CN115964619A - Method for removing cutting force signal noise, electronic equipment and storage medium - Google Patents
Method for removing cutting force signal noise, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115964619A CN115964619A CN202310114634.9A CN202310114634A CN115964619A CN 115964619 A CN115964619 A CN 115964619A CN 202310114634 A CN202310114634 A CN 202310114634A CN 115964619 A CN115964619 A CN 115964619A
- Authority
- CN
- China
- Prior art keywords
- target
- list
- cutting force
- force signal
- key
- 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.)
- Granted
Links
Images
Landscapes
- Numerical Control (AREA)
Abstract
The invention provides a method for removing cutting force signal noise, which comprises the steps of obtaining a cutting force signal vector list, obtaining target vector matrix lists with different embedded dimensions according to the cutting force signal vector, carrying out singular value decomposition on each target vector matrix, obtaining a target information quantity list, obtaining a target change rate list according to the target information quantity list, obtaining a key change rate information list according to the target change rate list, taking the target vector matrix corresponding to the specified embedded dimension as a key vector matrix, carrying out singular value decomposition on the key vector matrix, obtaining the key information quantity list, obtaining the maximum information quantity order, reconstructing A through an anti-diagonal method after setting the key information quantity of the maximum order in the key information quantity list to be 0, and obtaining the cutting force signal list after denoising. The recognition accuracy of the cutting tool state is improved.
Description
Technical Field
The present invention relates to the field of cutting force signal denoising technology, and in particular, to a method for removing cutting force signal noise, an electronic device, and a storage medium.
Background
In the prior art, cutting machining is actually a process of cutting a workpiece by mutual extrusion of a cutter and the workpiece. The cutting force signal is a key factor for measuring the machining state of the machine, and not only directly influences the machining precision of the workpiece, but also directly influences the production efficiency and the production cost. However, in ultra-precision turning, the cutting force is extremely weak and is only mN level, so that the acquired cutting force signal also belongs to a weak signal, the cutting force signal is easily affected by electromagnetic interference and sensor noise, a user cannot acquire an accurate cutting state of the cutting tool, misjudgment is generated, the working efficiency of the cutting tool is reduced due to the fact that the user changes parameters of the cutting tool according to the wrong cutting state of the cutting tool, the problem that the user cannot find the cutting tool in time and solve the problem of the cutting tool may be caused, and monitoring of the state of the cutting tool and evaluation of processing quality are seriously affected.
Disclosure of Invention
Aiming at the technical problem, the technical scheme adopted by the invention is as follows:
a method of removing cutting force signal noise, comprising the steps of:
s100, obtaining cutting force signal information list A = (A) 1 ，A 2 ，……，A i ，……，A m ) I =1,2, … …, m, m is the number of cutting force signal information; wherein A is i Cutting force signal information is the ith cutting force signal information, and the cutting force signal information at least comprises: acquiring the time of the corresponding cutting force signal and the corresponding signal intensity of the cutting force signal; the signal intensity is used for representing the magnitude of the cutting force;
s200, vectorizing A to obtain a cutting force signal vector list B = (B) 1 ，B 2 ，……，B i ，……，B m ) (ii) a Wherein, B i Is a pair A i Carrying out vectorization processing to obtain a cutting force signal vector;
s300, according to B, obtaining a target vector matrix list BZ = (BZ) 1 ，BZ 2 ，……，BZ j ，……，BZ n ) J =1,2, … …, n, n is the number of the target vector matrix obtained according to m; wherein, BZ j Embedding a target vector matrix with dimension j +1, wherein the target vector matrix is a token matrix, BZ j The following conditions are met:
wherein p satisfies the following condition p = m-j;
s400, singular value decomposition is carried out on each target vector matrix in the BZ, and a target information quantity list P = (P) is obtained 1 ，P 2 ，……，P j ，……，P n ) (ii) a Wherein, P j Is BZ j The corresponding target information quantity is a second information quantity of a diagonal line in a singular value matrix obtained after singular value decomposition is carried out on a target vector matrix;
s500, acquiring a target change rate list H = (H) according to P 1 ，H 2 ，……，H g ，……，H n-1 ) G =1,2, … …, n-1; wherein H g Is P j+1 Corresponding target rate of change, H g The following conditions are met:
H g ＝P g+1 -P g ；
s600, acquiring a key change rate information list Q = (Q) according to H 1 ，Q 2 ，……，Q k ，……，Q n-3 ) K =1,2, … …, n-3; wherein the kth key rate of change information Q k ＝(S k ，v 1 k ，v 2 k ，q k )，S k For the Key embedding dimension, v 1 k Is the first key rate of change, v 2 k Is the second key rate of change, q k Is S k Corresponding key change value, q k The following conditions are met:
S700, embedding dimension S to be specified 0 The corresponding target vector matrix is used as a key vector matrix XZ; wherein S is 0 Embedding the corresponding key dimension when the second key change value in the key change rate information list is 0;
s800, singular value decomposition is carried out on the XZ, and a key information amount list D = (D) is obtained 1 ，D 2 ，……，D f ，……，D F ) F =1,2, … …, F is the amount of key information amount corresponding to XZ; wherein D is f F, carrying out singular value decomposition on the XZ to obtain the f-th key information quantity;
s900, acquiring the maximum information quantity order t; wherein t meets the following condition:
t＝arg f min(d f ) And argmin () is a variable value function corresponding to the variable at the minimum value of the function, wherein d f The following conditions are met:
s1000, adding D into D t And after each key information quantity is set to be 0, reconstructing A through an anti-diagonal method to obtain a denoised cutting force signal information list.
