based on the ISO standard, as shown in fig. 2.
The NC program text style is as follows:
T09 M06
G00 G90 X-20.Y-10.S2300 M03
G43 Z30.H00
Z1.
G41 G01 X-20.462Y-10.192F500.M08
G17 G03 X-20.Y-12.5I.462J-1.108
#101＝-0.500
#102＝-21.500
WHILE[#101GE#102]DO1
X-20.Y-12.5Z#101I0.0 J2.5
#101＝-#101+[-1.500]
END1
I0.0 J2.5
X-19.538Y-10.192I0.0 J1.2
G40 G01 X-20.Y-10.Z-21.5
G00 Z30.
M02
％
the cutter processing path comprises an arc section and a straight line section, and is extracted according to the following steps:
inquiring the NC program using the cutter in an MES system according to the cutter number;
obtaining a linear interpolation motion starting point and a linear interpolation motion ending point by identifying a G01 instruction in an NC program;
obtaining a starting point, an ending point and a circular arc radius of circular arc interpolation motion by identifying a G02/G03 instruction in an NC program;
converting the circular arc and straight line parameters into three-dimensional curves and recording corresponding line numbers of each curve in an NC program;
obtaining tool movement radius correction parameters by identifying G41/G42 in an NC program, recording line numbers, and converting all correction parameters into a serial data format of [ { "line number": 12 }, { "line number": 42 }, "radius correction amount": -0.02} ];
obtaining cutter length correction parameters by identifying G43/G44 in an NC program, recording line numbers, and converting all the length correction parameters into a serial data format of [ { "line numbers": 20 ": length correction amount": 0.1}, { "line numbers": 50 ": radius correction amount": 0.0} ];
each row of correction parameters will influence the subsequent straight line interpolation and arc interpolation instructions until new correction parameters are replaced; and (3) offsetting the circular arc and the linear motion track according to the correction amount and the correction direction.
The cutting speed is obtained by the following steps:
inquiring the diameter D of the cutter from a cutter database according to the cutter model;
obtaining a spindle rotation parameter by identifying an M03/M04 instruction in an NC program, and recording an M03/M04 instruction line number so as to obtain a spindle rotation speed N (rpm);
according to the formula: cutting speed vc=pi×dn/1000, and converting the cutting speed parameter into a serial data format of [ { "line number": 10, "cutting speed": 0.05}, { "line number": 80, "cutting speed": 0.025} ].
The feeding speed is obtained through the following steps:
obtaining a spindle rotation parameter by identifying an M03/M04 instruction in an NC program and recording an M03/M04 instruction line number so as to obtain a spindle rotation number N (rpm);
obtaining a cutter feeding speed F (i.e. the cutter feeding amount per revolution, mm/rev) by identifying an F instruction in an NC program, and recording an F instruction line number;
according to the formula: feed speed v=n×f, calculate feed speed V, insert one feed speed each time there is a code line modified by N and F values: the feed rate conversion format is [ { "line number": 10, "feed rate": 0.2}, { "line number": 100, "feed rate": 0.5} ].
S2, reconstructing a workpiece machining process model: based on the digital model before each processing procedure of the cutter, calculating a process model of the workpiece geometry after each processing procedure according to the cutter processing path of each processing procedure. The method comprises the following steps: inquiring a digital model of the cutter from a cutter database according to the cutter number; creating a section line of the movement direction of the cutter based on the digital model of the cutter; using the offset cutter machining path as a guide line to sweep a section line in the cutter movement direction, and creating a sweep model; and performing difference solving by using the digitized model and the sweep model before workpiece processing to obtain a process model of the workpiece geometry after the current processing is completed.
S3, calculating cutting quantity:
dividing the cutter processing path according to unit time according to the process model of the workpiece geometry and the cutter processing path, and calculating the cutting quantity of each time point of the cutter.
