CN114004044B - Machine tool spindle thermal error quick identification method based on temperature sensitive points - Google Patents
Machine tool spindle thermal error quick identification method based on temperature sensitive points Download PDFInfo
- Publication number
- CN114004044B CN114004044B CN202111344767.2A CN202111344767A CN114004044B CN 114004044 B CN114004044 B CN 114004044B CN 202111344767 A CN202111344767 A CN 202111344767A CN 114004044 B CN114004044 B CN 114004044B
- Authority
- CN
- China
- Prior art keywords
- temperature
- moment
- value
- time
- error
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000001914 filtration Methods 0.000 claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 81
- 230000009965 odorless effect Effects 0.000 claims description 32
- 238000005259 measurement Methods 0.000 claims description 25
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000010219 correlation analysis Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 claims description 3
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 claims description 3
- 238000009529 body temperature measurement Methods 0.000 claims description 3
- 230000021615 conjugation Effects 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 230000000704 physical effect Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000003111 delayed effect Effects 0.000 description 2
- 238000003754 machining Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 238000004904 shortening Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Geometry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computing Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Feedback Control In General (AREA)
- Numerical Control (AREA)
- Automatic Control Of Machine Tools (AREA)
Abstract
The invention discloses a quick identification method of a machine tool spindle thermal error based on temperature sensitive points, which comprises the steps of calculating a bias correlation coefficient; based on a K-mean++ clustering algorithm and a bias correlation coefficient, selecting an initial clustering center to obtain a clustering combination; based on the partial correlation coefficient, selecting key temperature sensitive points from the clustering combination to obtain a temperature sensitive point combination; sending the data in the key temperature sensitive point combination into a BP neural network for training and selecting; establishing an exponential type machine tool temperature rise model based on data in the optimal temperature sensitive point combination, and performing self-adaptive adjustment based on a standard Kalman filtering algorithm; calculating to obtain a time delay time; calculating and obtaining identification time of each optimal temperature sensitive point of a machine tool spindle based on the delay time; and selecting the maximum value in the identification time to be unified as the identification time of the temperature prediction of each optimal temperature sensitive point, and obtaining the unified identification time. The invention realizes the quick identification of the thermal error of the machine tool spindle.
Description
Technical Field
The invention relates to the field of numerical control machine tool thermal error identification, in particular to a quick machine tool spindle thermal error identification method based on temperature sensitive points.
Background
The generation of thermal errors of the machine tool is unavoidable, and the thermal errors occupy a larger proportion in the total error source of the machine tool. Thermal balance is an effective method for reducing the influence of thermal errors and improving the machining precision of a machine tool. However, it takes a long time for the machine tool to reach a thermal equilibrium state from the start. The process is reasonably controlled, the time for the machine tool to reach the thermal equilibrium state is shortened, the machining efficiency of the machine tool is improved, and the machine tool is a problem which needs to be solved in the equipment manufacturing industry. The quick identification of the thermal error of the machine tool spindle is the basis for shortening the thermal balance time, and is one of the prerequisites for improving the precision.
The existing method needs to be further improved in terms of rapidity and high efficiency when identifying the thermal error of the machine tool spindle. Aiming at the problem of main shaft thermal error identification, most of the thermal error models are established by adopting methods of genetic neural network, grey theory, clustering ambiguity, linear regression and the like, and the models need a large amount of thermal error measurement data and are subjected to complex training. The whole process needs a large amount of calculation to obtain the complete machine tool spindle thermal error, and often the time required by model training is more than half of the total identification time. Therefore, research on a quick identification method of the thermal error of the machine tool spindle is an important part of shortening the thermal balance time of the machine tool.
Disclosure of Invention
Aiming at the defects in the prior art, the quick identification method for the thermal error of the machine tool spindle based on the temperature sensitive point solves the problem of long identification time of the thermal error in the prior art.
