Machine tool spindle taper hole repairing method based on multi-source data driving
Through the multi-source data-driven machine spindle convolutional neural network model, the cause of damage is quickly judged and repaired, solving the problem that spindle cone hole damage needs to be disassembled and repaired in the existing technology, and achieving a fast and accurate repair process , avoid waste of resources and additional costs.
Patent Information
- Application Number
- CN202411163564.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art requires disassembly of the entire machine tool for manual repair when the taper hole of the machine tool spindle is damaged, resulting in wasting production resources, manpower and time, and it is impossible to quickly determine the cause of the damage.
The machine tool spindle convolutional neural network model is adopted based on multi-source data-driven machine tool spindle cone hole fault convolutional neural network model. By acquiring and preprocessing multi-source data, the cause of damage is quickly judged and polished, so as to achieve rapid repair without disassembly of the spindle.
It quickly determines the cause of the damage after the spindle taper hole is damaged and repairs it, shortens the repair time, improves the accuracy, does not require special equipment, and avoids bearing replacement costs.
Smart Images

Figure CN118939975A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of machine tool maintenance, and in particular to a machine tool spindle taper hole repairing method based on multi-source data drive. Background Art
[0002] The taper hole of the machining center spindle is an important part of the machine tool spindle. The geometric size and shape and position accuracy of the spindle taper hole directly affect the machining accuracy of the machine tool and the service life of the tool. The main functions of the spindle taper hole are as follows: (1) Ensure the clamping accuracy and consistency of the tool. The taper and geometric accuracy of the machining center spindle taper hole can ensure the installation accuracy of the tool and the consistency of multiple tool exchanges, thereby ensuring the service life of the tool and the machining accuracy of the workpiece; (2) Improve the rigidity of the tool. The taper of the machining center spindle taper hole is currently mainstream in 7:24 or 1:10, which can take into account the rigidity of the tool after automatic tool exchange and the tool taper shank, and reduce the vibration and deformation of the tool; (3) Ensure the interchangeability of the tool. The standardized design of the machining center spindle taper hole can ensure that different tools have the same interface, thereby improving the efficiency of tool exchange.
[0003] The spindle taper hole is the location and clamping joint between the machine tool spindle and the tool holder. It transmits the power and torque of the spindle to the tool, and is a key link that affects the coaxiality and dynamic balance of the spindle and the tool. Once the spindle taper hole is damaged, it will cause adverse effects such as blade deflection, reduced rigidity, and reduced friction after clamping the tool holder. The machining center will frequently and automatically exchange tools during operation. The spindle taper hole will be separated and combined with the tool taper handle once each time the tool is exchanged. During processing, it must withstand the impact and vibration of various tool cutting. If these links fail, it is very likely to cause damage to the taper hole joint. Based on the existing technology, each time the spindle taper hole is damaged, the entire machine tool needs to be shut down, all spindle components need to be disassembled, and the taper hole damage needs to be repaired manually, which wastes a lot of production resources, manpower and time. Summary of the invention
[0004] In order to solve the above technical problems, a machine tool spindle taper hole repair method based on multi-source data drive is provided. This technical solution solves the problems raised in the above background technology.
[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:
[0006] A machine tool spindle taper hole repair method based on multi-source data drive, comprising:
[0007] Obtain at least one possible damage cause of the taper hole of the machine tool spindle and various operating data under the corresponding damage cause, wherein the various operating data include acceleration data of the drive end, acceleration data of the motor end, acceleration data of the bearing base, time series data and motor speed;
[0008] Based on various damage causes and the corresponding operating data under the damage causes, a convolutional neural network model of machine tool spindle taper hole fault is constructed;
[0009] Acquire multi-source data of the machine tool to be repaired, and perform data preprocessing on the multi-source data of the machine tool to be repaired to obtain a data set to be detected;
[0010] The data set to be detected is input into the convolutional neural network model of the machine tool spindle taper hole fault to obtain the cause of the damage to the machine tool spindle taper hole to be repaired;
[0011] Based on the damage cause of the spindle taper hole of the machine tool to be repaired, the spindle taper hole of the machine tool is ground to obtain a repaired machine tool;
[0012] Detect the rotation accuracy of the repaired machine tool spindle bearing and obtain the spindle radial runout;
[0013] The axial movement of the repaired machine tool spindle bearing is measured to obtain the spindle axial movement;
[0014] The radial rigidity of the repaired machine tool spindle bearing is tested to obtain the spindle radial rigidity;
[0015] Determine whether the various index values of the spindle bearing meet the standards. If so, output that the machine tool spindle taper hole status is healthy. If not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other damage causes. The various index values of the spindle bearing include spindle radial runout, spindle axial runout and spindle radial rigidity.
[0016] Preferably, the construction of a convolutional neural network model for a machine tool spindle taper hole fault based on various damage causes and various operating data corresponding to the damage causes specifically includes:
[0017] Summarize the various operating data under various damage causes into an m*n operating data matrix;
[0018] Construct a 3*3 convolution kernel matrix;
[0019] Split the running data matrix into at least one 3*3 convolution frame;
[0020] Perform matrix multiplication on each convolution frame and convolution kernel matrix to obtain the convolution layer output matrix;
[0021] Determine the activation layer function;
[0022] Input the convolution layer output matrix into the activation layer function to obtain the activation layer output value;
[0023] A connection is established between the output value of the activation layer and its corresponding damage cause, and a convolutional neural network model of machine tool spindle taper hole fault is obtained;
[0024] The convolution kernel matrix is:
[0025]
[0026] Where α is the convolution kernel matrix;
[0027] The activation layer function is:
[0028]
[0029] Where f)z) is the activation layer function, and z is the eigenvalue of the convolutional layer output matrix.
[0030] Preferably, the acquiring of multi-source data of the machine tool to be repaired and performing data preprocessing on the multi-source data of the machine tool to be repaired to obtain the data set to be detected specifically includes:
[0031] Obtain multi-source data of the machine tool to be repaired;
[0032] Set the weight decay speed;
[0033] The weight at each moment is calculated using the weight formula based on the weight decay rate;
[0034] Substituting the weights at each moment into a weighted recursive average filtering formula to obtain multi-source data of the machine tool after noise reduction;
[0035] The multi-source data of the machine tool after denoising is input into the Gaussian noise function to obtain the data set to be detected;
[0036] The weight formula is:
[0037] ω i =e -vi ,
[0038] In the formula, ω i is the weight at the i-th moment, and v is the weight decay rate;
[0039] The weighted recursive average filtering formula is:
[0040] Y(i)=(1-ω i )Y(i-1)+ω i X(i),
[0041] In the formula, Y(i) is the filtering result at the i-th moment, and X(i) is the input data at the i-th moment;
[0042] The Gaussian noise function is:
[0043] S ( i ) =Y ( i )+N(μ,σ 2 ),
[0044] In the formula, S ( i ) is the Gaussian noise function output result at the i-th moment, N(μ, σ 2 ) has a mean of μ and a variance of σ 2 Gaussian distributed random function.
[0045] Preferably, the grinding of the spindle taper hole of the machine tool to be repaired based on the cause of damage to the spindle taper hole of the machine tool to obtain the repaired machine tool specifically includes:
[0046] Determine whether the damage is caused by natural wear of the spindle taper hole. If so, use special grinding equipment to perform online mechanical grinding on the spindle taper hole. If not, no output is made.
[0047] Determine whether the cause of damage is chips entering the junction between the spindle taper hole and the tool taper shank. If so, connect high-pressure gas to blow away chips, dust and cutting fluid that may be attached to the surface of the spindle taper hole. If not, no output is made.
[0048] Determine whether the damage is caused by tool collision due to machining program or tool calling error. If so, repair the tapered hole by manual grinding combined with grinding without disassembling the spindle. If not, no output is made.
[0049] Determine whether the cause of damage is illegal vibration of the tool caused by failure of the spring of the broach mechanism or damage to other components. If so, the complete spindle set needs to be removed from the machine tool and disassembled, all failed or damaged components on the shaft core need to be replaced, the shaft core needs to be clamped on a precision internal hole grinder, and the damaged tapered hole needs to be ground. If not, no output will be made.
[0050] Preferably, the detecting the rotation accuracy of the repaired machine tool spindle bearing to obtain the spindle radial runout specifically includes:
[0051] Remove the locating key on the spindle end face;
[0052] Select the exposed outer cylindrical surface of the spindle or the undamaged tapered hole surface as the area to be measured;
[0053] Clean all surfaces to be measured;
[0054] Rotate the spindle more than one turn and measure the radial runout of the spindle with a micrometer.