The S3 specifically comprises the following steps:
3.1, performing triangle gridding on a workpiece and a digital model of a cutter before the current processing;
3.2 converting the workpiece triangular mesh and the cutter triangular mesh into octree mesh with the accuracy of S (for example, 0.1 mm), wherein the conversion process is as follows:
3.2.1 the center of the octree initial block takes the center of the cutter as a reference, and offsets X to the moving direction, wherein X is the simulation precision (for example, 2 mm), so that the octree initial block just completely envelops the cutter;
3.2.2 if the octree mesh comprises any triangle in the triangle mesh and the current octree mesh size is greater than the conversion accuracy S, continuously subdividing the octree;
3.2.3 if the octree grid does not contain any triangle in the triangle grid or the octree grid size is smaller than the conversion precision S, ending the current grid subdivision task;
3.3, calculating the fitting condition of each cell in the workpiece octree grid and the cutter octree grid, wherein the number of the cells in the workpiece octree grid, which are fitted with the cutter octree grid, is Tg;
3.4, the pressure area of the cutter is positively correlated with Tg, the cutting amount of the cutter is Tg x V, the cutting force of the cutter, namely the pressure Ty=k x Tg x V, k is the hardness coefficient of the workpiece material, and V is the feeding speed;
3.5 as shown in FIG. 3, setting the initial cutting point as A0, repeating the stress condition simulation calculation once for 3.1-3.4 at each interval L of cutting distance; each simulation point on the whole cutter processing path is marked with A0, A1 and A2 … … An;
3.6 comparing the stress difference ratio of A1 to A0: if the difference ratio is smaller than the standard value, the stress conditions of the two points are considered to be close to each other, and the two points are combined; processing time parameters of each stress section are obtained by combining the feeding speed of the cutter, and the combined serialization parameters are expressed as follows: { "processing time": 0,30 }, "cutting force": 200}; the cutting force is the average value of the cutting forces at the points A0 and A1;
3.7 continuously comparing the stress difference ratio of the point A2 and the point A0/A1:
and if the difference ratio is smaller than the standard value, combining the difference ratio with A0 and A1, and expressing the combined serialization parameters as follows: { "processing time": 0,45 }, "cutting force": 210}, continuing to compare the ratio of the difference between the stress at point A3 and the combined stress at point A0/A1/A2 according to 3.7, and thus, until An; if the difference is larger than the standard value, taking A2 as a new initial cutting point A0, and taking each simulation point A3 … … An after the A2 as a new A1 … … An-2, and repeating for 3.6;
and 3.8, finally obtaining a processing procedure stress condition parameter array and submitting the array to an MES system database, wherein the serialization format is as follows: [ { "processing time": 0,45], "cutting force": 200}, { "processing time": 45,60], "cutting force": 210} ].
Octree (Octree) is a tree data structure for describing a three-dimensional space. Each node of the octree represents a volumetric element of a cube, each node having eight child nodes, the volumetric elements represented by the eight child nodes together being equal to the volume of the parent node. The general center point serves as the bifurcation center of the nodes. If the octree is not a null tree, only eight or zero child nodes of any node in the tree exist, that is, the child nodes do not have numbers other than 0 and 8.
Octree leaf child nodes represent the highest resolution case. For example, if the resolution is set to 0.1cm, then each leaf is a 1mm square.
S4, storing the serialization parameters: the cutting speed, the feeding speed and the cutting quantity of all the cutters in the machining process are stored in an array mode.
S5, establishing a cutter life prediction model: and counting numbers of all scrapped tools in the MES system, combining S1-S4 to obtain historical processing parameter information of the scrapped tools, and establishing a tool life prediction model of the relation between the stress condition and the life of the tools based on deep learning according to the principle that the continuous work of the tools leads to deduction of the life.
The S5 specifically comprises the following steps:
5.1, counting numbers of all scrapped cutters in the MES system;
5.2, obtaining historical processing parameter information of the scrapped cutter according to the calculation steps in S1, S2 and S3;
5.3, continuously working the cutter to lead the deduction service life to be converted into a polynomial of degree 2 for fitting, and obtaining the formula: Σ (a1+a2+ty+a0) (b1+tn+b2+tn+b0); tn identifies the continuous cutting time; a1, a2, ty, a0, b1, b2 and b0 are respectively polynomial coefficients to be solved and are obtained by training a deep learning network.
