CN116690313B - Failure monitoring method for machining cutter of web plate of aircraft structural member - Google Patents
Failure monitoring method for machining cutter of web plate of aircraft structural member Download PDFInfo
- Publication number
- CN116690313B CN116690313B CN202310953555.7A CN202310953555A CN116690313B CN 116690313 B CN116690313 B CN 116690313B CN 202310953555 A CN202310953555 A CN 202310953555A CN 116690313 B CN116690313 B CN 116690313B
- Authority
- CN
- China
- Prior art keywords
- data
- monitoring
- classification model
- cutter
- web
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 154
- 238000000034 method Methods 0.000 title claims abstract description 88
- 238000003754 machining Methods 0.000 title claims abstract description 27
- 238000012545 processing Methods 0.000 claims abstract description 138
- 238000013145 classification model Methods 0.000 claims abstract description 133
- 238000012549 training Methods 0.000 claims abstract description 84
- 238000012360 testing method Methods 0.000 claims abstract description 22
- 230000011218 segmentation Effects 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 20
- 238000012706 support-vector machine Methods 0.000 claims description 7
- 101100241173 Caenorhabditis elegans dat-1 gene Proteins 0.000 claims description 3
- 230000035772 mutation Effects 0.000 abstract description 4
- 239000000463 material Substances 0.000 abstract description 3
- 238000005520 cutting process Methods 0.000 description 5
- 230000004927 fusion Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 2
- 238000003801 milling Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 239000000956 alloy Substances 0.000 description 1
- 229910045601 alloy Inorganic materials 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
- B23Q17/00—Arrangements for observing, indicating or measuring on machine tools
- B23Q17/09—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool
- B23Q17/0952—Arrangements for observing, indicating or measuring on machine tools for indicating or measuring cutting pressure or for determining cutting-tool condition, e.g. cutting ability, load on tool during machining
- B23Q17/0957—Detection of tool breakage
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a failure monitoring method for a machining cutter of a web plate of an aircraft structural member, which belongs to the technical field of numerical control machining of aviation materials and is characterized by comprising the following steps: s1, performing model training by using historical data, specifically referring to: inputting characteristic values of test batch data, calculating to obtain cutter states through a classification model, comparing the cutter states with label data, and calculating to obtain accuracy of accurate identification; judging the precision of the classification model; determining a signal characteristic value selection rule and storing classification model training data; s2, monitoring through a real-time data driving classification model. According to the web processing characteristics, the method and the device for monitoring the web processing tool state, the tool state monitoring strategy is formulated according to the web processing characteristics, the interference of the working condition mutation on the monitoring result is reduced, the general web processing tool monitoring model and strategy which can be suitable for different machine tools, tools with multiple types and specifications and time-varying parameters are built, and the problem that the learning model is low in accuracy caused by complex tool parameter combination and time-varying working condition data is solved.
Description
Technical Field
The invention relates to the technical field of numerical control machining of aviation materials, in particular to a failure monitoring method for a machining cutter of a web plate of an aircraft structural member.
Background
In the cutting process of aviation materials, the cutter can be gradually worn or even broken along with the extension of service time. The tool is in direct contact with the workpiece, and excessive wear and tear of the tool will reduce the dimensional accuracy and surface quality of the part. Therefore, during the machining process, the state of the tool needs to be paid attention to in real time, and the tool needs to be replaced in time when the tool is worn to a certain degree or damaged.
In the machining process of the aircraft structural part, web milling is an important process of numerical control machining of the structural part, milling of web features is required in rough machining, semi-finishing machining and finishing machining, most blank allowance is removed, and tools including indexable blades and integral hard alloy tools are used. The monitoring of the cutter state usually adopts a mode of learning and then comparing to carry out abnormal monitoring, but for the processing modes of various and small batches of airplane structural members, the data size for model learning is small, and signals are easily influenced by the fluctuation of the processing state in the processing process of complex parts of the airplane structural members, so that false alarms are frequently generated in the application process, and normal production is influenced. How to establish a tool state monitoring method which uses small sample size data and has high accuracy is particularly important aiming at the processing working conditions of small batches, multiple varieties, high discrete degree, complex structure, multiple tool types and changeable tracks of the aircraft structural member.
The Chinese patent document with publication number of CN110561193A and publication date of 2019, 12 and 13 discloses a method and a system for evaluating and monitoring cutter wear based on feature fusion, wherein the method comprises the following steps: the method comprises the steps of flexible data acquisition, working condition segmentation, data preprocessing, feature extraction, feature fusion and evaluation monitoring. The system comprises a working condition segmentation module, a data preprocessing module, a characteristic extraction module, a characteristic fusion module, an evaluation monitoring module and a self-learning module which are sequentially connected in series.
The method for evaluating and monitoring the cutter wear by the feature fusion disclosed in the patent document aims at working conditions of the same parts, the same processing features, the same cutters and the same parameters in a large-batch processing model, and numerical control processing of an aircraft structural member is a small-batch multi-variety processing mode, if the monitoring objects are divided according to the working conditions, more than 1000 monitoring objects are generated, large sample amount data required by model learning cannot be provided, and the accuracy of a learning model is reduced due to complex cutter parameter combination and time-varying working condition data.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a method for monitoring the failure of a web processing cutter of an aircraft structural member, which is used for making a cutter state monitoring strategy according to web processing characteristics, reducing the interference of working condition mutation on a monitoring result, establishing a general web processing cutter monitoring model and strategy applicable to different machine tools, multi-type specification cutters and time-varying parameters, and solving the problem of low accuracy of a learning model caused by complex cutter parameter combination and time-varying working condition data.