The invention has at least the following beneficial effects:
the method comprises the steps of obtaining a cutting force signal vector list, obtaining target vector matrix lists of different embedding dimensions according to cutting force signal vectors, performing singular value decomposition on each target vector matrix to obtain a target information quantity list, obtaining a target change rate list according to the target information quantity list, obtaining a key change rate information list according to the target change rate list, taking the target vector matrix corresponding to the specified embedding dimension as a key vector matrix, performing singular value decomposition on the key vector matrix to obtain a key information quantity list, obtaining the maximum information quantity order, reconstructing A through a deprecation angle method after setting the key information quantity of the maximum order in the key information quantity list to be 0, and obtaining the cutting force signal information list after denoising. The noise in the cutting force signal is removed through the processing, the interference of the noise to the cutting force signal is avoided, the user cannot acquire the accurate cutting state of the cutting tool, the error judgment is generated, the working efficiency of the cutting tool is reduced due to the fact that the user changes the parameters of the cutting tool according to the wrong cutting state of the cutting tool, the user may not find the cutting tool in time and solve the problem of the cutting tool, the cutting force signal after the noise is removed can more accurately show the cutting state of the cutting tool, the user can further optimize the parameters of the cutting tool in time according to the cutting state of the cutting tool, and therefore the working efficiency of the cutting tool is improved, and the identification precision of the cutting tool state is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a method for removing signal noise according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
The invention provides a method for removing cutting force signal noise, which is characterized by comprising the following steps:
s100, obtaining cutting force signal information list A = (A) 1 ，A 2 ，……，A i ，……，A m ) I =1,2, … …, m, m is the number of cutting force signal information; wherein, A i Cutting force signal information is the ith cutting force signal information, and the cutting force signal information at least comprises: acquiring the time of the corresponding cutting force signal and the corresponding signal intensity of the cutting force signal; the signal strength is used to indicate the magnitude of the cutting force.
Specifically, in ultra-precision turning, the cutting force is extremely weak and is only mN level, and the obtained cutting force signal is a weak signal, so that the cutting force signal is easily affected by electromagnetic interference and sensor noise, and the monitoring of the tool state and the evaluation of the machining quality are difficult, so that the cutting force signal with higher detection precision needs to be obtained through denoising processing.
In the embodiment of the present invention, a may be obtained through the following steps:
s101, acquiring target sampling frequency f 0 And a sampling duration t 0 。
In particular, f 0 Can be obtained by the following steps:
s1011, obtaining the center frequency f corresponding to the current cutting signal of the cutting tool c ；
S1012, when the current cutting signal indicates that the cutting tool is in the working state, executing S1013; when the current cutting signal indicates that the cutting tool is in a non-working state, executing S1014;
specifically, the working state is a state in which the cutter is cutting the workpiece, and the non-working state is a state in which the cutter is in the process of moving but is not cutting the workpiece.
And further, judging the change of the cutting state of the cutter through a preset sensor, and transmitting the change information of the cutting state to a server corresponding to the cutter after judging that the cutting state of the cutter is changed.
S1013, according to f c Obtaining f 0 ＝6f c ；
S1014, according to f c Obtaining f 0 ＝3f c 。
Determining a target sampling frequency through S1011 and S1013, and when the working state of the cutting tool is determined according to the current cutting force signal, setting the target sampling frequency to be 6 times of the center frequency when the working state of the cutting tool is determined; when the cutting tool is judged to be in a non-working state, setting the target sampling frequency to be 3 times of the central frequency; the target sampling frequency is changed according to the working state of the cutting tool, so that when the cutting tool is in the working state, the cutting force signal density acquired in unit time is greater than that acquired when the cutting tool is in the non-working state; the condition that the working state of the cutter is judged according to the cutting force signal to have larger errors caused by excessive useless cutting force signals is avoided.
S102, according to f 0 And t 0 Acquiring a target sampling number Num;
specifically, num satisfies the following condition:
Num＝f 0 *t 0 。
further, t 0 =1s, and inaccurate cutting state of the cutting tool in subsequent judgment caused by too few collected cutting force signals when the sampling time set by a user is too short is avoided; the problem that the workload of a processor is too large when the user sets the sampling time to be too large and the collected cutting force signal is too large to cause the subsequent processing of the cutting force signal is solved, so that the processing speed is reduced, the processing efficiency is too low, and the cutting state of the cutting tool is not timely judged.