5.4 deep learning network adopts a combination of a multi-layer full-connection layer network and an LSTM network, wherein the full-connection layer network is used for carrying out formula fitting, and the LSTM is used for recursively influencing the service life of the current cutter by the service condition of the cutter; the input of the deep learning network is the historical processing duration time and the processing process cutting force of a single cutter, and the output of the full-connection layer is single processing life loss; the input of the LSTM network is the accumulation of the life loss of the output of the full-connection layer, and the output of the LSTM network is the total life loss of the cutter after the current machining is executed;
and 5.5, setting a deep learning model training iteration termination condition, and directly outputting a cutter life prediction model by a full-connection layer in the training model when the iteration termination condition is met. In this embodiment, the iteration termination condition is: the accuracy of 1000 iterations or models on the test set reached 90%.
S6, predicting the service life of the current cutter: and predicting the processed simulation parameters of the current cutter by using a cutter life prediction model to obtain the residual service life of the current cutter.
The S6 specifically comprises the following steps:
inquiring a cutter number in a cutter library according to the model of a cutter used for a workpiece to be processed, and inquiring the accumulated loss service life U loss of the cutter according to the cutter number;
and (3) obtaining simulation parameters of the tool in the processing process of the workpiece through the S1, the S2 and the S3, substituting the simulation parameters into the tool life prediction model constructed in the S5 for calculation, and obtaining the residual service life expected Uin the current tool.
Further, the step S6 may further include: recording the theoretical life of the cutter as U, and if (U loss+U pre)/U is smaller than a first preset proportion, considering that the secondary cutter can be safely used in the processing; if (U loss+u pre)/U theory > = first preset proportion and (U loss+u pre)/U theory < = second preset proportion, giving an early warning in the MES system; if the (U loss+Upre)/U principle is greater than a second preset proportion, notifying the site to perform tool changing treatment; the first preset proportion is less than the second preset proportion < =100%. In this embodiment, the first preset ratio is 80% and the second preset ratio is 100%.
According to the invention, the NC program is used for acquiring and analyzing the parameter information of the workpiece, a cutter life prediction model is established, cutter life prediction is performed based on cutter track simulation, the prediction precision of cutter life is improved, and the prediction deviation of workpiece life is avoided.
It will be understood that modifications and variations will be apparent to those skilled in the art from the foregoing description, and it is intended that all such modifications and variations be included within the scope of the following claims.
Claims (9)
1. The cutter life prediction method based on the cutter path running track is characterized by comprising the following steps of:
s1, extracting processing information:
extracting all cutter processing paths and cutter processing parameters according to an NC program, wherein the cutter processing parameters comprise cutter type, cutting speed and feeding speed, and obtaining cutter numbers by butting with an MES system;
s2, reconstructing a workpiece machining process model:
calculating a process model of the geometry of the workpiece after each machining process according to the machining path of the cutter before each machining process based on the digital model of the cutter before each machining process;
s3, calculating cutting quantity:
dividing the cutter processing path according to unit time according to a process model of workpiece geometry and the cutter processing path, and calculating the cutting quantity of each time point of the cutter;
s4, storing the serialization parameters:
storing the cutting speed, the feeding speed and the cutting quantity of all the cutters in the machining process in an array mode;
s5, establishing a cutter life prediction model:
counting numbers of all scrapped tools in an MES system, combining S1-S4 to obtain historical processing parameter information of the scrapped tools, and establishing a tool life prediction model of the relation between the stress condition and the life of the tools based on deep learning according to the principle that the continuous work of the tools leads to deduction of the life; the S5 specifically comprises the following steps:
5.1, counting numbers of all scrapped cutters in the MES system;
5.2, obtaining historical processing parameter information of the scrapped cutter according to the calculation steps in S1, S2 and S3;
5.3, continuously working the cutter to lead the deduction service life to be converted into a polynomial of degree 2 for fitting, and obtaining the formula: Σ (a1+a2+ty+a0) (b1+tn+b2+tn+b0); tn identifies the continuous cutting time; a1, a2, ty, a0, b1, b2 and b0 are respectively polynomial coefficients to be solved and are obtained by training a deep learning network;
5.4 deep learning network adopts a combination of a multi-layer full-connection layer network and an LSTM network, wherein the full-connection layer network is used for carrying out formula fitting, and the LSTM is used for recursively influencing the service life of the current cutter by the service condition of the cutter; the input of the deep learning network is the historical processing duration time and the processing process cutting force of a single cutter, and the output of the full-connection layer is single processing life loss; the input of the LSTM network is the accumulation of the life loss of the output of the full-connection layer, and the output of the LSTM network is the total life loss of the cutter after the current machining is executed;
5.5 setting a deep learning model training iteration termination condition, and directly outputting a cutter life prediction model by a full-connection layer in the training model when the iteration termination condition is met;
s6, predicting the service life of the current cutter:
and predicting the processed simulation parameters of the current cutter by using a cutter life prediction model to obtain the residual service life of the current cutter.