The invention is realized by the following technical scheme:
the method for monitoring the failure of the machining cutter of the web plate of the aircraft structural member is characterized by comprising the following steps of:
s1, performing model training by using historical data, specifically referring to:
s101, collecting vibration signals of a web plate processing process of an aircraft structural member;
s102, calculating the root mean square sliding average value of each rotation of the vibration signal;
s103, judging the tool path characteristics and denoising;
s104, monitoring signal segmentation;
s105, calculating a signal characteristic value;
s106, dividing the data into a training group and a test group;
s107, selecting a signal characteristic value;
s108, selecting a support vector machine classification model, and training to obtain the classification model by using the characteristic values and the label data of the training group batch data;
s109, inputting characteristic values of test batch data, calculating to obtain a cutter state through a classification model, comparing the cutter state with label data, and calculating to obtain accuracy of accurate identification;
s110, judging the precision of the classification model;
the method comprises the steps of presetting precision, judging whether a classification model precision calculation value meets the preset precision, and if the classification model precision is greater than the preset precision, entering S111; if the precision of the classification model is smaller than the preset precision, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision is met;
s111, determining a signal characteristic value selection rule and storing classification model training data;
s2, monitoring through a real-time data driving classification model, specifically:
s201, when the process programming is carried out, setting web processing content in one process step, and adding an identification code in a digital control program of the process step;
s202, starting a cutter monitoring system when a program runs, and starting a web processing cutter state monitoring system after identifying a web processing step identification code in the part processing process;
s203, monitoring and dynamically updating the classification model;
s204, starting a web processing cutter for real-time monitoring;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
s207, monitoring signal segmentation, namely segmenting to obtain batch data by adopting the number N of root mean square values per rotation contained in each batch of data which is the same as that in the training stage of the classification model;
s208, calling model identification;
s209, calculating a signal duty ratio A of the identified tipping section;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value or not;
s211, monitoring and alarming, and controlling the machine tool.
In step S206, the tool path feature determination and denoising, specifically, separating the tool paths processed by each layer of web according to G0 and G1 codes, performing the tool path feature determination and denoising by using each layer of web processing tool path as a data object, and calculating to obtain the minimum Z coordinate value Z in the processing process of one layer of web min Setting Z value deviation DeltaZ, and processing each layer of web plate to meet Z min -△Z＜Z＜Z min +△And cutting out the tool path of Z and the corresponding processing signals as a data object for monitoring the web processing tool state.
In step S207, the monitoring signal segmentation specifically refers to setting the number of root mean square values per turn included in each batch of data as N, and if the number of root mean square values per turn included in each web processing monitoring data object is M, dividing each batch of monitoring data into K batches, where K is a value obtained by rounding M/N.
In step S104, the monitoring signal segmentation specifically refers to performing signal characteristic value calculation on each batch of data obtained, including deviation, kurtosis, standard deviation, mean value and maximum time-frequency domain signal characteristic value of root mean square data of each revolution.
In step S111, determining a rule for selecting signal feature values and storing training data of the classification model specifically refers to recording the signal feature values selected by the constructed classification model, storing training set data used in training the model, including signals corresponding to the states of intact cutters and failure cutters, using the same rule for selecting signal feature values when the features of the classification model are updated in the monitoring stage in real time, and reconstructing the classification model after processing the data stored in the rule.
In step S203, the monitoring of the dynamic update of the classification model specifically includes starting the classification model update module when the new tool is used, obtaining the first n layers of signals of the new tool cutting, performing the equal proportion processing on the training data of the classification model by using the real-time vibration signals, and reconstructing the classification model based on the updated data.
The method specifically comprises the steps of starting a classification model updating module when the new cutter is identified to be used, acquiring the front n-layer signals of web processing by the new cutter, and extracting each layer of signals by using the cutter rail characteristics in the classification model training stage to judge and denoise.
The use of real-time vibration signals to perform equal proportion processing on the training data of the classification model specifically refers to the use of real-time data to calculate and obtain the average value M of the root mean square of each turn of web processing signals of each layer 2 Using labels of the classification model training stage as tool healthData, calculating and obtaining the average value M of root mean square per rotation of each layer of web processing signal 1 Calculating to obtain signal proportion m=m 2 /M 1 And reading the root mean square Data1 of each rotation saved in the model training stage, and obtaining updated root mean square Data 2=Dat1×m of each rotation by using signal proportion calculation.
The reconstruction of the classification model based on the updated Data specifically refers to reconstructing the tool state monitoring model by using the new signal characteristic value by using the root mean square Data2 of each revolution obtained after updating according to the signal characteristic value selection rule determined in the training stage of the classification model and calculating to obtain the corresponding signal characteristic value.
In step S204, starting the web processing tool to monitor in real time specifically means that the monitoring model is used to determine the state of the tool when the next web is processed, and the processing process data of one web is used as the object, and the result of the determination is that the tool is intact or damaged.
In step S208, the calling model identification specifically refers to reading the calculated signal characteristic value for each batch of data, calling the monitoring model to identify, and outputting the discrimination result of each batch of data.
In step S209, the calculation of the signal duty ratio a identified as the tipping segment specifically refers to calculating the number Nt of batch data identified as the tipping segment according to the model output result, the number Nf of batch data contained in each layer of web processing data, and calculating the signal duty ratio a=nt/Nf identified as the tipping segment.
In the step S210, determining whether the signal duty ratio a of the tipping segment is greater than the monitoring threshold specifically means that all batch data of a web processing process is selected as an object to perform tool state determination, the signal duty ratio a of the tipping segment is selected as a monitoring index value, a tool state alarm threshold value Athr is set, if a is less than or equal to Athr, the step S205 is returned to execute the next web processing state identification, and if a is greater than Athr, a tool state alarm signal is sent.
The G0 is an unconditional jump instruction.
The invention discloses a linear feeding command G1.
The beneficial effects of the invention are mainly shown in the following aspects:
1. the invention S1, model training is carried out by using historical data, specifically: inputting characteristic values of test batch data, calculating to obtain cutter states through a classification model, comparing the cutter states with label data, and calculating to obtain accuracy of accurate identification; judging the precision of the classification model; determining a signal characteristic value selection rule and storing classification model training data; s2, monitoring is carried out through a real-time data driving classification model, compared with the prior art, a cutter state monitoring strategy is formulated according to web processing characteristics, interference of working condition mutation on monitoring results is reduced, a general web processing cutter monitoring model and strategy which can be suitable for different machine tools, multi-type specification cutters and time-varying parameters are built, and the problem that the learning model is low in accuracy caused by complex cutter parameter combination and time-varying working condition data is solved.