[0055] Preferably, the measuring of the axial movement of the repaired machine tool spindle bearing to obtain the spindle axial movement specifically includes:
[0056] Attach the magnetic base to the spindle housing;
[0057] The micrometer head is at any position on the spindle end face;
[0058] The tool switching method of loosening and tightening is adopted, and the difference of the micrometer pointer reading is observed and output as the spindle axial movement amount.
[0059] Preferably, the detecting of the radial rigidity of the repaired machine tool spindle bearing to obtain the spindle radial rigidity specifically includes:
[0060] Attach the magnetic base to the spindle housing;
[0061] The micrometer head contacts the outer surface of the spindle shaft diameter, and the value displayed by the micrometer at this time is recorded and output as the first measurement value;
[0062] The rigid measuring device is placed on the outer diameter of the spindle;
[0063] The force application point is 180° away from the contact point of the micrometer. A force of 50N is applied to the rigidity measuring device. The value displayed on the micrometer at this time is recorded and output as the second measurement value.
[0064] Cancel the applied force, record the value displayed on the dial gauge at this time and output it as the third measured value;
[0065] Performing a difference process on the second measured value and the first measured value to obtain a first rigidity comparison value;
[0066] The third measured value is subtracted from the first measured value to obtain a rigid second comparison value.
[0067] Preferably, the determining whether the various index values of the main shaft bearing meet the standards specifically includes:
[0068] Determine whether the spindle radial runout is less than 0.007mm. If so, no output is made. If not, the machine tool spindle taper hole status is output as unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes.
[0069] Determine whether the spindle axial movement is less than 0.003mm. If so, no output is made. If not, the machine tool spindle taper hole status is output as unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes.
[0070] Determine whether the first rigidity comparison value is not greater than 0.01mm and the second rigidity comparison value is 0mm. If so, output that the machine tool spindle taper hole status is healthy; if not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other causes of damage.
[0071] Compared with the prior art, the present invention has the following beneficial effects:
[0072] By inputting multi-source data into the convolutional neural network model of machine tool spindle taper hole failure, the cause of machine tool spindle taper hole damage can be quickly determined and corresponding measures can be taken. This technology has the advantages of short repair time, high precision, and no need for special equipment in most cases. It is a fast and effective repair method for machining center spindle taper hole damage without disassembling the spindle shaft to avoid the cost of bearing replacement. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] Figure 1 It is a flow chart of the machine tool spindle taper hole repair method based on multi-source data drive of the present invention;
[0074] Figure 2 A flow chart of the method for constructing a convolutional neural network model of a machine tool spindle taper hole fault according to the present invention;
[0075] Figure 3 It is a flow chart of a method for preprocessing multi-source data of a machine tool to be repaired according to the present invention;
[0076] Figure 4 This is a flow chart of the method for grinding the taper hole of a machine tool spindle according to the present invention;
[0077] Figure 5 A flow chart of a method for detecting the rotation accuracy of a repaired machine tool spindle bearing according to the present invention;
[0078] Figure 6 A flow chart of a method for measuring the axial movement of a repaired machine tool spindle bearing according to the present invention;
[0079] Figure 7 A flow chart of a method for detecting the radial rigidity of a repaired machine tool spindle bearing according to the present invention;
[0080] Figure 8 The present invention is a flow chart of a method for judging whether various index values of a spindle bearing meet the standards. DETAILED DESCRIPTION
[0081] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0082] Reference Figure 1 As shown, a machine tool spindle taper hole repair method based on multi-source data drive includes:
[0083] Obtain at least one possible damage cause of the taper hole of the machine tool spindle and various operating data under the corresponding damage cause, wherein the various operating data include acceleration data of the drive end, acceleration data of the motor end, acceleration data of the bearing base, time series data and motor speed;
[0084] Based on various damage causes and the corresponding operating data under the damage causes, a convolutional neural network model of machine tool spindle taper hole fault is constructed;
[0085] Acquire multi-source data of the machine tool to be repaired, and perform data preprocessing on the multi-source data of the machine tool to be repaired to obtain a data set to be detected;
[0086] The data set to be detected is input into the convolutional neural network model of the machine tool spindle taper hole fault to obtain the cause of the damage to the machine tool spindle taper hole to be repaired;
[0087] Based on the damage cause of the spindle taper hole of the machine tool to be repaired, the spindle taper hole of the machine tool is ground to obtain a repaired machine tool;
[0088] Detect the rotation accuracy of the repaired machine tool spindle bearing and obtain the spindle radial runout;
[0089] The axial movement of the repaired machine tool spindle bearing is measured to obtain the spindle axial movement;
[0090] The radial rigidity of the repaired machine tool spindle bearing is tested to obtain the spindle radial rigidity;
[0091] Determine whether the various index values of the spindle bearing meet the standards. If so, output that the machine tool spindle taper hole status is healthy. If not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other damage causes. The various index values of the spindle bearing include spindle radial runout, spindle axial runout and spindle radial rigidity.
[0092] Reference Figure 2 As shown in the figure, based on various damage causes and the operating data under the corresponding damage causes, the convolutional neural network model of machine tool spindle taper hole fault is constructed, which specifically includes:
[0093] Summarize the various operating data under various damage causes into an m*n operating data matrix;
[0094] Construct a 3*3 convolution kernel matrix;
[0095] Split the running data matrix into at least one 3*3 convolution frame;
[0096] Perform matrix multiplication on each convolution frame and convolution kernel matrix to obtain the convolution layer output matrix;
[0097] Determine the activation layer function;
[0098] Input the convolution layer output matrix into the activation layer function to obtain the activation layer output value;
[0099] A connection is established between the output value of the activation layer and its corresponding damage cause, and a convolutional neural network model of machine tool spindle taper hole fault is obtained;
[0100] The convolution kernel matrix is:
[0101]
[0102] Where α is the convolution kernel matrix;
[0103] The activation layer function is:
[0104]
[0105] Where f(z) is the activation layer function and z is the eigenvalue of the convolutional layer output matrix.
[0106] A convolutional neural network is a multi-level neural network constructed using mathematical convolution operations, modeled after the working principle of the biological visual system. Its purpose is to fit a highly nonlinear function that conforms to the data results based on existing data cases. The role of the convolution layer is to extract feature maps from the image, the role of the activation layer is to introduce nonlinear transformations into the network model, and the role of the final fully connected layer is to map the high-dimensional signals extracted by the previous layer to the final target result.
[0107] Reference Figure 3 As shown, obtaining multi-source data of the machine tool to be repaired, and performing data preprocessing on the multi-source data of the machine tool to be repaired to obtain the data set to be detected specifically includes:
[0108] Obtain multi-source data of the machine tool to be repaired;
[0109] Set the weight decay speed;
[0110] The weight at each moment is calculated using the weight formula based on the weight decay rate;
[0111] Substituting the weights at each moment into a weighted recursive average filtering formula to obtain multi-source data of the machine tool after noise reduction;
[0112] The multi-source data of the machine tool after denoising is input into the Gaussian noise function to obtain the data set to be detected;
[0113] The weight formula is:
[0114] ω i =e -vi ,
[0115] In the formula, ω i is the weight at the i-th moment, and v is the weight decay rate;
[0116] The weighted recursive average filtering formula is:
[0117] Y(i)=(1-ω i )Y(i-1)+ωi X(i),
[0118] In the formula, Y(i) is the filtering result at the i-th moment, and X(i) is the input data at the i-th moment;
[0119] The Gaussian noise function is:
[0120] S ( i ) =Y ( i ) +N(μ,σ 2 ),
[0121] In the formula, S ( i ) is the Gaussian noise function output result at the i-th moment, N(μ, σ 2 ) has a mean of μ and a variance of σ 2 Gaussian distributed random function.
[0122] The weighted recursive average filtering method is mainly used to eliminate or weaken the noise that is significantly different from the normal data. By limiting and eliminating trend items, outliers are filtered out from the software level to make the data smoother. Since there are various interferences in the signal acquisition process, the collected signal data contains Gaussian errors. By continuously and artificially applying weak Gaussian noise to the original signal, the data can be closer to the actual situation. Inputting these Gaussian noise-enhanced data into the network model can improve the generalization ability of the network model for fault diagnosis.
[0123] Reference Figure 4 As shown in the figure, based on the cause of the damage to the spindle taper hole of the machine tool to be repaired, the spindle taper hole of the machine tool is ground, and the repaired machine tool specifically includes:
[0124] Determine whether the damage is caused by natural wear of the spindle taper hole. If so, use special grinding equipment to perform online mechanical grinding on the spindle taper hole. If not, no output is made.
[0125] Determine whether the cause of damage is chips entering the junction between the spindle taper hole and the tool taper shank. If so, connect high-pressure gas to blow away chips, dust and cutting fluid that may be attached to the surface of the spindle taper hole. If not, no output is made.
[0126] Determine whether the damage is caused by tool collision due to machining program or tool calling error. If so, repair the tapered hole by manual grinding combined with grinding without disassembling the spindle. If not, no output is made.
[0127] Determine whether the cause of damage is illegal vibration of the tool caused by failure of the spring of the broach mechanism or damage to other components. If so, the complete spindle set needs to be removed from the machine tool and disassembled, all failed or damaged components on the shaft core need to be replaced, the shaft core needs to be clamped on a precision internal hole grinder, and the damaged tapered hole needs to be ground. If not, no output will be made.
[0128] The visible points of the spindle taper hole after grinding should be evenly distributed over an area greater than 70%. The grinding can be stopped when the runout at 300mm of the detection mandrel is less than 0.03mm, record the direction of the high point of the taper hole, and switch to the grinding process. The grinding process can not only correct the geometric shape and axis deviation of the spindle taper hole, but also improve the smoothness of the taper hole. It is a very important link. When the runout at 300mm of the detection mandrel is less than 0.015mm, use a taper inspection rod to paint and check the taper hole. At this time, there should be no obvious high points, and the visible points are mainly surface contact. Stop grinding when the contact area is greater than 80%.
[0129] Reference Figure 5 As shown in the figure, the rotation accuracy of the repaired machine tool spindle bearing is detected, and the spindle radial runout is obtained, which specifically includes:
[0130] Remove the locating key on the spindle end face;
[0131] Select the exposed outer cylindrical surface of the spindle or the undamaged tapered hole surface as the area to be measured;
[0132] Clean all surfaces to be measured;
[0133] Rotate the spindle more than one turn and measure the radial runout of the spindle with a micrometer.
[0134] Since some high spots will appear on the inner wall of the damaged spindle tapered hole, the use of a standard core rod to detect the spindle radial runout will result in a large value, which is of no reference value for evaluating the spindle bearing condition. At this time, choose the exposed outer cylindrical surface of the spindle or the undamaged tapered hole surface, and use a micrometer to detect the spindle radial runout.
[0135] Reference Figure 6 As shown in the figure, the axial movement of the repaired machine tool spindle bearing is measured, and the spindle axial movement is specifically including:
[0136] Attach the magnetic base to the spindle housing;
[0137] The micrometer head is at any position on the spindle end face;
[0138] The tool switching method of loosening and tightening is adopted, and the difference of the micrometer pointer reading is observed and output as the spindle axial movement amount.
[0139] In the process of measuring the axial play of the machine tool spindle bearing, the purpose of adsorbing the magnetic base on the spindle housing is to avoid bringing the deformation of other moving parts into the measurement results when the spindle is subjected to force.
[0140] Reference Figure 7 As shown in the figure, the radial rigidity of the repaired machine tool spindle bearing is tested, and the spindle radial rigidity includes:
[0141] Attach the magnetic base to the spindle housing;
[0142] The micrometer head contacts the outer surface of the spindle shaft diameter, and the value displayed by the micrometer at this time is recorded and output as the first measurement value;
[0143] The rigid measuring device is placed on the outer diameter of the spindle;
[0144] The force application point is 180° away from the contact point of the micrometer. A force of 50N is applied to the rigidity measuring device. The value displayed on the micrometer at this time is recorded and output as the second measurement value.
[0145] Cancel the applied force, record the value displayed on the dial gauge at this time and output it as the third measured value;
[0146] Performing a difference process on the second measured value and the first measured value to obtain a first rigidity comparison value;
[0147] The third measured value is subtracted from the first measured value to obtain a rigid second comparison value.
[0148] Reference Figure 8 As shown in the figure, judging whether the index values of the spindle bearing meet the standards specifically includes:
[0149] Determine whether the spindle radial runout is less than 0.007mm. If so, no output is made. If not, the machine tool spindle taper hole status is output as unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes.
[0150] Determine whether the spindle axial movement is less than 0.003mm. If so, no output is made. If not, the machine tool spindle taper hole status is output as unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes.
[0151] Determine whether the first rigidity comparison value is not greater than 0.01mm and the second rigidity comparison value is 0mm. If so, output that the machine tool spindle taper hole status is healthy; if not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other causes of damage.
[0152] By quickly determining the cause of the damage to the machine tool spindle taper hole and taking corresponding measures, the purpose of short repair time, high precision and no need for special equipment can be achieved in most cases. When the spindle taper hole of the machining center is damaged, the spindle shaft does not need to be disassembled to avoid the cost of bearing replacement.
[0153] Furthermore, the present solution also proposes a computer-readable storage medium on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned machine tool spindle taper hole repair method based on multi-source data drive is executed.
[0154] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).
[0155] To sum up, the advantages of the present invention are: by inputting multi-source data into the convolutional neural network model of the machine tool spindle taper hole fault, the cause of the machine tool spindle taper hole damage can be quickly determined and corresponding measures can be taken. This technology has the advantages of short repair time, high precision, and no need for special equipment in most cases. It is a fast and effective repair method for the machining center spindle taper hole damage without disassembling the spindle shaft to avoid the cost of bearing replacement.
[0156] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A machine tool spindle taper hole repair method based on multi-source data drive, It is characterized in that include: Obtain at least one possible damage cause of the taper hole of the machine tool spindle and various operating data under the corresponding damage cause, wherein the various operating data include acceleration data of the drive end, acceleration data of the motor end, acceleration data of the bearing base, time series data and motor speed; Based on various damage causes and the corresponding operating data under the damage causes, a convolutional neural network model of machine tool spindle taper hole fault is constructed; Acquire multi-source data of the machine tool to be repaired, and perform data preprocessing on the multi-source data of the machine tool to be repaired to obtain a data set to be detected; The data set to be detected is input into the convolutional neural network model of the machine tool spindle taper hole fault to obtain the cause of the damage to the machine tool spindle taper hole to be repaired; Based on the damage cause of the spindle taper hole of the machine tool to be repaired, the spindle taper hole of the machine tool is ground to obtain a repaired machine tool; Detect the rotation accuracy of the repaired machine tool spindle bearing and obtain the spindle radial runout; The axial movement of the repaired machine tool spindle bearing is measured to obtain the spindle axial movement; The radial rigidity of the repaired machine tool spindle bearing is tested to obtain the spindle radial rigidity; Determine whether the various index values of the spindle bearing meet the standards. If so, output that the machine tool spindle taper hole status is healthy. If not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other damage causes. The various index values of the spindle bearing include spindle radial runout, spindle axial runout and spindle radial rigidity.
2. According to claim 1, a machine tool spindle taper hole repair method based on multi-source data drive, It is characterized in that The construction of the machine tool spindle taper hole fault convolutional neural network model based on various damage causes and various operating data under the corresponding damage causes specifically includes: Summarize the various operating data under various damage causes into an m*n operating data matrix; Construct a 3*3 convolution kernel matrix; Split the running data matrix into at least one 3*3 convolution frame; Perform matrix multiplication on each convolution frame and convolution kernel matrix to obtain the convolution layer output matrix; Determine the activation layer function; Input the convolution layer output matrix into the activation layer function to obtain the activation layer output value; A connection is established between the output value of the activation layer and its corresponding damage cause, and a convolutional neural network model of machine tool spindle taper hole fault is obtained; The convolution kernel matrix is: Where α is the convolution kernel matrix; The activation layer function is: Where f(z) is the activation layer function and z is the eigenvalue of the convolutional layer output matrix.
3. According to claim 2, a machine tool spindle taper hole repair method based on multi-source data drive, It is characterized in that The step of acquiring multi-source data of the machine tool to be repaired and performing data preprocessing on the multi-source data of the machine tool to be repaired to obtain a data set to be detected specifically includes: Obtain multi-source data of the machine tool to be repaired; Set the weight decay speed; The weight at each moment is calculated using the weight formula based on the weight decay rate; Substituting the weights at each moment into a weighted recursive average filtering formula to obtain multi-source data of the machine tool after noise reduction; The multi-source data of the machine tool after denoising is input into the Gaussian noise function to obtain the data set to be detected; The weight formula is: oh i =e -vi , In the formula, ω i is the weight at the i-th moment, and v is the weight decay rate; The weighted recursive average filtering formula is: Y(i)=(1-ω i )Y(i-1)+ω i X(i), In the formula, Y(i) is the filtering result at the i-th moment, and X(i) is the input data at the i-th moment; The Gaussian noise function is: S ( I ) =Y ( I ) +N(μ,σ 2 ), In the formula, S ( i ) is the Gaussian noise function output result at the i-th moment, N(μ, σ 2 ) has a mean of μ and a variance of σ 2 Gaussian distributed random function.