2. The tool life prediction method based on the tool path running path according to claim 1, wherein the S1 is based on the ISO standard by reading NC files of data processing, parsing text data therein to obtain a tool processing path and tool processing parameters.
3. The tool life prediction method based on the tool path running track according to claim 2, wherein the tool processing path comprises a circular arc section and a straight line section, and the method is extracted according to the following steps:
inquiring the NC program using the cutter in an MES system according to the cutter number;
obtaining a linear interpolation motion starting point and a linear interpolation motion ending point by identifying a G01 instruction in an NC program;
obtaining a starting point, an ending point and a circular arc radius of circular arc interpolation motion by identifying a G02/G03 instruction in an NC program;
converting the circular arc and straight line parameters into three-dimensional curves and recording corresponding line numbers of each curve in an NC program;
obtaining tool movement radius correction parameters by identifying G41/G42 in an NC program, recording line numbers, and converting all correction parameters into a serial data format of [ { "line number": 12 }, { "line number": 42 }, "radius correction amount": -0.02} ];
obtaining cutter length correction parameters by identifying G43/G44 in an NC program, recording line numbers, and converting all the length correction parameters into a serial data format of [ { "line numbers": 20 ": length correction amount": 0.1}, { "line numbers": 50 ": radius correction amount": 0.0} ];
each row of correction parameters will influence the subsequent straight line interpolation and arc interpolation instructions until new correction parameters are replaced; and (3) offsetting the circular arc and the linear motion track according to the correction amount and the correction direction.
4. The tool life prediction method based on the path-movement trajectory according to claim 2, wherein the cutting speed is obtained by:
inquiring the diameter D of the cutter from a cutter database according to the cutter model;
obtaining a spindle rotation parameter by identifying an M03/M04 instruction in an NC program, and recording an M03/M04 instruction line number so as to obtain a spindle rotation number N;
according to the formula: cutting speed vc=pi×dn/1000, and converting the cutting speed parameter into a serial data format of [ { "line number": 10, "cutting speed": 0.05}, { "line number": 80, "cutting speed": 0.025} ].
5. The tool life prediction method based on the path-movement trajectory according to claim 2, wherein the feed speed is obtained by:
the spindle rotation parameters are obtained by identifying M03/M04 instructions in an NC program, and the line numbers of the M03/M04 instructions are recorded, so that the spindle rotation number N is obtained;
obtaining a cutter feeding speed F by identifying an F instruction in an NC program, and recording an F instruction line number;
according to the formula: feed speed v=n×f, calculate feed speed V, insert one feed speed each time there is a code line modified by N and F values: the feed rate conversion format is [ { "line number": 10, "feed rate": 0.2}, { "line number": 100, "feed rate": 0.5} ].
6. The tool life prediction method based on the tool path running track according to claim 3, wherein the step S2 specifically comprises:
inquiring a digital model of the cutter from a cutter database according to the cutter number;
creating a section line of the movement direction of the cutter based on the digital model of the cutter;
using the offset cutter machining path as a guide line to sweep a section line in the cutter movement direction, and creating a sweep model;
and performing difference solving by using the digitized model and the sweep model before workpiece processing to obtain a process model of the workpiece geometry after the current processing is completed.