2. According to the invention, the monitoring model is dynamically updated based on real-time data and by using rules and data stored in a training stage, and the signal segmentation monitoring is carried out on each layer of web processing process which uses the cutter, so that the state of the cutter can be effectively monitored.
3. According to the invention, the duty ratio threshold of the tipping signal is set, and the alarm is given and the machine tool is controlled after the signal exceeds the threshold, so that the damage of parts or equipment can be effectively prevented.
4. Compared with the Chinese patent document with the publication number of CN110561193A and the publication date of 2019, 12 and 13, the invention can realize accurate on-line identification of the states of cutters under different types and specifications of cutters and different combination parameter working conditions in the web processing process.
Detailed Description
Example 1
A method for monitoring failure of a machining cutter of a web plate of an aircraft structural member comprises the following steps:
s1, performing model training by using historical data, specifically referring to:
s101, collecting vibration signals of a web plate processing process of an aircraft structural member;
s102, calculating the root mean square sliding average value of each rotation of the vibration signal;
s103, judging the tool path characteristics and denoising;
s104, monitoring signal segmentation;
s105, calculating a signal characteristic value;
s106, dividing the data into a training group and a test group;
s107, selecting a signal characteristic value;
s108, selecting a support vector machine classification model, and training to obtain the classification model by using the characteristic values and the label data of the training group batch data;
s109, inputting characteristic values of test batch data, calculating to obtain a cutter state through a classification model, comparing the cutter state with label data, and calculating to obtain accuracy of accurate identification;
s110, judging the precision of the classification model;
the method comprises the steps of presetting precision, judging whether a classification model precision calculation value meets the preset precision, and if the classification model precision is greater than the preset precision, entering S111; if the precision of the classification model is smaller than the preset precision, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision is met;
s111, determining a signal characteristic value selection rule and storing classification model training data;
s2, monitoring through a real-time data driving classification model, specifically:
s201, when the process programming is carried out, setting web processing content in one process step, and adding an identification code in a digital control program of the process step;
s202, starting a cutter monitoring system when a program runs, and starting a web processing cutter state monitoring system after identifying a web processing step identification code in the part processing process;
s203, monitoring and dynamically updating the classification model;
s204, starting a web processing cutter for real-time monitoring;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
s207, monitoring signal segmentation, namely segmenting to obtain batch data by adopting the number N of root mean square values per rotation contained in each batch of data which is the same as that in the training stage of the classification model;
s208, calling model identification;
s209, calculating a signal duty ratio A of the identified tipping section;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value or not;
s211, monitoring and alarming, and controlling the machine tool.
Compared with the prior art, the method and the device for monitoring the web processing tool have the advantages that a tool state monitoring strategy is formulated according to web processing characteristics, interference of working condition mutation on monitoring results is reduced, a general web processing tool monitoring model and strategy which can be suitable for different machine tools, tools with multiple types and time-varying parameters are built, and the problem that the learning model is low in accuracy caused by complex tool parameter combination and time-varying working condition data is solved.
Example 2
A method for monitoring failure of a machining cutter of a web plate of an aircraft structural member comprises the following steps:
s1, performing model training by using historical data, specifically referring to:
s101, collecting vibration signals of a web plate processing process of an aircraft structural member;
s102, calculating the root mean square sliding average value of each rotation of the vibration signal;
s103, judging the tool path characteristics and denoising;
s104, monitoring signal segmentation;
s105, calculating a signal characteristic value;
s106, dividing the data into a training group and a test group;
s107, selecting a signal characteristic value;
s108, selecting a support vector machine classification model, and training to obtain the classification model by using the characteristic values and the label data of the training group batch data;
s109, inputting characteristic values of test batch data, calculating to obtain a cutter state through a classification model, comparing the cutter state with label data, and calculating to obtain accuracy of accurate identification;
s110, judging the precision of the classification model;
the method comprises the steps of presetting precision, judging whether a classification model precision calculation value meets the preset precision, and if the classification model precision is greater than the preset precision, entering S111; if the precision of the classification model is smaller than the preset precision, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision is met;
s111, determining a signal characteristic value selection rule and storing classification model training data;
s2, monitoring through a real-time data driving classification model, specifically:
s201, when the process programming is carried out, setting web processing content in one process step, and adding an identification code in a digital control program of the process step;
s202, starting a cutter monitoring system when a program runs, and starting a web processing cutter state monitoring system after identifying a web processing step identification code in the part processing process;
s203, monitoring and dynamically updating the classification model;
s204, starting a web processing cutter for real-time monitoring;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
s207, monitoring signal segmentation, namely segmenting to obtain batch data by adopting the number N of root mean square values per rotation contained in each batch of data which is the same as that in the training stage of the classification model;
s208, calling model identification;
s209, calculating a signal duty ratio A of the identified tipping section;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value or not;
s211, monitoring and alarming, and controlling the machine tool.
Further, in the step S206, the tool path feature determination and denoising are specifically performed on the tool path for processing each layer of web according to the G0 and G1 codesSeparating, taking each layer of web processing cutter track as a data object to judge cutter track characteristics and denoise, and calculating to obtain a Z coordinate minimum Z in the web processing process min Setting Z value deviation DeltaZ, and processing each layer of web plate to meet Z min -△Z＜Z＜Z min The tool path with the angle delta Z and the corresponding processing signals are intercepted and used as the data object for monitoring the web processing tool state.
In step S207, the monitoring signal segmentation specifically refers to setting the number of root mean square values per turn included in each batch of data as N, and if the number of root mean square values per turn included in each web processing monitoring data object is M, dividing each batch of monitoring data into K batches, where K is a value obtained by rounding M/N.
In step S104, the monitoring signal segmentation specifically refers to performing signal characteristic value calculation on each batch of data obtained, including deviation, kurtosis, standard deviation, mean value and maximum time-frequency domain signal characteristic value of root mean square data of each revolution.