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
the quick identification method for the thermal error of the machine tool spindle based on the temperature sensitive point comprises the following steps:
S1, calculating a simple correlation coefficient between each temperature variable and each thermal error, and establishing a correlation coefficient matrix according to the simple correlation coefficient;
s2, calculating an inverse matrix of the correlation coefficient matrix based on a partial correlation analysis theory;
s3, calculating a bias correlation coefficient based on an inverse matrix of the correlation coefficient matrix;
s4, selecting an initial clustering center based on a K-mean++ clustering algorithm and a bias correlation coefficient to obtain a clustering combination;
S5, selecting key temperature sensitive points from the clustering combination based on the partial correlation coefficient to obtain a temperature sensitive point combination;
S6, sending the data in the key temperature sensitive point combination into a BP neural network for training and selecting to obtain an optimal temperature sensitive point combination;
s7, establishing an exponential type machine tool temperature rise model based on a thermal model theory and a machine tool heat conduction theory, and obtaining a temperature rise state equation by combining a standard odorless Kalman filtering algorithm;
S8, acquiring temperature actual measurement values in an initial time period of each optimal temperature sensitive point in the optimal temperature sensitive point combination, and performing self-adaptive adjustment based on a temperature rise state equation and a standard odorless Kalman filtering algorithm to obtain an adjusted optimal temperature predicted value;
S9, calculating an average absolute percentage error between the actual temperature measured value and the optimal temperature predicted value to obtain a time delay time;
S10, carrying out root mean square error calculation on the temperature actual measured value and the temperature predicted value based on the delay time to respectively obtain identification time of each optimal temperature sensitive point;
s11, selecting the maximum value in the identification time of each optimal temperature sensitive point to be unified as the identification time of the temperature prediction of each optimal temperature sensitive point, and obtaining unified identification time;
And S12, establishing a temperature-thermal error relation model, and respectively calculating thermal errors in three directions of a machine tool spindle based on the optimal temperature predicted value corresponding to the unified identification time to finish thermal error identification.
Further, the specific method for calculating the bias correlation coefficient in step S3 is as follows:
according to the formula:
Obtaining a bias correlation coefficient c ij between the i-th set of temperatures and the j-th set of thermal errors; wherein lambda ij is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, lambda ii is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, and lambda jj is the jth row and jth column parameter in the inverse matrix of the correlation coefficient matrix.
Further, the specific method for obtaining the temperature rise state equation in step S7 is as follows:
according to the formula:
yk＝Tk+vk
Obtaining a temperature rise state equation; wherein the temperature state vector x k, the temperature predicted value y k,T∞,k-1 is the ambient temperature at the moment of k-1, T k-1 is the predicted temperature at the moment of k-1, ε k-1 is the coefficient related to the physical property and the initial temperature of the system at the moment of k-1, deltat is the sampling time, e is the natural number logarithm, and omega k-1 is the system noise at the moment of k-1; t k is the predicted temperature at time k, and v k is the measurement noise.
Further, the specific method of step S8 is as follows:
S8-1, acquiring a temperature actual measurement value and a temperature predicted value in an initial time period of each optimal temperature sensitive point, and calculating a residual error r t between the temperature actual measurement value and the temperature predicted value;
S8-2, judging whether the absolute value of the residual error r t is smaller than or equal to a positive threshold value, if yes, not processing and entering a step S9; otherwise, entering a step S8-3;
S8-3, judging whether the residual error r t is larger than a positive threshold value, and if so, entering a step S8-4; otherwise, enter step S8-7;
S8-4, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-5; otherwise, enter step S8-6;
S8-5, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, decreasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
S8-6, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-7, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-8; otherwise, enter step S8-9;
s8-8, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-9, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, reducing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9.
Further, the specific method for adjusting the Q covariance matrix and the R covariance matrix in step S8 is as follows:
The result of the adjustment of the Q covariance matrix in step S8 is Q update: when the Q covariance matrix is to be increased, When Q covariance matrix is reduced,/>
Where k is the imaginary part, Q is the pre-adjustment Q covariance,Conjugation to the pre-adjustment Q covariance;
The result after the adjustment of the R covariance matrix in step S8 is R update:
Rupdate＝hjR
Wherein h is an adjustment coefficient, j is the adjustment times, and R is the covariance of the actual temperature measurement value before the first adjustment; when the R covariance matrix is increased, h is larger than 1, and when the R covariance matrix is reduced, h is larger than 0 and smaller than 1.