In the embodiment of the invention, f is paired 0 And t 0 The Num is obtained by calculation, and the situation that t is arbitrarily set by a user is avoided 0 The quantity of cutting force signals obtained by sampling is insufficient, so that the subsequent calculation is inaccurate.
S103, using f 0 The cutting signal of the cutting tool is carried out for a time period t 0 To obtain a.
S200, vectorizing A to obtain a cutting force signal vector list B = (B) 1 ，B 2 ，……，B i ，……，B m ) (ii) a Wherein, B i Is to A i And carrying out vectorization processing to obtain a cutting force signal vector.
Specifically, since the cutting force signal in the present invention is a time-domain discrete signal, the cutting force signal can be naturally expressed as a cutting force signal vector corresponding to the cutting force signal based on the cutting force signal information.
In the embodiment of the invention, single Instruction Multiple Data is adopted to carry out vectorization processing on A to obtain B.
Furthermore, cutting force signal information is converted into a cutting force signal vector through vectorization processing, so that the expression of the cutting force signal is more uniform and simpler, and the acquisition of a target vector matrix is easier to express.
S300, according to B, obtaining a target vector matrix list BZ = (BZ) 1 ，BZ 2 ，……，BZ j ，……，BZ n ) J =1,2, … …, n, n is the number of the target vector matrix obtained according to m; wherein, BZ j Embedding a target vector matrix with dimension j +1, wherein the target vector matrix is a token matrix, BZ j The following conditions are met:
wherein p satisfies the following condition p = m-j.
Specifically, n ≦ m-1, it is understood that the embedding dimension of the target vector matrix is less than or equal to the number of cutting force signal vectors that make up the target vector matrix.
Preferably, the first and second electrodes are formed of a metal,
wherein it is present>
Rounding the upper symbol; it can be understood that the number of the target vector matrix is half of the number of the cutting force signal vectors; the method avoids the inaccurate processing result caused by too small quantity of the target vector matrixes; the excessive workload of the processor during the subsequent processing caused by the excessive quantity of the target vector matrixes is avoided, the processing speed is reduced, and the processing efficiency is improvedThe rate is too low, so that the cutting state of the cutting tool is not judged in time.
S400, singular value decomposition is carried out on each target vector matrix in the BZ, and a target information quantity list P = (P) is obtained 1 ，P 2 ，……，P j ，……，P n ) (ii) a Wherein, P j Is BZ j And the corresponding target information quantity is the second information quantity of the diagonal line in the singular value matrix obtained by performing singular value decomposition on the target vector matrix.
Specifically, those skilled in the art know that any method for performing singular value decomposition on the target vector matrix falls within the scope of the present invention, and details thereof are not repeated herein.
Further, after singular value decomposition is performed on the target vector matrixes, in an information quantity list corresponding to each target vector matrix, the information quantities are arranged in a descending order.
Further, a second information amount of a diagonal line in the singular value matrix obtained by performing singular value decomposition on each target vector matrix is obtained, which can be understood as obtaining an information amount of a second position of the information amount obtained by performing singular value decomposition on each target vector matrix as a target information amount; since the first two information contents in the information content list corresponding to each target vector matrix generally contain most of cutting force signal information, in order to highlight observation amplitude variation and detection accuracy, the information content with the information content arranged at the second position obtained by performing singular value decomposition on each target vector matrix is selected as the target information content.
S500, according to the P, obtaining a target change rate list H = (H) 1 ，H 2 ，……，H g ，……，H n-1 ) G =1,2, … …, n-1; wherein H g Is P j+1 Corresponding target rate of change, H g The following conditions are met:
H g ＝P g+1 -P g 。
s600, acquiring a key change rate information list Q = (Q) according to H 1 ，Q 2 ，……，Q k ，……，Q n-3 )，k =1,2, … …, n-3; wherein the kth key rate of change information Q k ＝(S k ，v 1 k ，v 2 k ，q k )，S k For the critical embedding dimension, v 1 k Is the first key rate of change, v 2 k Is the second key rate of change, q k Is S k Corresponding key change value, q k The following conditions are met:
Specifically, by obtaining the key change value, a target change rate curve with the ordinate being the target change rate and the abscissa being the embedding dimension corresponding to the target change rate can be obtained, the embedding dimension in the target change rate curve, which obviously decreases as the embedding dimension increases, can be obtained according to the target change rate curve, and the embedding dimension in the target change rate curve, which obviously decreases as the embedding dimension increases, is the preferred embedding dimension according to the embedding dimension characteristics k When =0, q k Corresponding S k Is the embedding dimension in the target rate of change curve that decreases significantly as the embedding dimension increases.
S700, embedding dimension S to be specified 0 The corresponding target vector matrix is used as a key vector matrix XZ; wherein S is 0 And embedding the corresponding key dimension when the second key change value in the key change rate information list is 0.
Specifically, a key embedding dimension corresponding to second change rate information with a second key change value of 0 in a second change rate information list is obtained; in order to ensure the accuracy of the reconstructed signal, the key embedding dimension corresponding to the second change rate information with the second key change value of 0 in the second change rate information list is selected.
S800, singular value decomposition is carried out on the XZ, and a key information content list D = (C) is obtainedD 1 ，D 2 ，……，D f ，……，D F ) F =1,2, … …, F is the amount of key information amount corresponding to XZ; wherein D is f The f-th key information quantity is obtained after singular value decomposition is carried out on the XZ.
In particular, D 1 ＞D 2 ＞……＞D f ＞……＞D F 。
S900, acquiring the maximum information quantity order t; wherein t meets the following condition:
t＝arg f min(d f ) And argmin () is a variable value-taking function of the corresponding variable when the function is at the minimum value, wherein d f The following conditions are met:
specifically, D1/F is a diagonal slope in a key information amount graph with the ordinate generated according to the key information amount as the key information amount and the abscissa as the order corresponding to the key information amount; d f and/F is the slope corresponding to the F-th key information content.
Further, d f The slope of the curve chart of the f-th key information quantity and the key information quantity is D 1 The numerical value corresponding to the key information quantity with the minimum vertical distance between diagonals of the/F is used as the maximum information quantity order; since the numerical value representing the key information quantity is larger before the key information quantity is in the higher order, only the key information quantity before t is reserved by taking the maximum information quantity order t as a boundary, and most information of the cutting force signal is reserved while the influence of noise on the cutting force signal caused by the key information quantity after the maximum information quantity order is avoided, so that the noise of the cutting force signal is removed, and the information of the cutting force signal is reserved as much as possible.
In the embodiment of the invention, the slope of the curve graph of the selection and the key information content is D 1 Vertical distance between diagonals of/F
S1000, adding D into D t And after each key information amount is set to be 0, reconstructing A by an anti-diagonal method, and acquiring a denoised cutting force signal information list.
The cutting force signal information list after denoising is obtained by obtaining a cutting force signal vector list, obtaining target vector matrix lists of different embedding dimensions according to the cutting force signal vector, performing singular value decomposition on each target vector matrix to obtain a target information quantity list, obtaining a target change rate list according to the target information quantity list, obtaining a key change rate information list according to the target change rate list, then performing singular value decomposition on the key vector matrix by using the target vector matrix corresponding to the specified embedding dimension as a key vector matrix, obtaining the key information quantity list, obtaining the maximum information quantity order, reconstructing A by a negative angle method after setting the key information quantity of the maximum order in the key information quantity list to be 0. The noise in the cutting force signal is removed through the processing, the interference of the noise to the cutting force signal is avoided, the user cannot acquire the accurate cutting state of the cutting tool, the error judgment is generated, the working efficiency of the cutting tool is reduced due to the fact that the user changes the parameters of the cutting tool according to the wrong cutting state of the cutting tool, the cutting force signal after the noise is removed can more accurately show the cutting state of the cutting tool, the user can more timely find the problems of the cutting tool, and then the user can optimize the parameters of the cutting tool according to the cutting state of the cutting tool, therefore, the working efficiency of the cutting tool is improved, and the identification precision of the cutting tool state is improved.
In the specific embodiment of the present invention, before S100, the method further includes the following steps:
s001, obtaining a target origin Z corresponding to a target coordinate axis where a cutting tool is located 0 (ii) a Wherein Z is 0 And the tool setting safety position is located on the target coordinate axis.
In particular, the target coordinate axis may be understood as an axis generated by connecting a cutting tool in a turning tool and a work piece to be cut and extending to both ends.
Further, the safe tool setting position can be understood as that the cutting tool is far away from the workpiece to be cut on the target coordinate axis and is at a distance d from the workpiece to be cut 1 The position of (a).
Further, d can be set by those skilled in the art according to actual requirements 1 And will not be described herein.
S002, acquiring a first position point Z on the target coordinate axis 1 And controlling the cutting tool in an axial direction along the target coordinate axis from Z 0 To Z 1 Moving; wherein Z is 1 The following conditions are met:
Z 1 ＝W 1 -k, wherein W 1 And k is a distance adjusting factor, and is a last tool setting position point corresponding to the current target coordinate axis.
Specifically, k satisfies the following condition:
k = β × γ, where β is an installation error corresponding to replacement of the cutting tool, and γ is a safety factor corresponding to replacement of the cutting tool.
In the embodiment of the invention, β satisfies the following condition:
β＝a*R 0 ，R 0 and a is a preset cutting tool installation error coefficient, wherein the radius is the cutting edge radius of the current cutting tool.
Further, a ∈ [0.4,0.6].
Further, a =0.5; due to the structure of the turning tool, the distance between the current tool setting position point on the target coordinate axis and the previous tool setting position point is close, the current tool setting position point may be far away from the target original point on the basis of the previous tool setting position point, and the current tool setting position point may also be close to the target original point on the basis of the previous tool setting position point, so that a spacing distance needs to be set between the first position point and the previous tool setting position point, and the spacing distance is related to the cutting edge radius of the cutting tool, therefore, the installation error of the cutting tool is set to be 0.5 times of the cutting tool radius, the cutter error in the subsequent process of the cutting tool is avoided, and the current tool setting position point is more accurate.
In the embodiment of the present invention, a person skilled in the art can set a safety factor corresponding to the replacement of the cutting tool according to actual requirements, which is not described herein again.
S003, from Z in the axial direction of the cutting tool along the target coordinate axis 0 To Z 1 During the movement, a first cutting force signal list QA = (QA) is acquired 1 ，QA 2 ，……，QA i ，……，QA m ) I =1,2, … …, m, m is the number of first cutting force signals; among them, QA i Is the i-th first cutting force signal acquired.
Specifically, the first cutting force signal is a denoised cutting force signal.
In the embodiment of the present invention, after S900, the method further includes the following steps:
s004, acquiring a first spectrum entropy list H1= (H1) according to QA 1 ，H1 2 ，……，H1 i ，……，H1 m ) (ii) a Wherein, H1 i Is A i Corresponding first spectral entropy, H1 i The following conditions are met:
In the embodiment of the present invention, H1 may be obtained by:
s041, fourier transform is performed on each first cutting force signal in QA, and a first amplitude value list set f = (f) corresponding to A is obtained 1 ，f 2 ，……，f i ，……，f m ) (ii) a Wherein A is i First amplitude value list f obtained after Fourier transform i ＝(f i1 ，f i2 ，……，f ir ，……，f iR(i) )；
Specifically, a fast fourier transform is performed on each first cutting force signal in QA to obtain a list of first amplitude values corresponding to QA.
And S042, acquiring a first spectrum entropy list H1 according to the f.
In the above, a first amplitude value list is obtained by performing fourier transform on the first cutting force signal, and a first spectrum entropy list is calculated according to the first amplitude value list. The first cutting force signal is a signal acquired by the cutting tool moving between the target origin and the safe tool setting position, so that the calculated first frequency spectrum entropy is the frequency spectrum entropy corresponding to the non-tool setting position.
S005, acquiring a target frequency spectrum entropy range H according to H1 0 ＝[h1，h2](ii) a Wherein H1 is the upper limit value corresponding to the target spectrum entropy range, H2 is the lower limit value corresponding to the target spectrum entropy range, and H 0 The ratio between the amount of the corresponding first spectral entropy and m is larger than 0.99.
In the present example, H 0 Can be obtained by the following steps:
s051, acquiring a key mean value v based on H1; wherein v satisfies the following condition:
v＝∑ m i＝1 (H1 i )/m；
s052, acquiring a key standard deviation c based on H1; wherein c meets the following conditions:
s053, acquiring the key frequency F according to c and v 0 (ii) a Wherein, F 0 The following conditions are met:
in embodiments of the invention, x = v-3c or v +3c; since the first spectrum entropy in the first spectrum entropy list conforms to the normal distribution, the key frequency values corresponding to v-3c and v +3c are the same.
In particular, the critical frequency of occurrence may be understood as the number of occurrences in the first spectral entropy list, e.g. when F 0 =1, this represents spectral entropies of only 1 identical value in the first spectral entropy list.
S054, acquiring the occurrence frequency of H1 as F 0 Third spectral entropy list H3= (H3) 1 ，H3 2 ，……，H3 t ，……，H3 T ) T =1,2, … …, T is the frequency of occurrence in H1, F 0 The amount of the first spectral entropy of (a); wherein, H3 t Is the t-th third spectral entropy in H3;
specifically, the third spectral entropy may be understood as F occurring in the first spectral entropy list 0 The first spectral entropy of (1).
Further, T is less than or equal to m.
S055, based on H3, obtaining maximum spectrum entropy H3 max And minimum spectral entropy H3 min (ii) a Wherein, H3 max And H3 min The following conditions are met:
H3 max = max (H3), max () being a maximum value determining function;
H3 min = min (H3), min () being a minimum value determining function;
s056, mixing H3 min Taking the corresponding third spectral entropy as h1; h3 is to be max Corresponding third spectrum entropy is taken as H2 to obtain H 0 。
Obtaining a key frequency by obtaining a key mean value and a key standard deviation, obtaining a third spectral entropy list generated by obtaining a third spectral entropy list with the occurrence frequency as the key frequency in the first spectral entropy list, obtaining a maximum spectral entropy and a minimum spectral entropy in the third spectral entropy list, and taking the third spectral entropy corresponding to the minimum spectral entropy as an upper limit value corresponding to a target spectral entropy range; and taking the third spectral entropy corresponding to the maximum spectral entropy as a lower limit value corresponding to the target spectral entropy range. Since the first spectral entropy in the first spectral entropy list conforms to normal distribution, and the first spectral entropies in the first spectral entropy list are all the spectral entropies corresponding to the non-tool-setting positions, the distribution value of the first frequency under normal distribution is set to be c-3v or c +3v according to the three-sigma criterion under normal distribution, and the target spectral entropy range obtained through the steps can include more than ninety percent of the first spectral entropies, H is therefore more than ninety percent of the first spectral entropies 0 The corresponding first spectral entropy contains spectral entropy corresponding to a majority of the non-tool-set positions.
S006, according to Z 1 And Z 0 Acquiring a target position point list D = (D) on a target coordinate axis 1 ，D 2 ，……，D t ，……，D T ) T =1,2, … …, T is the number of target location points in D; wherein D is t Is the t-th target position point on the target coordinate axis, D t And Z 1 Is greater than D t-1 And Z 1 And D is a distance between t And Z 0 Is greater than D t-1 And Z 0 A distance between, D 1 And Z 1 Being a point of the same position, D t -D t-1 ＝d 0 ；d 0 Is a preset distance threshold.
In particular, d 0 ＝10nm。
S007, acquiring initial parameters g1=1 and obtaining the initial parameters according to D g1 Controlling the cutting tool to carry out tool setting operation;
the tool setting operation comprises the following steps:
s071, controlling the cutting tool to move from D g1 Move to D g1+1 ；
S072, obtaining the axial direction slave D of the cutting tool along the target coordinate axis g1 To D g1+1 During the movement, a second cutting force signal list XB = (XB) is acquired 1 ，XB 2 ，……，XB q ，……，XB Q ) Q =1,2, … …, Q is the number of second cutting force signals; wherein, XB q The obtained q-th second cutting force signal is obtained;
specifically, the second cutting force signal is a denoised cutting force signal.
And further, the second cutting force signal is processed in the process that the cutter moves to the next target position point, and when the cutting cutter moves to the next target position point but the second cutting force signal is not completely processed, the cutting cutter stops moving to avoid contacting with the workpiece and generating abrasion on the cutting cutter and the workpiece.
S073, acquiring a second spectrum entropy list H2= (H2) according to QB 1 ，H2 2 ，……，H2 q ，……，H2 Q ) (ii) a Wherein, the first and the second end of the pipe are connected with each other,H2 q is B q A corresponding second spectral entropy;
s074, when H 2 In the presence of a compound other than H 0 At a second spectral entropy of (D) g+1 Carrying out tool setting at the position points; otherwise, g = g +1 is set, and S071 is executed.
In the specific embodiment of the invention, a second cutting force signal list acquired by the cutting tool in the moving process of moving the current target position point to the next target position point is acquired, a second frequency spectrum entropy corresponding to the second cutting force signal is acquired to obtain a second frequency spectrum entropy list, and when a second frequency spectrum entropy which does not belong to the target frequency spectrum entropy range exists in the second frequency spectrum entropy list, tool setting is carried out at the next target position point; otherwise, the cutting tool continues to move forwards, the steps are executed again, and when the fact that the second spectrum entropy in the second spectrum entropy list does not belong to the target spectrum entropy range is known, the cutting tool is stopped to move and the cutting tool is set at the next target position point of the current position point of the cutting tool. The cutter is automatically controlled to move, so that manual operation is avoided, the cost is reduced, and the efficiency is improved.
Embodiments of the present invention also provide a non-transitory computer-readable storage medium, which may be disposed in an electronic device to store at least one instruction or at least one program for implementing a method of the method embodiments, where the at least one instruction or the at least one program is loaded into and executed by a processor to implement the method provided by the above embodiments.
Embodiments of the present invention also provide an electronic device comprising a processor and the aforementioned non-transitory computer-readable storage medium.
Embodiments of the present invention also provide a computer program product comprising program code means for causing an electronic device to carry out the steps of the method according to various exemplary embodiments of the invention described above in the present description, when said program product is run on the electronic device.
Although some specific embodiments of the present invention have been described in detail by way of illustration, it should be understood by those skilled in the art that the above illustration is only for the purpose of illustration and is not intended to limit the scope of the invention. It will also be appreciated by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.