In step S111, determining a rule for selecting signal feature values and storing training data of the classification model specifically refers to recording the signal feature values selected by the constructed classification model, storing training set data used in training the model, including signals corresponding to the states of intact cutters and failure cutters, using the same rule for selecting signal feature values when the features of the classification model are updated in the monitoring stage in real time, and reconstructing the classification model after processing the data stored in the rule.
In this embodiment, the monitoring model is dynamically updated based on real-time data and using rules and data stored in a training stage, so that the state of the tool can be effectively monitored by performing signal segment monitoring on each layer of web processing process which uses the tool.
Example 3
A method for monitoring failure of a machining cutter of a web plate of an aircraft structural member comprises the following steps:
s1, performing model training by using historical data, specifically referring to:
s101, collecting vibration signals of a web plate processing process of an aircraft structural member;
s102, calculating the root mean square sliding average value of each rotation of the vibration signal;
s103, judging the tool path characteristics and denoising;
s104, monitoring signal segmentation;
s105, calculating a signal characteristic value;
s106, dividing the data into a training group and a test group;
s107, selecting a signal characteristic value;
s108, selecting a support vector machine classification model, and training to obtain the classification model by using the characteristic values and the label data of the training group batch data;
s109, inputting characteristic values of test batch data, calculating to obtain a cutter state through a classification model, comparing the cutter state with label data, and calculating to obtain accuracy of accurate identification;
s110, judging the precision of the classification model;
the method comprises the steps of presetting precision, judging whether a classification model precision calculation value meets the preset precision, and if the classification model precision is greater than the preset precision, entering S111; if the precision of the classification model is smaller than the preset precision, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision is met;
s111, determining a signal characteristic value selection rule and storing classification model training data;
s2, monitoring through a real-time data driving classification model, specifically:
s201, when the process programming is carried out, setting web processing content in one process step, and adding an identification code in a digital control program of the process step;
s202, starting a cutter monitoring system when a program runs, and starting a web processing cutter state monitoring system after identifying a web processing step identification code in the part processing process;
s203, monitoring and dynamically updating the classification model;
s204, starting a web processing cutter for real-time monitoring;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
s207, monitoring signal segmentation, namely segmenting to obtain batch data by adopting the number N of root mean square values per rotation contained in each batch of data which is the same as that in the training stage of the classification model;
s208, calling model identification;
s209, calculating a signal duty ratio A of the identified tipping section;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value or not;
s211, monitoring and alarming, and controlling the machine tool.
In step S206, the tool path feature determination and denoising, specifically, separating the tool paths processed by each layer of web according to G0 and G1 codes, performing the tool path feature determination and denoising by using each layer of web processing tool path as a data object, and calculating to obtain the minimum Z coordinate value Z in the processing process of one layer of web min Setting Z value deviation DeltaZ, and processing each layer of web plate to meet Z min -△Z＜Z＜Z min The tool path with the angle delta Z and the corresponding processing signals are intercepted and used as the data object for monitoring the web processing tool state. In step S207, the monitoring signal segmentation specifically refers to setting the number of root mean square values per turn included in each batch of data as N, and if the number of root mean square values per turn included in each web processing monitoring data object is M, dividing each batch of monitoring data into K batches, where K is a value obtained by rounding M/N. In step S104, the monitoring signal segmentation specifically refers to performing signal characteristic value calculation on each batch of data obtained, including deviation, kurtosis, standard deviation, mean value and maximum time-frequency domain signal characteristic value of root mean square data of each revolution. In step S111, determining a rule for selecting signal feature values and storing training data of the classification model specifically refers to recording the signal feature values selected by the constructed classification model, storing training set data used in training the model, including signals corresponding to the complete and failure states of the tool, using the same rule for selecting signal feature values when the features of the classification model are updated during the monitoring stage in real time, and using the stored data for processingAnd reconstructing the classification model. In step S203, the monitoring of the dynamic update of the classification model specifically includes starting the classification model update module when the new tool is used, obtaining the first n layers of signals of the new tool cutting, performing the equal proportion processing on the training data of the classification model by using the real-time vibration signals, and reconstructing the classification model based on the updated data.
Further, the acquiring the first n layers of signals of the new cutter cutting specifically means that when the new cutter is identified to be used, a classification model updating module is started, the first n layers of signals of the new cutter for web processing are acquired, and the cutter track feature discrimination and denoising are performed in the classification model training stage for extracting each layer of signals.
The use of real-time vibration signals to perform equal proportion processing on the training data of the classification model specifically refers to the use of real-time data to calculate and obtain the average value M of the root mean square of each turn of web processing signals of each layer 2 Using the label of the training stage of the classification model as the data of the intact state of the cutter, calculating and obtaining the average value M of the root mean square of each rotation of the web processing signals of each layer 1 Calculating to obtain signal proportion m=m 2 /M 1 And reading the root mean square Data1 of each rotation saved in the model training stage, and obtaining updated root mean square Data 2=Dat1×m of each rotation by using signal proportion calculation.
The reconstruction of the classification model based on the updated Data specifically refers to reconstructing the tool state monitoring model by using the new signal characteristic value by using the root mean square Data2 of each revolution obtained after updating according to the signal characteristic value selection rule determined in the training stage of the classification model and calculating to obtain the corresponding signal characteristic value.
Further, in the step S204, starting the web processing tool to monitor in real time specifically means that the monitoring model is used to determine the state of the tool when the next layer of web is processed, and the data of the web processing process of one layer is used as the object, and the result of the determination is that the tool is intact or damaged.
In step S208, the calling model identification specifically refers to reading the calculated signal characteristic value for each batch of data, calling the monitoring model to identify, and outputting the discrimination result of each batch of data.
In step S209, the calculation of the signal duty ratio a identified as the tipping segment specifically refers to calculating the number Nt of batch data identified as the tipping segment according to the model output result, the number Nf of batch data contained in each layer of web processing data, and calculating the signal duty ratio a=nt/Nf identified as the tipping segment.
In the step S210, determining whether the signal duty ratio a of the tipping segment is greater than the monitoring threshold specifically means that all batch data of a web processing process is selected as an object to perform tool state determination, the signal duty ratio a of the tipping segment is selected as a monitoring index value, a tool state alarm threshold value Athr is set, if a is less than or equal to Athr, the step S205 is returned to execute the next web processing state identification, and if a is greater than Athr, a tool state alarm signal is sent.
In the embodiment, the optimal implementation mode is that the tipping signal duty ratio threshold is set, and after the signal exceeds the threshold, an alarm is given and the machine tool is controlled, so that damage to parts or equipment can be effectively prevented.
The method can realize accurate online identification of the states of the cutters under different types and specifications of cutters and different combination parameter working conditions in the web processing process.
The following describes the implementation of the present invention:
s1, performing model training phase by using historical data
S101, collecting vibration signals of web plate processing process of airplane structural part
Acquiring a machining process vibration signal by using a vibration acceleration sensor, wherein the sensor is arranged on a machine tool spindle, acquiring a three-way machining process vibration signal, recording the cutter state corresponding to the acquired web machining vibration data, classifying the cutter state after measurement, and recording in a form of a label value;
s102, calculating the root mean square of each rotation of the vibration signal according to the root mean square of each rotation of the vibration signal, wherein the calculating method is as follows:
number of vibration signal samples per rotation of machine tool spindleThe calculation is as follows:
wherein,for the sampling frequency of the vibration signal, < >>The spindle rotation speed;
machine tool spindlekRoot mean square value per revolutionObtained by calculation of the formula:
wherein,is the first of the acquired vibration signalsiA number of values.
Then calculate the firstmSliding average of root mean square value per revolutionAs shown in the following formula;
wherein,is the first obtained by calculationmA running average of root mean square values per revolution,jnumber of values per cycle calculated as a running average, +.>For the machine tool spindle>Root mean square value per revolution;
obtaining a sliding average value by using the root mean square value of each turn of the vibration signal as a data source for identifying the failure of the cutter in the subsequent step;
s103, knife track feature discrimination and denoising
Separating the tool path processed by each layer of web according to G0 and G1 codes, and taking each layer of web processing tool path as a data object to perform tool path characteristic discrimination and denoising;
calculating to obtain Z coordinate minimum Z in the processing process of one layer of web min =216, and considering the uneven web surface, setting a Z value deviation ΔZ=2, and processing each layer of web to satisfy Z min -△Z＜Z＜Z min The tool path with the angle delta Z and the corresponding processing signals are intercepted and used as a data object for monitoring the state of a web processing tool;
s104, monitoring signal segmentation
On the basis of the data object of each layer of web processing cutter state monitoring obtained through cutter track characteristics, carrying out batch processing on the data set for subsequent signal characteristic value calculation;
setting the number of root mean square values per rotation contained in each batch of data as N=28, and if the number of root mean square values per rotation contained in each layer of web processing monitoring data object is M=13445, dividing each layer of monitoring data into K batches, wherein K is a value obtained by rounding M/N, and K=480;
s105, calculating signal characteristic value
Calculating signal characteristic values for each batch of data, wherein the signal characteristic values comprise deviation, kurtosis, standard deviation, mean value and maximum time-frequency domain signal characteristic values of root mean square data of each rotation;
s106, dividing the data into a training group and a test group
Dividing all signal batch data into a training group and a testing group, wherein the training group and the testing group respectively comprise a certain amount of data corresponding to the cutter intact and cutter damage state labels, the number of the testing data group accounts for 30% of the total number during classification, and the data corresponding to the cutter intact and cutter damage state labels of the training group and the testing group account for half;
s107, selecting signal characteristic value
In the example, selecting and calculating the maximum value, the mean value, the standard deviation, the deviation and the kurtosis as signal characteristic values;
s108: modeling using support vector machines
Selecting a support vector machine classification model, and training to obtain the classification model by using signal characteristic values and tag data of training group batch data;
s109: testing classification model accuracy
Inputting signal characteristic values of test batch data by using the classification model constructed in the step S108, calculating to obtain a cutter state, comparing the cutter state with the label data, and calculating to obtain the accuracy of accurate identification;
s110, discriminating precision of classification model
Judging whether the precision calculation value of the classification model meets the preset precision requirement or not, and if the precision of the classification model is greater than the preset precision requirement, entering S111; if the precision of the classification model is smaller than the preset precision requirement, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision requirement is met;
s111, determining a signal characteristic value selection rule and storing classification model training data
Recording signal characteristic values selected by constructing the classification model on the basis that the constructed classification model meets the preset precision requirement through testing, storing training set data used in training the classification model, using the same signal characteristic value selection rule when monitoring model characteristic updating in a real-time monitoring stage, and reconstructing the classification model after processing the stored data;
s2, monitoring stage through real-time data driving classification model
S201, adding identification during process programming
When the process programming is carried out, web processing content is set in one process step, and an identification code is added in a digital control program of the process step, so that the process step is analyzed and identified as a web processing process step in real-time monitoring;
s202, starting cutter monitoring system during program running
In the part machining process, after the web machining step identification code is identified, starting a web machining cutter state monitoring system;
s203: monitoring model dynamic updates
After the web processing cutter state monitoring system is started, a monitoring model dynamic updating module is firstly executed, a classification model updating module is triggered when the use of a new cutter is identified, the front n layers of signals of the new cutter for web processing are obtained, n=3 is set, and the extraction of each layer of signals uses the cutter track characteristic distinguishing and denoising method in the model training stage;
obtaining an average value M of root mean square per revolution of each layer web processing signal using real-time data calculation 2 Calculating and obtaining the average value M of root mean square per rotation of each layer of web processing signal by using labels in model training stage as data of cutter perfect state 1 Then the signal proportion m=m is calculated 2 /M 1 ；
Reading root mean square Data1 of each rotation stored in a model training stage, calculating to obtain Data2=Data1 by using signal proportion, calculating to obtain corresponding signal characteristic values according to a signal characteristic value selection rule determined in a classification model training stage, and reconstructing a cutter state monitoring model by using new signal characteristic values;
s204, starting web processing cutter real-time monitoring
For each new cutter, starting real-time monitoring after the monitoring model is dynamically updated, namely judging the state of the cutter by using the classification model when the next layer of web is processed, and outputting a judging result: the cutter is intact or damaged;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
the method for discriminating and denoising the cutter track features in the training stage of the classification model is mainly used for removing the interference of cutter feeding and retracting signals on monitoring results;
s207, monitoring signal segmentation
The same root mean square value number N per rotation contained in each batch of data as the training stage of the classification model, namely N=28, is adopted to obtain batch data in a segmented mode;
s208, calling model identification
For each batch of data, reading the calculated signal characteristic value, calling a monitoring model to identify, and outputting a discrimination result of each batch of data;
s209, calculating the signal duty ratio A of the identified tipping section
Calculating the number Nt of batch data identified as the tipping section according to the model output result, wherein the number Nf of batch data contained in each layer of web processing data is calculated, and the signal duty ratio A=Nt/Nf of the tipping section can be obtained;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value
Selecting all batch data of a web processing process as an object to judge the state of the cutter, selecting the signal proportion A of the tipping section as a monitoring index value, and setting a cutter state alarm threshold value Athr=90%;
if A is less than or equal to Athr, returning to S205, and executing the next-layer web processing state identification;
if A is more than Athr, a tool state alarm signal is sent out, and the machine tool is controlled to be stopped, so that the tool is prevented from being further severely worn or damaged, and damage to parts and equipment is avoided.