7. The tool life prediction method based on the tool path running track according to claim 1, wherein the step S3 specifically comprises:
3.1, performing triangle gridding on a workpiece and a digital model of a cutter before the current processing;
3.2, converting the workpiece triangular mesh and the cutter triangular mesh into octree meshes with the precision of S, wherein the conversion process is as follows:
3.2.1, taking the center of the cutter as a reference, shifting X to the motion direction, wherein X is the simulation precision, so that the cutter is completely enveloped by the octree initial square;
3.2.2 if the octree mesh comprises any triangle in the triangle mesh and the current octree mesh size is greater than the conversion accuracy S, continuously subdividing the octree;
3.2.3 if the octree grid does not contain any triangle in the triangle grid or the octree grid size is smaller than the conversion precision S, ending the current grid subdivision task;
3.3, calculating the fitting condition of each cell in the workpiece octree grid and the cutter octree grid, wherein the number of the cells in the workpiece octree grid, which are fitted with the cutter octree grid, is Tg;
3.4, the pressed area of the cutter is positively correlated with Tg, the cutting quantity of the cutter is Tg x V, the cutting force Ty=k of the cutter is Tg x V, k is the hardness coefficient of the workpiece material, and V is the feeding speed;
3.5, setting an initial cutting point as A0, repeating the steps of 3.1-3.4 at intervals of L cutting distances, and performing simulation calculation on the stress condition once; each simulation point on the whole cutter processing path is marked with A0, A1 and A2 … … An;
3.6 comparing the stress difference ratio of A1 to A0: if the difference ratio is smaller than the standard value, the stress conditions of the two points are considered to be close to each other, and the two points are combined; processing time parameters of each stress section are obtained by combining the feeding speed of the cutter, and the combined serialization parameters are expressed as follows: { "processing time": 0,30 }, "cutting force": 200}; the cutting force is the average value of the cutting forces at the points A0 and A1;
3.7 continuously comparing the stress difference ratio of the point A2 and the point A0/A1: and if the difference ratio is smaller than the standard value, combining the difference ratio with A0 and A1, and expressing the combined serialization parameters as follows: { "processing time": 0,45 }, "cutting force": 210}, continuing to compare the ratio of the difference between the stress at point A3 and the combined stress at point A0/A1/A2 according to 3.7, and thus, until An; if the difference is larger than the standard value, taking A2 as a new initial cutting point A0, and taking each simulation point A3 … … An after the A2 as a new A1 … … An-2, and repeating for 3.6;
and 3.8, finally obtaining a processing procedure stress condition parameter array and submitting the array to an MES system database, wherein the serialization format is as follows: [ { "processing time": 0,45], "cutting force": 200}, { "processing time": 45,60], "cutting force": 210} ].
8. The tool life prediction method based on the tool path running track according to claim 1, wherein the step S6 specifically comprises:
inquiring a cutter number in a cutter library according to the model of a cutter used for a workpiece to be processed, and inquiring the accumulated loss service life U loss of the cutter according to the cutter number;
and (3) obtaining simulation parameters of the tool in the processing process of the workpiece through the S1, the S2 and the S3, substituting the simulation parameters into the tool life prediction model constructed in the S5 for calculation, and obtaining the residual service life expected Uin the current tool.