Claims (9)
1. A method of removing cutting force signal noise, the method comprising the steps of:
s100, a cutting force signal information list a = (a) is acquired 1 ，A 2 ，……，A i ，……，A m ) I =1,2, … …, m, m is the number of cutting force signal information; wherein A is i Cutting force signal information is the ith cutting force signal information, and the cutting force signal information at least comprises: acquiring the time of the corresponding cutting force signal and the corresponding signal intensity of the cutting force signal; the signal intensity is used for representing the magnitude of the cutting force;
s200, vectorizing A to obtain a cutting force signal vector list B = (B) 1 ，B 2 ，……，B i ，……，B m ) (ii) a Wherein, B i Is a pair A i Carrying out vectorization processing to obtain a cutting force signal vector;
s300, according to B, obtaining a target vector matrix list BZ = (BZ) 1 ，BZ 2 ，……，BZ j ，……，BZ n ) J =1,2, … …, n, n is the number of the target vector matrix obtained according to m; wherein BZ j Embedding a target vector matrix with dimension j +1, wherein the target vector matrix is a token matrix, BZ j The following conditions are met:
wherein p satisfies the following condition p = m-j;
s400, singular value decomposition is carried out on each target vector matrix in the BZ, and a target information quantity list P = (P) is obtained 1 ，P 2 ，……，P j ，……，P n ) (ii) a Wherein, P j Is BZ j The corresponding target information quantity is a second information quantity of a diagonal line in a singular value matrix obtained after singular value decomposition is carried out on a target vector matrix;
s500, acquiring a target change rate list H = (H) according to P 1 ，H 2 ，……，H g ，……，H n-1 ) G =1,2, … …, n-1; wherein H g Is P j+1 Corresponding target rate of change, H g The following conditions are met:
H g ＝P g+1 -P g ；
s600, acquiring a key change rate information list Q = (Q) according to H 1 ，Q 2 ，……，Q k ，……，Q n-3 ) K =1,2, … …, n-3; wherein the kth key rate of change information Q k ＝(S k ，v 1 k ，v 2 k ，q k )，S k For the critical embedding dimension, v 1 k Is the first key rate of change, v 2 k Is the second key rate of change, q k Is S k Corresponding key change value, q k The following conditions are met:
S700, embedding dimension S to be specified 0 The corresponding target vector matrix is used as a key vector matrix XZ; wherein S is 0 Embedding the corresponding key dimension when the second key change value in the key change rate information list is 0;
s800, singular value decomposition is carried out on the XZ, and a key information amount list D = (D) is obtained 1 ，D 2 ，……，D f ，……，D F ) F =1,2, … …, F is the amount of key information amount corresponding to XZ; wherein D is f For singular values of XZF, obtaining the key information amount after decomposition;
s900, acquiring the maximum information quantity order t; wherein t meets the following condition:
t＝arg f min(d f ) And argmin () is a variable value function corresponding to the variable at the minimum value of the function, wherein d f The following conditions are met:
/>
s1000, adding D into D t And after each key information quantity is set to be 0, reconstructing A through an anti-diagonal method to obtain a denoised cutting force signal information list.
2. The method of claim 1, wherein a is obtained by:
s101, acquiring target sampling frequency f 0 And a sampling duration t 0 ；
S102, according to f 0 And t 0 Acquiring a target sampling number Num;
s103, using f 0 The cutting signal of the cutting tool is carried out for a time period t 0 To obtain a.
3. Method according to claim 2, characterized in that f 0 Can be obtained by the following steps:
s1011, obtaining the center frequency f corresponding to the current cutting signal of the cutting tool c ；
S1012, when the current cutting signal indicates that the cutting tool is in the working state, executing S1013; when the current cutting signal indicates that the cutting tool is in a non-working state, executing S1014;
s1013, according to f c Obtaining f 0 ＝6f c ；
S1014, according to f c Obtaining f 0 ＝3f c 。
4. The method of claim 3, wherein Num satisfies the following condition:
Num＝f 0 *t 0 。
5. method according to claim 4, characterized in that t is 0 ＝1s。
6. The method of claim 1,
wherein +>
Rounding up the symbol.
7. The method of claim 1, wherein in S200, a is vectorized with Single Instruction Multiple Data to obtain B.
8. A non-transitory computer readable storage medium having stored therein at least one instruction or at least one program, the at least one instruction or the at least one program being loaded and executed by a processor to implement the method of any one of claims 1-7.
9. An electronic device comprising a processor and the non-transitory computer readable storage medium of claim 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310114634.9A CN115964619B (en) | 2023-02-15 | 2023-02-15 | Method for removing cutting force signal noise, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310114634.9A CN115964619B (en) | 2023-02-15 | 2023-02-15 | Method for removing cutting force signal noise, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115964619A true CN115964619A (en) | 2023-04-14 |
| CN115964619B CN115964619B (en) | 2023-09-22 |
Family
ID=87353112
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310114634.9A Active CN115964619B (en) | 2023-02-15 | 2023-02-15 | Method for removing cutting force signal noise, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115964619B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106849131A (en) * | 2017-04-01 | 2017-06-13 | 福州大学 | A kind of low-frequency oscillation modal identification method based on quadravalence mixing average accumulated amount with improvement TLS ESPRIT algorithms |
| US20200049782A1 (en) * | 2017-02-14 | 2020-02-13 | The United States of America, as represented by the Secretary, Department of Health and Human Ser.. | Denoising of dynamic magnetic resonance spectroscopic imaging using low rank approximations in the kinetic domain |
| CN113779914A (en) * | 2021-09-08 | 2021-12-10 | 福州大学 | TLS-ESPRIT-based transformer oil paper insulation expansion Debye equivalent circuit parameter identification method |
| CN114626193A (en) * | 2021-12-30 | 2022-06-14 | 南昌大学 | Method for reducing noise of leakage structure vibration signal based on improved variational modal decomposition |
| CN114977216A (en) * | 2022-06-13 | 2022-08-30 | 国网河北省电力有限公司电力科学研究院 | Parameter identification method of oscillation signal and terminal |
-
2023
- 2023-02-15 CN CN202310114634.9A patent/CN115964619B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200049782A1 (en) * | 2017-02-14 | 2020-02-13 | The United States of America, as represented by the Secretary, Department of Health and Human Ser.. | Denoising of dynamic magnetic resonance spectroscopic imaging using low rank approximations in the kinetic domain |
| CN106849131A (en) * | 2017-04-01 | 2017-06-13 | 福州大学 | A kind of low-frequency oscillation modal identification method based on quadravalence mixing average accumulated amount with improvement TLS ESPRIT algorithms |
| CN113779914A (en) * | 2021-09-08 | 2021-12-10 | 福州大学 | TLS-ESPRIT-based transformer oil paper insulation expansion Debye equivalent circuit parameter identification method |
| CN114626193A (en) * | 2021-12-30 | 2022-06-14 | 南昌大学 | Method for reducing noise of leakage structure vibration signal based on improved variational modal decomposition |
| CN114977216A (en) * | 2022-06-13 | 2022-08-30 | 国网河北省电力有限公司电力科学研究院 | Parameter identification method of oscillation signal and terminal |
Non-Patent Citations (5)
| Title |
|---|
| RAJ RAO NADAKUDITI: "Matrix Denoising by Optimal，Data-Driven Singular Value Shrinkage", 《IEEE TRANSACTIONS ON INFORMATION THEORY》, vol. 60, no. 5, pages 3002 - 3018, XP011545791, DOI: 10.1109/TIT.2014.2311661 * |
| XINGXIAN BAO 等: "Modal parameters identification based on noise reduction for jacket type offshore platforms", 《JOURNAL OF VIBROENGINEERING》, pages 1219 - 1230 * |
| 朱锐;杭晓晨;姜东;费庆国;靳文冰;: "基于奇异值分解的ERA改进算法及模态定阶", 《振动.测试与诊断》, vol. 38, no. 01, pages 115 - 122 * |
| 李华;刘韬;伍星;陈庆;: "基于SVD和熵优化频带熵的滚动轴承故障诊断研究", 《振动工程学报》, vol. 31, no. 02, pages 358 - 364 * |
| 王树青;林裕裕;孟元栋;高志强;: "一种基于奇异值分解技术的模型定阶方法", 《振动与冲击》, vol. 31, no. 15, pages 87 - 91 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115964619B (en) | 2023-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109597315B (en) | Method, equipment and system for identifying health degradation state of mechanical equipment | |
| CN111898443B (en) | Flow monitoring method for wire feeding mechanism of FDM type 3D printer | |
| CN108445838A (en) | A kind of numerically-controlled machine tool processing quality analysis method, grader and equipment | |
| CN112518425B (en) | Intelligent machining cutter wear prediction method based on multi-source sample migration reinforcement learning | |
| CN116939047B (en) | Data intelligent communication method for numerical control machine tool system | |
| CN111168569B (en) | Grinding material removal amount prediction method, device, equipment and storage medium | |
| CN113159121B (en) | Priori knowledge model-based robot polishing removal prediction method and device | |
| CN115374822A (en) | Fault diagnosis method and system based on multi-level feature fusion | |
| CN115964619A (en) | Method for removing cutting force signal noise, electronic equipment and storage medium | |
| CN110084301B (en) | Hidden Markov model-based multi-working-condition process working condition identification method | |
| CN117093857B (en) | Rapid processing method for intelligent ring test data | |
| CN110543869A (en) | Ball screw service life prediction method and device, computer equipment and storage medium | |
| CN113408696A (en) | Fixed point quantization method and device of deep learning model | |
| CN115157005B (en) | Method, device, equipment and storage medium for monitoring tool wear based on strain | |
| CN114997242B (en) | Extremum positioning waveform extension LMD signal decomposition method | |
| CN116176531A (en) | Method and device for determining performance index of opening degree adjustment and storage medium | |
| CN111444578B (en) | Automatic calibration method of variable modulus model parameters based on bending process | |
| Van Herreweghe et al. | A machine learning-based approach for predicting tool wear in industrial milling processes | |
| CN116038425A (en) | Method for determining tool setting position of cutting tool, electronic equipment and storage medium | |
| CN115034389A (en) | Neural network quantization and processing method and device, electronic equipment and storage medium | |
| CN114186191A (en) | Method, device and equipment for calculating coordinate transformation matrix and readable storage medium | |
| CN112241787A (en) | Neural network model pruning method and device | |
| CN113110972A (en) | Method, device and medium for detecting time sequence data abnormity | |
| CN111047039A (en) | Deep neural network clipping method and device and electronic equipment | |
| CN115685170B (en) | Active sonar target echo detection method based on reinforcement learning |
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 |