For parts with more web forms and characteristics, the change of the web processing tool track is frequent, a monitoring threshold is increased, the number of layers of web continuous monitoring alarm is set to be Ncthr=5, if the number of layers exceeds 5, an alarm signal is sent to a machine tool, otherwise, the monitoring is continuously executed;
s211, monitoring and alarming and controlling machine tool
After receiving the alarm signal, the machine tool is controlled to execute the alarm function and stop operation, and damage of parts and equipment is prevented.
Claims (5)
1. The method for monitoring the failure of the machining cutter of the web plate of the aircraft structural member is characterized by comprising the following steps of:
s1, performing model training by using historical data, specifically referring to:
s101, collecting vibration signals of a web plate processing process of an aircraft structural member;
s102, calculating the root mean square sliding average value of each rotation of the vibration signal;
s103, judging the tool path characteristics and denoising;
separating the tool path processed by each layer of web, and taking each layer of web processing tool path as a data object to perform tool path characteristic discrimination and denoising;
s104, monitoring signal segmentation;
on the basis of the data object of each layer of web processing cutter state monitoring obtained through cutter track characteristics, carrying out batch processing on the data set for subsequent signal characteristic value calculation;
s105, calculating a signal characteristic value;
s106, dividing the data into a training group and a test group;
dividing all signal batch data into a training group and a testing group, wherein the training group and the testing group respectively comprise a certain amount of data corresponding to the intact cutter and the damaged cutter state labels;
s107, selecting a signal characteristic value;
s108, selecting a support vector machine classification model, and training to obtain the classification model by using the characteristic values and the label data of the training group batch data;
s109, inputting characteristic values of test batch data, calculating to obtain a cutter state through a classification model, comparing the cutter state with label data, and calculating to obtain accuracy of accurate identification;
s110, judging the precision of the classification model;
the method comprises the steps of presetting precision, judging whether a classification model precision calculation value meets the preset precision, and if the classification model precision is greater than the preset precision, entering S111; if the precision of the classification model is smaller than the preset precision, returning to the S107 signal characteristic value selection stage, adjusting the selected signal characteristic value, reconstructing the classification model and calculating the precision until the preset precision is met;
s111, determining a signal characteristic value selection rule and storing classification model training data;
s2, monitoring through a real-time data driving classification model, specifically:
s201, when the process programming is carried out, setting web processing content in one process step, and adding an identification code in a digital control program of the process step;
s202, starting a cutter monitoring system when a program runs, and starting a web processing cutter state monitoring system after identifying a web processing step identification code in the part processing process;
s203, monitoring and dynamically updating the classification model;
the method specifically comprises the steps of starting a classification model updating module when a new cutter is used, acquiring a first n-layer signal cut by the new cutter, performing equal proportion processing on classification model training data by using a real-time vibration signal, and reconstructing a classification model based on the updated data;
the method specifically comprises the steps of starting a classification model updating module when the new cutter is identified to be used, acquiring the front n-layer signals of web processing by the new cutter, and extracting each layer of signals by using the cutter rail characteristics in the classification model training stage to judge and denoise;
the use of real-time vibration signals to perform equal proportion processing on the training data of the classification model specifically refers to the use of real-time data to calculate and obtain the average value M of the root mean square of each turn of web processing signals of each layer 2 Using the label of the training stage of the classification model as the data of the intact state of the cutter, calculating and obtaining the average value M of the root mean square of each rotation of the web processing signals of each layer 1 Calculating to obtain signal proportion m=m 2 /M 1 Reading the root mean square Data1 of each rotation saved in the model training stage, and calculating the signal proportion to obtain updated root mean square Data 2=Dat1;
the reconstruction of the classification model based on the updated Data specifically refers to reconstructing a cutter state monitoring model by using new signal characteristic values by using the updated root mean square Data2 per revolution to calculate and obtain corresponding signal characteristic values according to a signal characteristic value selection rule determined in a classification model training stage;
s204, starting a web processing cutter for real-time monitoring;
the method specifically refers to judging the state of a cutter by using a monitoring model when the next layer of web is processed, taking the processing process data of one layer of web as a target, and outputting a judging result to be that the cutter is intact or damaged;
s205, obtaining a vibration signal of a cutter for processing a layer of web;
s206, executing tool path feature discrimination and denoising;
s207, monitoring signal segmentation, namely segmenting to obtain batch data by adopting the number N of root mean square values per rotation contained in each batch of data which is the same as that in the training stage of the classification model;
s208, calling model identification;
s209, calculating a signal duty ratio A of the identified tipping section;
s210, judging whether the signal duty ratio A of the tipping section is larger than a monitoring threshold value or not;
s211, monitoring and alarming, and controlling the machine tool.