9. The method for predicting the life of a tool based on a path trajectory as claimed in claim 1, wherein said S6 further comprises: recording the theoretical life of the cutter as U, and if (U loss+U pre)/U is smaller than a first preset proportion, considering that the secondary cutter can be safely used in the processing; if (U loss+u pre)/U theory > = first preset proportion and (U loss+u pre)/U theory < = second preset proportion, giving an early warning in the MES system; if the (U loss+Upre)/U principle is greater than a second preset proportion, notifying the site to perform tool changing treatment;
the first preset proportion is less than the second preset proportion < =100%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210713491.9A CN115167276B (en) | 2022-06-22 | 2022-06-22 | Cutter life prediction method based on cutter path running track |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210713491.9A CN115167276B (en) | 2022-06-22 | 2022-06-22 | Cutter life prediction method based on cutter path running track |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115167276A CN115167276A (en) | 2022-10-11 |
| CN115167276B true CN115167276B (en) | 2023-05-23 |
Family
ID=83486441
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210713491.9A Active CN115167276B (en) | 2022-06-22 | 2022-06-22 | Cutter life prediction method based on cutter path running track |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115167276B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116165968A (en) * | 2023-04-24 | 2023-05-26 | 成都航利航空科技有限责任公司 | Numerical control procedure processing parameter recording method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107728577B (en) * | 2017-09-28 | 2019-09-27 | 大连理工大学 | Instantaneous cutting output planing method based on thin-wall curved-surface machining deformation |
| TWI650625B (en) * | 2017-11-16 | 2019-02-11 | 財團法人工業技術研究院 | Tool wear detecting device, detecting method thereof and tool wear compensation method |
| CN109465676B (en) * | 2018-09-13 | 2020-10-16 | 成都数之联科技有限公司 | Tool life prediction method |
| CN109724785A (en) * | 2018-12-29 | 2019-05-07 | 中铁工程装备集团有限公司 | A kind of tool condition monitoring and life prediction system based on Multi-source Information Fusion |
| CN110900307B (en) * | 2019-11-22 | 2020-12-15 | 北京航空航天大学 | Numerical control machine tool cutter monitoring system driven by digital twin |
| CN110948287B (en) * | 2019-12-07 | 2022-03-11 | 上海爱堃智能系统有限公司 | Intelligent cutter management method and system |
| CN111832624A (en) * | 2020-06-12 | 2020-10-27 | 上海交通大学 | Tool remaining life prediction method based on anti-migration learning |
-
2022
- 2022-06-22 CN CN202210713491.9A patent/CN115167276B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115167276A (en) | 2022-10-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115167276B (en) | Cutter life prediction method based on cutter path running track | |
| CN114237155B (en) | Error prediction and compensation method, system and medium for multi-axis numerical control machining | |
| CN101477351B (en) | Intelligent numerical control method with three-stage process self-optimization function | |
| CN113341882B (en) | Numerical control process design and optimization method based on processing knowledge | |
| CN109176168B (en) | Tool sharpening tune machine device and method | |
| CN105488282A (en) | Cutting parameter segmentation and variable cutting depth optimizing method based on dynamic machining feature | |
| CN102455679A (en) | Machining time predicting apparatus of numerically controlled machine tool | |
| CN109249284B (en) | Tool sharpening parameter compensation device and method | |
| CN113051683B (en) | Numerical control machine tool life prediction method, system, equipment and storage medium | |
| CN112147950B (en) | Numerical control side milling system and accurate synchronization method of threshold value and signal to be monitored thereof | |
| CN112100810B (en) | Flat-bottom milling cutter multi-axis milling force prediction method | |
| CN104318022A (en) | Method for predicting workpiece surface roughness and increasing cutting efficiency | |
| CN111724203B (en) | Product processing price estimation method and system based on multiple regression model | |
| Chu et al. | Generation of reciprocating tool motion in 5-axis flank milling based on particle swarm optimization | |
| CN111650885A (en) | Automatic programming method for numerical control machine tool | |
| CN110069379A (en) | The screening technique and screening plant of monitor control index | |
| CN111736530B (en) | Method and system for simulating tool wear morphology in machining process | |
| CN113869625A (en) | Cutter optimization evaluation model and method | |
| CN110083126A (en) | A kind of die industry machining period assessment system, computation model and calculation method | |
| CN114862287B (en) | Risk benefit analysis method, system, terminal and medium for cascade power station group scheduling | |
| CN113885439B (en) | Blade curvature identification and program optimization method based on machine tool indexing | |
| CN111167900B (en) | Method and device for determining parameters of bent pipe machining process | |
| CN117195152A (en) | Under-dam tributary jacking condition analysis system based on deep learning | |
| CN116796974A (en) | Additive manufacturing and scheduling method, device, equipment and storage medium | |
| Balic et al. | Model of an integrated intelligent design and manufacturing system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |