CN109634254A - A kind of on-line fault diagnosis method and device of intelligence manufacture equipment - Google Patents

Abstract

The invention relates to the technical field of fault diagnosis, in particular to an online fault diagnosis method and device for intelligent manufacturing equipment. By collecting historical data of numerically controlled equipment, an inspection threshold value for fault diagnosis is obtained, and a fault mode database under different fault modes is established. ; By collecting the running data of the numerical control equipment in real time, and making fault judgment and fault mode diagnosis on the running data, the ability of diagnosing the online faults of the numerical control equipment is improved.

Description

Online fault diagnosis method and device for intelligent manufacturing equipment

technical field

The invention relates to the technical field of fault diagnosis, in particular to an on-line fault diagnosis method and device for intelligent manufacturing equipment.

Background technique

Fault diagnosis methods can usually be divided into three categories: model-driven fault diagnosis, knowledge judgment-based fault diagnosis, and data-driven fault diagnosis. Compared with model-driven and knowledge-based fault diagnosis methods, data-driven fault diagnosis The fault diagnosis method is to directly use the data or status of the monitoring system, and use data analysis, statistical decision-making, deep learning and other technologies for fault diagnosis. The characteristics and fault attributes can be effectively analyzed, so as to realize the identification of faults.

The fault problem-driven system modeling and diagnosis itself is a process of benign interaction. When a complex manufacturing system fails, it is necessary to diagnose the fault in time to ensure the operation safety of the complex manufacturing system, and how to timely confirm whether a fault has occurred and where it occurred. The failure, what level or magnitude of failure occurred, in order to gain time in the process of safety management, and timely implement maintenance or troubleshooting, become the key issues worth solving.

SUMMARY OF THE INVENTION

The present invention provides an on-line fault diagnosis method and device for intelligent manufacturing equipment, which improves the ability of diagnosing on-line faults of numerical control equipment.

An online fault diagnosis method for intelligent manufacturing equipment provided by the present invention is characterized in that it includes the following steps:

Step S1, collecting the historical data of the numerical control equipment, and obtaining the inspection threshold value used for fault diagnosis;

Step S2, establishing a failure mode database;

Step S3, collecting the running data of the numerical control equipment in real time;

Step S4: Perform fault judgment and fault mode diagnosis on the operating data.

Further, the step S1 specifically includes the following steps:

Step S11, collecting historical data of numerical control equipment, and extracting stationary residuals for fault diagnosis;

Let historical data Y=(y ₁ , y ₂ ,...y _m ), where n _y is the dimension of the output data, m is the sample size, and Y is decomposed into a non-stationary trend term and a stationary residual term,

in, is a non-stationary trend term, is the stationary residual term;

Step S12, record and respectively and The i-th column of , computes the covariance matrix of the stationary residual term,

in, for The transposed matrix of ;

Step S13, calculate the detection residual of the historical data y

in, is the transpose matrix of y;

Step S14, constructing a statistic T ² ,

in, for The transposed matrix of ;

Let the detection significance level be α, then the corresponding test threshold is:

Among them, F _α ( _ny , mn _y ) is the historical data with a significance level of α.

Further, the step S2 specifically includes:

Let f be the failure mode of historical data y, f ₀ represents the mode in which the operating data is normal, and F={f _i , i=1, 2, . . . , n _f } represents the historical failure with n _f types of different failure modes Pattern library, f _i represents the i-th type of historical fault; establish a one-to-one mapping relationship between the collected historical data and the detected fault modes, so as to establish a known fault mode library.

Further, the operation data of the numerical control device in the step S1 includes: power y ₁ , temperature y ₂ , humidity y ₃ , pressure y ₄ , displacement y ₅ , and rotational speed y ₆ of the numerical control device.

Further, the specific method for extracting the stationary residual for fault diagnosis in the step S11 is: using the steady-state determination method to obtain the steady-state factor of the historical sample data corresponding to the final diagnosis model parameters in each time period, and setting the steady-state tolerance. , remove the data whose steady-state factor is less than the steady-state tolerance, and obtain the steady-state residuals of each time period.

Further, the step S4 specifically includes:

Step S41, perform fault detection on the running data, if Then f∈{f ₀ }, the running data y is normal data; otherwise, The running data y is abnormal data;

Step S42, calculating the degree of deviation between the abnormal data and all the failure modes f _i , and discriminating the failure mode with the smallest deviation degree as the failure mode of the abnormal data.

Further, the step S42 specifically includes:

Denote r _ij (j=1,2,...,n _i ) as the sample direction corresponding to the running fault, where,

Then the relationship between the running fault characteristic direction r _ij and the sample direction is:

Wherein, R _i is the characteristic direction corresponding to the failure mode f _i , i=1, 2,...,n _j ;

Denote θ(r, T _i ) as the included angle between the operating fault direction r _ij and the characteristic direction R _i corresponding to the i-th fault mode f _i in the known fault mode library, then

θ(r,R _i )=arccos[|r ^T R _i |/||r||·||R _i ||]

If the included angle between the running fault direction r _ij and the characteristic direction R _i corresponding to the ith fault mode is the smallest, the current fault is judged as the ith fault mode.

An online fault diagnosis device for intelligent manufacturing equipment, the device includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, triggering The apparatus performs any of the methods described above.

The beneficial effects of the present invention are as follows: the present invention discloses an on-line fault diagnosis method and device for intelligent manufacturing equipment. By collecting historical data of numerical control equipment, a test threshold for fault diagnosis is obtained, and a fault mode database under different fault modes is established. ; By collecting the running data of the numerical control equipment in real time, and making fault judgment and fault mode diagnosis on the running data, the ability of diagnosing the online faults of the numerical control equipment is improved.

Description of drawings

The present invention will be further described below in conjunction with the accompanying drawings and examples.

FIG. 1 is a schematic flowchart of an online fault diagnosis method for an intelligent manufacturing device according to an embodiment of the present invention.

Detailed ways

Referring to FIG. 1, an online fault diagnosis method for intelligent manufacturing equipment provided by the present invention is characterized in that, it includes the following steps:

Step S1, collecting the historical data of the numerical control equipment, and obtaining the inspection threshold value used for fault diagnosis;

Step S2, establishing a failure mode database;

Step S3, collecting the running data of the numerical control equipment in real time;

Step S4: Perform fault judgment and fault mode diagnosis on the operating data.

Further, the step S1 specifically includes the following steps:

Step S11, collecting historical data of numerical control equipment, and extracting stationary residuals for fault diagnosis;

Let historical data Y=(y ₁ , y ₂ ,...y _m ), where n _y is the dimension of the output data, m is the sample size, and Y is decomposed into a non-stationary trend term and a stationary residual term,

in, is a non-stationary trend term, is the stationary residual term;

Step S12, record and respectively and The i-th column of , computes the covariance matrix of the stationary residual term,

in, for The transposed matrix of ;

Step S13, calculate the detection residual of the historical data y

in, is the transpose matrix of y;

Step S14, constructing a statistic T ² ,

in, for The transposed matrix of ;

The running data obeys the multivariate normal distribution, and its detection significance level is α, then the corresponding test threshold is:

Among them, F _α ( _ny , mn _y ) is the historical data with a significance level of α.

Further, the step S2 specifically includes:

Let f be the failure mode of historical data y, f ₀ represents the mode in which the operating data is normal, and F={f _i , i=1, 2, . . . , n _f } represents the historical failure with n _f types of different failure modes Pattern library, f _i represents the i-th type of historical fault; establish a one-to-one mapping relationship between the collected historical data and the detected fault modes, so as to establish a known fault mode library.

Further, the operation data of the numerical control device in the step S1 includes: power y ₁ , temperature y ₂ , humidity y ₃ , pressure y ₄ , displacement y ₅ , and rotational speed y ₆ of the numerical control device.

Further, the specific method for extracting the stationary residual for fault diagnosis in the step S11 is: using the steady-state determination method to obtain the steady-state factor of the historical sample data corresponding to the final diagnosis model parameters in each time period, and setting the steady-state tolerance. , remove the data whose steady-state factor is less than the steady-state tolerance, and obtain the steady-state residuals of each time period.

Further, the step S4 specifically includes:

Step S41, perform fault detection on the running data, if Then f∈{f ₀ }, the running data y is normal data; otherwise, The running data y is abnormal data;

Step S42, calculating the degree of deviation between the abnormal data and all the failure modes f _i , and discriminating the failure mode with the smallest deviation degree as the failure mode of the abnormal data.

After the fault detection, the sample direction r _ij of the current fault is obtained, and the known fault identification is to use the position distribution or direction distribution of the detection residual to extract the fault feature, which is used to construct the deviation degree of the fault identification. The direction information is used as the amount of information to construct the deviation degree of fault identification.

Further, the step S42 specifically includes:

Denote r _ij (j=1,2,...,n _i ) as the sample direction corresponding to the running fault, where,

Then the relationship between the running fault characteristic direction r _ij and the sample direction is:

Wherein, R _i is the characteristic direction corresponding to the failure mode f _i , i=1, 2,...,n _j ;

Denote θ(r, T _i ) as the included angle between the operating fault direction r _ij and the characteristic direction R _i corresponding to the i-th fault mode f _i in the known fault mode library, then

θ(r,R _i )=arccos[|r ^T R _i |/||r||·||R _i ||]

If the included angle between the running fault direction r _ij and the characteristic direction R _i corresponding to the ith fault mode is the smallest, the current fault is judged as the ith fault mode.

The present invention provides an online fault diagnosis method and device for intelligent manufacturing equipment. The device includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the When the processor executes, the device is triggered to execute any one of the methods described above.

The above descriptions are only preferred embodiments of the present invention, and the present invention is not limited to the above-mentioned embodiments, as long as the technical effects of the present invention are achieved by the same means, they should all belong to the protection scope of the present invention.

Claims (8)

1. an on-line fault diagnosis method of intelligent manufacturing equipment, is characterized in that, comprises the following steps: Step S1, collecting the historical data of the numerical control equipment, and obtaining the inspection threshold value used for fault diagnosis; Step S2, establishing a failure mode database; Step S3, collecting the running data of the numerical control equipment in real time; Step S4: Perform fault judgment and fault mode diagnosis on the operating data. 2. The online fault diagnosis method of an intelligent manufacturing device according to claim 1, wherein the step S1 specifically comprises the following steps: Step S11, collecting historical data of numerical control equipment, and extracting stationary residuals for fault diagnosis; Let historical data Y=(y ₁ , y ₂ ,...y _m ), where n _y is the dimension of the output data, m is the sample size, and Y is decomposed into a non-stationary trend term and a stationary residual term, in, is a non-stationary trend term, is the stationary residual term; Step S12, record and respectively and The i-th column of , computes the covariance matrix of the stationary residual term, in, for The transposed matrix of ; Step S13, calculate the detection residual of the historical data y in, is the transpose matrix of y; Step S14, constructing a statistic T ² , in, for The transposed matrix of ; Let the detection significance level be α, then the corresponding test threshold is: Among them, F _α ( _ny , mn _y ) is the historical data with a significance level of α. 3 . The online fault diagnosis method for intelligent manufacturing equipment according to claim 2 , wherein the operation data of the numerical control equipment in the step S1 comprises: the power y ₁ , the temperature y ₂ , the humidity y of the numerical control equipment running. 4 . _3. Pressure y ₄ , displacement y ₅ , rotational speed y ₆ . 4. The online fault diagnosis method of an intelligent manufacturing equipment according to claim 2, wherein the specific method for extracting stationary residuals for fault diagnosis in the step S11 is: using a steady state determination method to obtain a final diagnosis The steady state factor of the historical sample data corresponding to the model parameters in each time period is set, the steady state tolerance is set, the data with the steady state factor smaller than the steady state tolerance is eliminated, and the steady residual error of each time period is obtained. 5. The online fault diagnosis method for an intelligent manufacturing device according to claim 1, wherein the step S2 specifically comprises: Let f be the failure mode of historical data y, where f ₀ represents the normal mode of operation data, F={f _i , i=1, 2, ..., n _f } represents the failure mode with n _f types of different failure modes Historical failure mode database, f _i represents the i-th historical failure; establish a one-to-one mapping relationship between the collected historical data and the detected failure modes, thereby establishing a failure mode database. 6. The online fault diagnosis method for an intelligent manufacturing device according to claim 1, wherein the step S4 specifically comprises: Step S41, perform fault detection on the running data, if Then f∈{f ₀ }, the running data y is normal data; otherwise, The running data y is abnormal data; Step S42, calculating the degree of deviation between the abnormal data and all the failure modes f _i , and discriminating the failure mode with the smallest deviation degree as the failure mode of the abnormal data. 7. The online fault diagnosis method for an intelligent manufacturing device according to claim 6, wherein the step S42 specifically comprises: Denote r _ij (j=1,2,...,n _i ) as the sample direction corresponding to the running fault, where, Then the relationship between the running fault characteristic direction r _ij and the sample direction is: Wherein, R _i is the characteristic direction corresponding to the failure mode f _i , i=1, 2,...,n _j ; Denote θ(r, T _i ) as the included angle between the operating fault direction r _ij and the characteristic direction R _i corresponding to the i-th fault mode f _i in the known fault mode library, then θ(r,R _i )=arccos[|r ^T R _i |/||r||·||R _i ||] If the included angle between the running fault direction r _ij and the characteristic direction R _i corresponding to the ith fault mode is the smallest, the current fault is judged as the ith fault mode. 8. An online fault diagnosis device for intelligent manufacturing equipment, characterized in that the device comprises a memory for storing computer program instructions and a processor for executing program instructions, wherein when the computer program instructions are executed by the When the processor executes, the apparatus is triggered to execute the method according to any one of claims 1 to 7 .