2. The method for monitoring the failure of a machining tool for a web of an aircraft structural member according to claim 1, wherein the method comprises the following steps: in step S206, the tool path feature determination and denoising, specifically, separating the tool paths processed by each layer of web according to G0 and G1 codes, performing the tool path feature determination and denoising by using each layer of web processing tool path as a data object, and calculating to obtain the minimum Z coordinate value Z in the processing process of one layer of web min Setting Z value deviation DeltaZ, and processing each layer of web plate to meet Z min -△Z＜Z＜Z min The tool path with the angle delta Z and the corresponding processing signals are intercepted and used as the data object for monitoring the web processing tool state.
3. The method for monitoring the failure of a machining tool for a web of an aircraft structural member according to claim 1, wherein the method comprises the following steps: in step S207, the monitoring signal segmentation specifically refers to setting the number of root mean square values per turn included in each batch of data as N, and if the number of root mean square values per turn included in each web processing monitoring data object is M, dividing each batch of monitoring data into K batches, where K is a value obtained by rounding M/N.
4. The method for monitoring the failure of a machining tool for a web of an aircraft structural member according to claim 1, wherein the method comprises the following steps: in step S104, the monitoring signal segmentation specifically refers to performing signal characteristic value calculation on each batch of data obtained, including deviation, kurtosis, standard deviation, mean value and maximum time-frequency domain signal characteristic value of root mean square data of each revolution.
5. The method for monitoring the failure of a machining tool for a web of an aircraft structural member according to claim 1, wherein the method comprises the following steps: in step S111, determining a rule for selecting signal feature values and storing training data of the classification model specifically refers to recording the signal feature values selected by the constructed classification model, storing training set data used in training the model, including signals corresponding to the states of intact cutters and failure cutters, using the same rule for selecting signal feature values when the features of the classification model are updated in the monitoring stage in real time, and reconstructing the classification model after processing the data stored in the rule.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310953555.7A CN116690313B (en) | 2023-08-01 | 2023-08-01 | Failure monitoring method for machining cutter of web plate of aircraft structural member |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310953555.7A CN116690313B (en) | 2023-08-01 | 2023-08-01 | Failure monitoring method for machining cutter of web plate of aircraft structural member |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116690313A CN116690313A (en) | 2023-09-05 |
| CN116690313B true CN116690313B (en) | 2023-12-12 |
Family
ID=87839518
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310953555.7A Active CN116690313B (en) | 2023-08-01 | 2023-08-01 | Failure monitoring method for machining cutter of web plate of aircraft structural member |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116690313B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117170306B (en) * | 2023-10-31 | 2024-03-15 | 成都飞机工业（集团）有限责任公司 | Cutter state monitoring method, cutter state monitoring system, electronic equipment and storage medium |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104723171A (en) * | 2015-03-17 | 2015-06-24 | 洛阳理工学院 | Cutter wear monitoring method based on current and acoustic emission compound signals |
| CN107584334A (en) * | 2017-08-25 | 2018-01-16 | 南京航空航天大学 | A kind of complex structural member numerical control machining cutter status real time monitor method based on deep learning |
| JP2018120406A (en) * | 2017-01-25 | 2018-08-02 | Ntn株式会社 | State monitoring method and state monitoring apparatus |
| CN110456731A (en) * | 2019-08-13 | 2019-11-15 | 成都飞机工业（集团）有限责任公司 | A kind of numerical control cutting geometry knife rail feature extracting method |
| CN110561193A (en) * | 2019-09-18 | 2019-12-13 | 杭州友机技术有限公司 | Cutter wear assessment and monitoring method and system based on feature fusion |
| CN111007801A (en) * | 2019-12-27 | 2020-04-14 | 中国一拖集团有限公司 | Cutter life real-time acquisition method based on multi-dimensional attribute state judgment |
| CN111782624A (en) * | 2020-06-16 | 2020-10-16 | 中译语通科技（青岛）有限公司 | Machine tool cutter vibration monitoring and analyzing method based on artificial intelligence and big data |
| CN112008495A (en) * | 2020-07-28 | 2020-12-01 | 成都飞机工业（集团）有限责任公司 | Cutter damage identification method based on vibration monitoring |
| CN112372372A (en) * | 2020-12-07 | 2021-02-19 | 哈尔滨理工大学 | Efficient milling cutter accumulated friction wear boundary identification and verification method |
| CN112720071A (en) * | 2021-01-27 | 2021-04-30 | 赛腾机电科技（常州）有限公司 | Cutter real-time state monitoring index construction method based on intelligent fusion of multi-energy domain signals |
| CN113378725A (en) * | 2021-06-15 | 2021-09-10 | 山东大学 | Cutter fault diagnosis method, equipment and storage medium based on multi-scale-channel attention network |
| CN113798920A (en) * | 2021-09-23 | 2021-12-17 | 大连理工大学 | Cutter wear state monitoring method based on variational automatic encoder and extreme learning machine |
| CN114227382A (en) * | 2022-01-18 | 2022-03-25 | 湖北汽车工业学院 | Cutter damage monitoring system and method based on novel capsule network |
| CN114850969A (en) * | 2022-07-08 | 2022-08-05 | 成都飞机工业（集团）有限责任公司 | Cutter failure monitoring method based on vibration signals |
| CN114986258A (en) * | 2022-07-22 | 2022-09-02 | 中机智能装备创新研究院(宁波)有限公司 | Cutter state monitoring method and device, computer equipment and system |
| CN115356985A (en) * | 2022-08-18 | 2022-11-18 | 成都飞机工业（集团）有限责任公司 | Machining process self-adaptive optimization method based on power signals |
| CN115922442A (en) * | 2022-12-24 | 2023-04-07 | 西安交通大学 | Cutter grinding damage real-time monitoring method based on spindle vibration signal and related device |