Further, the specific method of step S9 is as follows:
S9-1, obtaining an actual measurement value of the thermal error, and according to the formula:
Obtaining an average absolute percentage error MAPE; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first thermal error actual measurement value at time t, and n is a positive integer;
S9-2, continuously calculating the average absolute percentage error until the average absolute percentage error is smaller than the percentage parameter threshold value at a certain moment, and taking the moment as the delay moment.
Further, the specific method of step S10 is as follows:
S10-1, starting with a delay time K and according to the formula:
Performing root mean square error calculation on the thermal error actual measured value and the thermal error predicted value to obtain root mean square error beta; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first actual thermal error measurement value at time t, and M is the total number of measurements;
S10-2, searching a moment to keep the root mean square error to be minimum, and taking the moment as identification time.
Further, the specific method for establishing the temperature-thermal error relation model in step S12 is as follows:
according to the formula:
δ(t)＝δ(t-1)+αs·α·ΔT+κ
establishing a temperature-thermal error relation model; wherein delta (T) is a thermal error at the time T, delta (T-1) is a thermal error at the time T-1, alpha s is a linear expansion coefficient of the material, delta T is a difference value between an optimal temperature predicted value corresponding to uniform identification time and an ambient temperature, and alpha and kappa are undetermined coefficients of an equation.
The beneficial effects of the invention are as follows:
1. By combining partial correlation analysis, a K-mean++ clustering algorithm and a back propagation neural network (BP neural network), a comprehensive selection method of temperature sensitive points is provided, and a temperature-thermal error relation model based on the temperature of the sensitive points is established, so that the follow-up self-adaptive adjustment and quick identification are facilitated;
2. aiming at the problem that prediction divergence possibly occurs in a standard Kalman filtering algorithm, an adaptive adjustment rule is added to enhance the prediction robustness;
3. Designing a delay criterion: the delay time is obtained by calculation based on the average absolute percentage error and is used for delay processing, and a time judgment criterion is designed: and calculating the identification time of the delayed system based on the root mean square error, further calculating the identification time of the machine tool spindle in three directions, and selecting the maximum time as the unified identification time of the three directions to realize rapid thermal error identification.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a root mean square error plot of the shortest identification time for finding a selected sensitive point according to the present invention;
FIG. 3 is a graph of temperature predictions for selected sensitive spots in accordance with the present invention;
FIG. 4 is a graph showing the result of the identification of thermal errors in the X-direction of the spindle according to the present invention;
FIG. 5 is a graph showing the result of the thermal error recognition in the Y-direction of the spindle according to the present invention;
FIG. 6 is a graph showing the result of the thermal error recognition in the Z direction of the spindle according to the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, the quick identification method for the thermal error of the machine tool spindle based on the temperature sensitive point comprises the following steps:
S1, calculating a simple correlation coefficient between each temperature variable and each thermal error, and establishing a correlation coefficient matrix according to the simple correlation coefficient;
s2, calculating an inverse matrix of the correlation coefficient matrix based on a partial correlation analysis theory;
s3, calculating a bias correlation coefficient based on an inverse matrix of the correlation coefficient matrix;
s4, selecting an initial clustering center based on a K-mean++ clustering algorithm and a bias correlation coefficient to obtain a clustering combination;
S5, selecting key temperature sensitive points from the clustering combination based on the partial correlation coefficient to obtain a temperature sensitive point combination;
S6, sending the data in the key temperature sensitive point combination into a BP neural network for training and selecting to obtain an optimal temperature sensitive point combination;
s7, establishing an exponential type machine tool temperature rise model based on data in the optimal temperature sensitive point combination and a standard odorless Kalman filtering algorithm to obtain a temperature rise state equation;
S8, acquiring temperature actual measurement values in an initial time period of each optimal temperature sensitive point, and carrying out self-adaptive adjustment based on a temperature rise state equation and a standard odorless Kalman filtering algorithm to obtain an adjusted optimal temperature predicted value;
S9, calculating an average absolute percentage error between the actual temperature measured value and the optimal temperature predicted value to obtain a time delay time;
S10, carrying out root mean square error calculation on the temperature actual measured value and the temperature predicted value based on the delay time to respectively obtain identification time of each optimal temperature sensitive point;
s11, selecting the maximum value in the identification time of each optimal temperature sensitive point to be unified as the identification time of the temperature prediction of each optimal temperature sensitive point, and obtaining unified identification time;
And S12, establishing a temperature-thermal error relation model, and respectively calculating thermal errors in three directions of a machine tool spindle based on the optimal temperature predicted value corresponding to the unified identification time to finish thermal error identification.
The specific method for calculating the bias correlation coefficient in the step S3 is as follows:
according to the formula:
Obtaining a bias correlation coefficient c ij between the i-th set of temperatures and the j-th set of thermal errors; wherein lambda ij is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, lambda ii is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, and lambda jj is the jth row and jth column parameter in the inverse matrix of the correlation coefficient matrix.
The specific method for obtaining the temperature rise state equation in the step S7 is as follows:
according to the formula:
yk＝Tk+vk
Obtaining a temperature rise state equation; wherein the temperature state vector x k, the temperature predicted value y k,T∞,k-1 is the ambient temperature at the moment of k-1, T k-1 is the predicted temperature at the moment of k-1, ε k-1 is the coefficient related to the physical property and the initial temperature of the system at the moment of k-1, deltat is the sampling time, e is the natural number logarithm, and omega k-1 is the system noise at the moment of k-1; t k is the predicted temperature at time k, and v k is the measurement noise.
The specific method of step S8 is as follows:
S8-1, acquiring a temperature actual measurement value and a temperature predicted value in an initial time period of each optimal temperature sensitive point, and calculating a residual error r t between the temperature actual measurement value and the temperature predicted value;
S8-2, judging whether the absolute value of the residual error r t is smaller than or equal to a positive threshold value, if yes, not processing and entering a step S9; otherwise, entering a step S8-3;
S8-3, judging whether the residual error r t is larger than a positive threshold value, and if so, entering a step S8-4; otherwise, enter step S8-7;
S8-4, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-5; otherwise, enter step S8-6;
S8-5, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, decreasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
S8-6, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-7, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-8; otherwise, enter step S8-9;
s8-8, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-9, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, reducing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9.
The specific method for adjusting the Q covariance matrix and the R covariance matrix in the step S8 is as follows:
The result of the adjustment of the Q covariance matrix in step S8 is Q update: when the Q covariance matrix is to be increased, When Q covariance matrix is reduced,/>
Where k is the imaginary part, Q is the pre-adjustment Q covariance,Conjugation to the pre-adjustment Q covariance; the result after the adjustment of the R covariance matrix in step S8 is R update:
Rupdate＝hjR
Wherein h is an adjustment coefficient, j is the adjustment times, and R is the covariance of the actual temperature measurement value before the first adjustment; when the R covariance matrix is increased, h is larger than 1, and when the R covariance matrix is reduced, h is larger than 0 and smaller than 1.
The specific method of step S9 is as follows:
S9-1, obtaining an actual measurement value of the thermal error, and according to the formula:
Obtaining an average absolute percentage error MAPE; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first thermal error actual measurement value at time t, and n is a positive integer;
S9-2, continuously calculating the average absolute percentage error until the average absolute percentage error is smaller than the percentage parameter threshold value at a certain moment, and taking the moment as the delay moment.
The specific method of step S10 is as follows:
S10-1, starting with a delay time K and according to the formula:
Performing root mean square error calculation on the thermal error actual measured value and the thermal error predicted value to obtain root mean square error beta; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first actual thermal error measurement value at time t, and M is the total number of measurements;
S10-2, searching a moment to keep the root mean square error to be minimum, and taking the moment as identification time.
The specific method for establishing the temperature-thermal error relation model in the step S12 is as follows:
according to the formula:
δ(t)＝δ(t-1)+αs·α·ΔT+κ
establishing a temperature-thermal error relation model; wherein delta (T) is a thermal error at the time T, delta (T-1) is a thermal error at the time T-1, alpha s is a linear expansion coefficient of the material, delta T is a difference value between an optimal temperature predicted value corresponding to uniform identification time and an ambient temperature, and alpha and kappa are undetermined coefficients of an equation.
As shown in fig. 2, the delay time is 18min, i.e. 18min, lt1=21 min, lt15=26 min, lt25=23 min, lt26=27 min, and then 27min is selected as the thermal error recognition time, which is superior to the three to four hour time of the conventional technology.
As shown in fig. 3, the temperature below the predicted value of T25 is most gentle.
As shown in fig. 4 to 6, the thermal error prediction results of the machine tool spindle X, Y, Z in the three directions at the recognition time and the residual results with the actual measured values are respectively shown.
The invention combines partial correlation analysis, a K-mean++ clustering algorithm and a back propagation neural network (BP neural network), provides a comprehensive selection method of temperature sensitive points, establishes a temperature-thermal error relation model based on the temperature of the sensitive points, and is convenient for subsequent self-adaptive adjustment and rapid identification;
Aiming at the problem that prediction divergence possibly occurs in a standard Kalman filtering algorithm, an adaptive adjustment rule is added to enhance the prediction robustness;
Designing a delay criterion: the delay time is obtained by calculation based on the average absolute percentage error and is used for delay processing, and a time judgment criterion is designed: and calculating the identification time of the delayed system based on the root mean square error, further calculating the identification time of the machine tool spindle in three directions, and selecting the maximum time as the unified identification time of the three directions to realize rapid thermal error identification.
Claims (5)
1. A machine tool spindle thermal error quick identification method based on temperature sensitive points is characterized by comprising the following steps:
S1, calculating a simple correlation coefficient between each temperature variable and each thermal error, and establishing a correlation coefficient matrix according to the simple correlation coefficient;
s2, calculating an inverse matrix of the correlation coefficient matrix based on a partial correlation analysis theory;
s3, calculating a bias correlation coefficient based on an inverse matrix of the correlation coefficient matrix;
s4, selecting an initial clustering center based on a K-mean++ clustering algorithm and a bias correlation coefficient to obtain a clustering combination;
S5, selecting key temperature sensitive points from the clustering combination based on the partial correlation coefficient to obtain a temperature sensitive point combination;
S6, sending the data in the key temperature sensitive point combination into a BP neural network for training and selecting to obtain an optimal temperature sensitive point combination;
s7, establishing an exponential type machine tool temperature rise model based on a thermal model theory and a machine tool heat conduction theory, and obtaining a temperature rise state equation by combining a standard odorless Kalman filtering algorithm;
S8, acquiring temperature actual measurement values in an initial time period of each optimal temperature sensitive point in the optimal temperature sensitive point combination, and carrying out self-adaptive adjustment based on a temperature rise state equation and a standard odorless Kalman filtering algorithm to obtain an adjusted optimal temperature prediction value, wherein the method specifically comprises the following steps of:
S8-1, acquiring a temperature actual measurement value and a temperature predicted value in an initial time period of each optimal temperature sensitive point, and calculating a residual error r t between the temperature actual measurement value and the temperature predicted value;
S8-2, judging whether the absolute value of the residual error r t is smaller than or equal to a positive threshold value, if yes, not processing and entering a step S9; otherwise, entering a step S8-3;
S8-3, judging whether the residual error r t is larger than a positive threshold value, and if so, entering a step S8-4; otherwise, enter step S8-7;
S8-4, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-5; otherwise, enter step S8-6;
S8-5, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, decreasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
S8-6, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, increasing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-7, acquiring a temperature priori estimated value, judging whether the temperature actual measured value is smaller than the temperature priori estimated value, and if so, entering a step S8-8; otherwise, enter step S8-9;
s8-8, reducing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, reducing a t-moment temperature state vector, reducing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, reducing covariance of a t-moment temperature predicted value, increasing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
s8-9, increasing a Q covariance matrix of a t-moment standard odorless Kalman filtering algorithm, increasing a t-moment temperature state vector, increasing a R covariance matrix of the t-moment standard odorless Kalman filtering algorithm, increasing covariance of a t-moment temperature predicted value, reducing a t-moment Kalman gain, reducing an optimal estimated value of the t-moment temperature state vector to obtain an optimal temperature predicted value, and entering step S9;
the specific method for adjusting the Q covariance matrix and the R covariance matrix comprises the following steps:
The result of the adjustment of the Q covariance matrix in step S8 is Q update: when the Q covariance matrix is to be increased, When Q covariance matrix is reduced,/>
Where k is the imaginary part, Q is the pre-adjustment Q covariance,Conjugation to the pre-adjustment Q covariance;
The result after the adjustment of the R covariance matrix in step S8 is R update:
Rupdate＝hjR
Wherein h is an adjustment coefficient, j is the adjustment times, and R is the covariance of the actual temperature measurement value before the first adjustment; when the R covariance matrix is increased, h is larger than 1, and when the R covariance matrix is reduced, h is larger than 0 and smaller than 1;
S9, calculating an average absolute percentage error between the actual temperature measured value and the optimal temperature predicted value to obtain a time delay time;
S10, carrying out root mean square error calculation on the temperature actual measured value and the temperature predicted value based on the delay time to respectively obtain identification time of each optimal temperature sensitive point;
s11, selecting the maximum value in the identification time of each optimal temperature sensitive point to be unified as the identification time of the temperature prediction of each optimal temperature sensitive point, and obtaining unified identification time;
S12, establishing a temperature-thermal error relation model, and respectively calculating thermal errors in three directions of a machine tool spindle based on optimal temperature predicted values corresponding to unified identification time to complete thermal error identification;
The specific method for establishing the temperature-thermal error relation model comprises the following steps:
according to the formula:
δ(t)＝δ(t-1)+αs·α·ΔT+κ
establishing a temperature-thermal error relation model; wherein delta (T) is a thermal error at the time T, delta (T-1) is a thermal error at the time T-1, alpha s is a linear expansion coefficient of the material, delta T is a difference value between an optimal temperature predicted value corresponding to uniform identification time and an ambient temperature, and alpha and kappa are undetermined coefficients of an equation.
2. The quick identification method of machine tool spindle thermal error based on temperature sensitive points according to claim 1, wherein the specific method for calculating the partial correlation coefficient in step S3 is as follows:
according to the formula:
Obtaining a bias correlation coefficient c ij between the i-th set of temperatures and the j-th set of thermal errors; wherein lambda ij is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, lambda ii is the ith row and jth column parameter in the inverse matrix of the correlation coefficient matrix, and lambda jj is the jth row and jth column parameter in the inverse matrix of the correlation coefficient matrix.
3. The quick identifying method for the machine tool spindle temperature based on the temperature sensitive point according to claim 1, wherein the specific method for obtaining the temperature rise state equation in the step S7 is as follows:
according to the formula:
yk＝Tk+vk
Obtaining a temperature rise state equation; wherein the temperature state vector x k, the temperature predicted value y k,T∞,k-1 is the ambient temperature at the moment of k-1, T k-1 is the predicted temperature at the moment of k-1, ε k-1 is the coefficient related to the physical property and the initial temperature of the system at the moment of k-1, deltat is the sampling time, e is the natural number logarithm, and omega k-1 is the system noise at the moment of k-1; t k is the predicted temperature at time k, and v k is the measurement noise.
4. The fast recognition method of machine tool spindle thermal error based on odorless kalman filter algorithm according to claim 1, wherein the specific method of step S9 is as follows:
S9-1, obtaining an actual measurement value of the thermal error, and according to the formula:
Obtaining an average absolute percentage error MAPE; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first thermal error actual measurement value at time t, and n is a positive integer;
S9-2, continuously calculating the average absolute percentage error until the average absolute percentage error is smaller than the percentage parameter threshold value at a certain moment, and taking the moment as the delay moment.
5. The fast recognition method of machine tool spindle thermal error based on odorless kalman filter algorithm according to claim 1, wherein the specific method of step S10 is as follows:
S10-1, starting with a delay time K and according to the formula:
Performing root mean square error calculation on the thermal error actual measured value and the thermal error predicted value to obtain root mean square error beta; wherein y l (t) is the first adjusted thermal error prediction value at time t, a l (t) is the first actual thermal error measurement value at time t, and M is the total number of measurements;
S10-2, searching a moment to keep the root mean square error to be minimum, and taking the moment as identification time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111344767.2A CN114004044B (en) | 2021-11-15 | 2021-11-15 | Machine tool spindle thermal error quick identification method based on temperature sensitive points |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111344767.2A CN114004044B (en) | 2021-11-15 | 2021-11-15 | Machine tool spindle thermal error quick identification method based on temperature sensitive points |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114004044A CN114004044A (en) | 2022-02-01 |
| CN114004044B true CN114004044B (en) | 2024-04-30 |
Family
ID=79928941
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111344767.2A Active CN114004044B (en) | 2021-11-15 | 2021-11-15 | Machine tool spindle thermal error quick identification method based on temperature sensitive points |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114004044B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114021368B (en) * | 2021-11-15 | 2024-08-27 | 西南交通大学 | Machine tool spindle thermal error quick identification method based on odorless Kalman filtering algorithm |
| CN114967592B (en) * | 2022-05-31 | 2024-06-21 | 安徽工业大学 | Self-adaptive selection method for machine tool thermal error temperature sensitive point |
| CN118092313B (en) * | 2024-02-28 | 2024-08-27 | 西南交通大学 | Machine tool temperature rise and main shaft thermal error look-ahead prediction method based on rotating speed region division |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109765846A (en) * | 2019-03-11 | 2019-05-17 | 哈尔滨理工大学 | The selection method of double turntable five-axle number control machine tool thermal error modeling temperature sensitive points |
| CN110231801A (en) * | 2019-06-12 | 2019-09-13 | 内蒙古工业大学 | A kind of thermotropic position error key temperatures measuring point selection method of numerically-controlled machine tool |
| CN111123833A (en) * | 2019-12-03 | 2020-05-08 | 三明学院 | Thermal error compensation calculation method, device, equipment and storage medium of machine tool |
| CN111459094A (en) * | 2020-03-09 | 2020-07-28 | 西南交通大学 | Regional selection method for temperature sensitive point combination in machine tool spindle thermal error modeling |
-
2021
- 2021-11-15 CN CN202111344767.2A patent/CN114004044B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109765846A (en) * | 2019-03-11 | 2019-05-17 | 哈尔滨理工大学 | The selection method of double turntable five-axle number control machine tool thermal error modeling temperature sensitive points |
| CN110231801A (en) * | 2019-06-12 | 2019-09-13 | 内蒙古工业大学 | A kind of thermotropic position error key temperatures measuring point selection method of numerically-controlled machine tool |
| CN111123833A (en) * | 2019-12-03 | 2020-05-08 | 三明学院 | Thermal error compensation calculation method, device, equipment and storage medium of machine tool |
| CN111459094A (en) * | 2020-03-09 | 2020-07-28 | 西南交通大学 | Regional selection method for temperature sensitive point combination in machine tool spindle thermal error modeling |
Non-Patent Citations (7)
| Title |
|---|
| Improved unscented Kalman filter algorithm-based rapid identification of thermal errors of machine tool spindle for shortening thermal equilibrium time;Guoqiang Fu等;Measurement 195 (2022) 111121;20220401;全文 * |
| Look-ahead prediction of spindle thermal errors with on-machine measurement and the cubic exponential smoothing-unscented Kalman filtering-based temperature prediction model of the machine tools;Guoqiang Fu等;Measurement 210 (2023) 112536;20230124;全文 * |
| Temperature-sensitive point selection for thermal error modeling of machine tool spindle by considering heat source regions;Guoqiang Fu等;The International Journal of Advanced Manufacturing Technology (2021) 112;20210107;第2447–2460页 * |
| 基于改进的自适应渐消UKF机床主轴热平衡试验;余文利;邓小雷;姚鑫骅;傅建中;;农业机械学报;20190306(第04期);第370-380页 * |
| 基于过程数据时段特性的数控机床热误差预测研究;邢美峰;秦波;秦岩;袁媛;王春暖;;机床与液压;20150315(第05期);第101-105页 * |
| 基于集成BP神经网络的数控机床主轴热误差建模;谭峰;殷鸣;彭骥;卫亚斌;殷国富;;计算机集成制造系统;20170703(第06期);第61-68页 * |
| 数控机床主轴温升特性快速辨识方法及新型温控结构的研究;夏晨晖;中国博士学位论文全文数据库 工程科技I辑;20160115;文章第2-4章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114004044A (en) | 2022-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114004044B (en) | Machine tool spindle thermal error quick identification method based on temperature sensitive points | |
| CN108959728B (en) | Radio frequency device parameter optimization method based on deep learning | |
| CN110163429B (en) | Short-term load prediction method based on similarity day optimization screening | |
| CN108764568B (en) | Data prediction model tuning method and device based on LSTM network | |
| CN113012766B (en) | Self-adaptive soft measurement modeling method based on online selective integration | |
| CN109670625B (en) | NOx emission concentration prediction method based on unscented Kalman filtering least square support vector machine | |
| CN110045606B (en) | Increment space-time learning method for online modeling of distributed parameter system | |
| CN111310348A (en) | Material constitutive model prediction method based on PSO-LSSVM | |
| Wen et al. | Characterizing the gap between actor-critic and policy gradient | |
| CN112418522B (en) | Industrial heating furnace steel temperature prediction method based on three-branch integrated prediction model | |
| CN117587403A (en) | Aluminum alloy template surface treatment method and system based on chemical conversion | |
| CN115329823A (en) | Dynamic depth migration prediction method and device for furnace tube temperature of delayed coking heating furnace | |
| CN116842439A (en) | Semiconductor quality prediction method based on model | |
| CN112241832A (en) | Product quality grading evaluation standard design method and system | |
| CN114021368B (en) | Machine tool spindle thermal error quick identification method based on odorless Kalman filtering algorithm | |
| CN114442557B (en) | Quick identification method and system for machine tool temperature field | |
| CN114567288B (en) | Distribution collaborative nonlinear system state estimation method based on variable decibels | |
| CN114967625A (en) | Industrial process quality prediction method based on space-time convolution latent variable optimization neighbor component analysis | |
| CN115525697A (en) | Process optimization method based on traditional Chinese medicine production data mining | |
| Blum et al. | Practical issues in applying artificial neural networks for identification in model based predictive control | |
| CN114139677A (en) | Unequal interval time sequence data prediction method based on improved GRU neural network | |
| Zhang et al. | A novel strategy of correntropy-based iterative neural networks for data reconciliation and gross error estimation in semiconductor industry | |
| Ma et al. | Applying hierarchical genetic algorithm based neural network and multiple objective evolutionary algorithm to optimize parameter design with dynamic characteristics | |
| KR19980051203A (en) | Automatic combustion control method by heating by neural network | |
| Souza et al. | Mixture of elastic net experts and its application to a polymerization batch process |
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 |