4. According to claim 3, a machine tool spindle taper hole repair method based on multi-source data drive, It is characterized in that The method of grinding the taper hole of the spindle of the machine tool to be repaired based on the cause of the damage to the taper hole of the spindle of the machine tool to be repaired to obtain the repaired machine tool specifically includes: Determine whether the damage is caused by natural wear of the spindle taper hole. If so, use special grinding equipment to perform online mechanical grinding on the spindle taper hole. If not, no output is made. Determine whether the cause of damage is chips entering the junction between the spindle taper hole and the tool taper shank. If so, connect high-pressure gas to blow away chips, dust and cutting fluid that may be attached to the surface of the spindle taper hole. If not, no output is made. Determine whether the damage is caused by tool collision due to machining program or tool calling error. If so, repair the tapered hole by manual grinding combined with grinding without disassembling the spindle. If not, no output is made. Determine whether the cause of damage is illegal vibration of the tool caused by failure of the spring of the broach mechanism or damage to other components. If so, the complete spindle set needs to be removed from the machine tool and disassembled, all failed or damaged components on the shaft core need to be replaced, the shaft core needs to be clamped on a precision internal hole grinder, and the damaged tapered hole needs to be ground. If not, no output will be made.
5. According to claim 4, a machine tool spindle taper hole repair method based on multi-source data drive, It is characterized in that The detecting of the rotation accuracy of the repaired machine tool spindle bearing to obtain the spindle radial runout specifically includes: Remove the locating key on the spindle end face; Select the exposed outer cylindrical surface of the spindle or the undamaged tapered hole surface as the area to be measured; Clean all surfaces to be measured; Rotate the spindle more than one turn and measure the radial runout of the spindle with a micrometer.
6. A machine tool spindle taper hole repair method based on multi-source data drive according to claim 5, It is characterized in that The axial movement of the repaired machine tool spindle bearing is measured to obtain the spindle axial movement specifically including: Attach the magnetic base to the spindle housing; The micrometer head is at any position on the spindle end face; The tool switching method of loosening and tightening is adopted, and the difference of the micrometer pointer reading is observed and output as the spindle axial movement amount.
7. A machine tool spindle taper hole repair method based on multi-source data drive according to claim 6, It is characterized in that The radial rigidity of the repaired machine tool spindle bearing is detected to obtain the spindle radial rigidity, which specifically includes: Attach the magnetic base to the spindle housing; The micrometer head contacts the outer surface of the spindle shaft diameter, and the value displayed by the micrometer at this time is recorded and output as the first measurement value; The rigid measuring device is placed on the outer diameter of the spindle; The force application point is 180° away from the contact point of the micrometer. A force of 50N is applied to the rigidity measuring device. The value displayed on the micrometer at this time is recorded and output as the second measurement value. Cancel the applied force, record the value displayed on the dial gauge at this time and output it as the third measured value; Performing a difference process on the second measured value and the first measured value to obtain a first rigidity comparison value; The third measured value is subtracted from the first measured value to obtain a rigid second comparison value.
8. A machine tool spindle taper hole repair method based on multi-source data drive according to claim 7, It is characterized in that The determination of whether the various index values of the spindle bearing meet the standards specifically includes: Determine whether the spindle radial runout is less than 0.007mm. If so, no output is made. If not, the machine tool spindle taper hole status is output as unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes. Determine whether the spindle axial movement is less than 0.003mm. If so, no output is made. If not, the machine tool spindle taper hole status is unhealthy, and the machine tool spindle taper hole fault convolutional neural network model is re-input to detect other damage causes; Determine whether the first rigidity comparison value is not greater than 0.01mm and the second rigidity comparison value is 0mm. If so, output that the machine tool spindle taper hole status is healthy; if not, output that the machine tool spindle taper hole status is unhealthy, and re-input the machine tool spindle taper hole fault convolutional neural network model to detect other causes of damage.