| CN116061006A (en) * | 2023-04-03 | 2023-05-05 | 成都飞机工业（集团）有限责任公司 | Cutter monitoring method, device, equipment and medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210197335A1 (en) * | 2019-12-26 | 2021-07-01 | Dalian University Of Technology | Data Augmentation Method Based On Generative Adversarial Networks In Tool Condition Monitoring |
-
2023
- 2023-08-01 CN CN202310953555.7A patent/CN116690313B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104723171A (en) * | 2015-03-17 | 2015-06-24 | 洛阳理工学院 | Cutter wear monitoring method based on current and acoustic emission compound signals |
| JP2018120406A (en) * | 2017-01-25 | 2018-08-02 | Ntn株式会社 | State monitoring method and state monitoring apparatus |
| CN107584334A (en) * | 2017-08-25 | 2018-01-16 | 南京航空航天大学 | A kind of complex structural member numerical control machining cutter status real time monitor method based on deep learning |
| CN110456731A (en) * | 2019-08-13 | 2019-11-15 | 成都飞机工业（集团）有限责任公司 | A kind of numerical control cutting geometry knife rail feature extracting method |
| CN110561193A (en) * | 2019-09-18 | 2019-12-13 | 杭州友机技术有限公司 | Cutter wear assessment and monitoring method and system based on feature fusion |
| CN111007801A (en) * | 2019-12-27 | 2020-04-14 | 中国一拖集团有限公司 | Cutter life real-time acquisition method based on multi-dimensional attribute state judgment |
| CN111782624A (en) * | 2020-06-16 | 2020-10-16 | 中译语通科技（青岛）有限公司 | Machine tool cutter vibration monitoring and analyzing method based on artificial intelligence and big data |
| CN112008495A (en) * | 2020-07-28 | 2020-12-01 | 成都飞机工业（集团）有限责任公司 | Cutter damage identification method based on vibration monitoring |
| CN112372372A (en) * | 2020-12-07 | 2021-02-19 | 哈尔滨理工大学 | Efficient milling cutter accumulated friction wear boundary identification and verification method |
| CN112720071A (en) * | 2021-01-27 | 2021-04-30 | 赛腾机电科技（常州）有限公司 | Cutter real-time state monitoring index construction method based on intelligent fusion of multi-energy domain signals |
| CN113378725A (en) * | 2021-06-15 | 2021-09-10 | 山东大学 | Cutter fault diagnosis method, equipment and storage medium based on multi-scale-channel attention network |
| CN113798920A (en) * | 2021-09-23 | 2021-12-17 | 大连理工大学 | Cutter wear state monitoring method based on variational automatic encoder and extreme learning machine |
| CN114227382A (en) * | 2022-01-18 | 2022-03-25 | 湖北汽车工业学院 | Cutter damage monitoring system and method based on novel capsule network |
| CN114850969A (en) * | 2022-07-08 | 2022-08-05 | 成都飞机工业（集团）有限责任公司 | Cutter failure monitoring method based on vibration signals |
| CN114986258A (en) * | 2022-07-22 | 2022-09-02 | 中机智能装备创新研究院(宁波)有限公司 | Cutter state monitoring method and device, computer equipment and system |
| CN115356985A (en) * | 2022-08-18 | 2022-11-18 | 成都飞机工业（集团）有限责任公司 | Machining process self-adaptive optimization method based on power signals |
| CN115922442A (en) * | 2022-12-24 | 2023-04-07 | 西安交通大学 | Cutter grinding damage real-time monitoring method based on spindle vibration signal and related device |
| CN116061006A (en) * | 2023-04-03 | 2023-05-05 | 成都飞机工业（集团）有限责任公司 | Cutter monitoring method, device, equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116690313A (en) | 2023-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109909804B (en) | Tool wear damage online monitoring method based on spindle driving current and process steps | |
| CN110587377B (en) | Method for monitoring defects of milling cutter on line | |
| CN110908334B (en) | Cutter wear monitoring method | |
| CN108490880B (en) | Method for monitoring wear state of cutting tool of numerical control machine tool in real time | |
| CN110059442B (en) | Turning tool changing method based on part surface roughness and power information | |
| CN109507952B (en) | Method for monitoring abnormal state of numerical control machining of complex part based on cutting load | |
| CN116690313B (en) | Failure monitoring method for machining cutter of web plate of aircraft structural member | |
| CN101477351B (en) | Intelligent numerical control method with three-stage process self-optimization function | |
| CN111113150B (en) | Method for monitoring state of machine tool cutter | |
| CN114850969B (en) | Cutter failure monitoring method based on vibration signals | |
| CN111660141B (en) | Milling cutter wear state identification method based on spindle driving current and irrelevant to working conditions | |
| CN104781740A (en) | Modular system for real-time evaluation and monitoring of machining production-line overall performances calculated from each given workpiece, tool and machine | |
| CN113146359B (en) | Numerical control machine tool control system for intelligently monitoring and diagnosing cutting state | |
| Jemielniak | Contemporary challenges in tool condition monitoring | |
| CN108527005A (en) | A kind of CNC cutting tool states detection method and system | |
| CN114273977A (en) | MES-based cutter wear detection method and system | |
| CN116061006B (en) | Cutter monitoring method, device, equipment and medium | |
| CN114326593B (en) | Cutter life prediction system and method | |
| CN115922442A (en) | Cutter grinding damage real-time monitoring method based on spindle vibration signal and related device | |
| CN112381388B (en) | Method for monitoring operation of spindle motor of numerical control machine tool | |
| CN112008495B (en) | Cutter damage identification method based on vibration monitoring | |
| US11226613B2 (en) | Anomaly detection device, anomaly detection server and anomaly detection method | |
| CN112548678A (en) | Cutter detection method and device of numerical control machining center | |
| Park et al. | A study on the predictive maintenance of 5 axis CNC machine tools for cutting of large aircraft parts | |
| CN118081479B (en) | Machine tool operation fault online monitoring method and 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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Jiang Zhenxi Inventor after: Mou Wenping Inventor after: Zhu Shaowei Inventor after: Liu Kuan Inventor after: Mao Yiyan Inventor after: Zhao Zhonggang Inventor after: Yu Zhiyong Inventor after: Zhang Shaokun Inventor after: Li Weidong Inventor before: Jiang Zhenxi Inventor before: Mou Wenping Inventor before: Zhu Shaowei Inventor before: Liu Kuan Inventor before: Mao Yiyan Inventor before: Zhao Zhonggang Inventor before: Yu Zhiyong Inventor before: Yin Shaokun Inventor before: Li Weidong |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |