Precision machining process optimization method of machine tool spindle system

Abstract

The invention discloses a precision machining process optimization method for a machine tool spindle system, and relates to the technical field of machining. Comprising the steps that multi-dimensional dynamic data of the whole machining process of a machine tool spindle system are obtained, and the multi-dimensional dynamic data comprise the spindle real-time rotating speed, the cutting feeding speed, the cutting depth, the spindle radial force, the axial force, the torque, spindle box temperature field distribution, the spindle vibration frequency and the tool rear tool face abrasion width. According to the method, time-space synchronization of multi-source data is realized by adopting a timestamp alignment algorithm, feature enhancement is performed through a high-dimensional feature matrix, data time deviation and redundant information are effectively eliminated, the pertinence and precision of feature extraction are improved, and meanwhile, a deep belief network model of an attention mechanism is fused, so that the accuracy of feature extraction is improved. Through a three-layer restricted Boltzmann machine structure and contrast divergence algorithm training, key processing features can be adaptively focused.

Description

Technical Field

[0001] This invention relates to the field of machining technology, specifically to a method for optimizing the precision machining process of machine tool spindle systems. Background Technology

[0002] With the increasing demands of modern industry for machine tool processing quality and efficiency, intelligentization and automation are the development trends of the manufacturing industry. The processing technology of machine tool spindle systems also needs to adapt to this trend. Improving production efficiency and product quality is also one of the important goals pursued by enterprises.

[0003] Currently, the optimization of machine tool spindle machining processes in existing technologies largely relies on experience accumulation or trial-and-error adjustments of single parameters, which presents the following problems:

[0004] The data collection dimensions are limited, focusing mainly on basic parameters such as spindle speed and cutting feed rate, while ignoring key influencing factors such as temperature field distribution, vibration spectrum, and dynamic changes in tool wear, resulting in an incomplete basis for process optimization.

[0005] The lack of spatiotemporal synchronization and feature enhancement processing for multi-source data leads to significant time deviations and feature redundancy issues in data collected from different sensors, affecting the accuracy of subsequent optimization judgments.

[0006] The ability to extract high-dimensional processing features is limited, making it difficult to adapt to the needs of multi-objective collaborative optimization in complex processing scenarios. It lacks effective virtual simulation and closed-loop correction mechanisms. After the process parameters are optimized, they are directly applied to actual processing, which can easily lead to unstable processing accuracy due to fluctuations in working conditions. In addition, the parameter iteration cycle is long and the cost is high. Summary of the Invention

[0007] The purpose of this invention is to provide a method for optimizing the precision machining process of machine tool spindle systems in order to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the precision machining process of a machine tool spindle system, comprising the following steps:

[0009] Acquire multi-dimensional dynamic data of the entire machining process of the machine tool spindle system. The multi-dimensional dynamic data includes real-time spindle speed, cutting feed rate, cutting depth, spindle radial force, axial force, torque, spindle box temperature field distribution, spindle vibration frequency, and tool flank wear width.

[0010] The multi-dimensional dynamic data undergoes spatiotemporal synchronization processing and feature enhancement. The spatiotemporal synchronization processing is implemented using a timestamp alignment algorithm, and the feature enhancement is performed to obtain the high-dimensional processed feature matrix using the following formula:

[0011] ,in, This is the high-dimensional machining feature matrix, where n is the spindle speed. For cutting feed rate, For cutting depth, The radial force of the main shaft, T is the axial force of the main spindle, and T is the torque of the main spindle. These are characteristic values ​​of the temperature field distribution. The peak value of the vibration spectrum is represented by VB, where VB is the wear width of the tool's flank face. These are the feature weight coefficients, and ;

[0012] Based on the aforementioned high-dimensional processing feature matrix, a deep belief network process optimization judgment model with fused attention mechanism is constructed. The model makes optimization judgments using the following formula:

[0013] Where O is the optimization judgment result vector. Let W be the activation function and W be the weight matrix. Let b be the attention mechanism function, and b be the bias term.

[0014] The optimal combination of process parameters is output based on the optimization judgment result vector. The optimal combination of process parameters satisfies the optimization objectives of machining accuracy error ≤ 0.0008 mm, surface roughness Ra ≤ 0.4 μm, and machining energy consumption reduction ≥ 15%.

[0015] As a specific embodiment of the technical solution of this application, the timestamp alignment algorithm for spatiotemporal synchronization processing corrects time deviations using the following formula:

[0016] ,in, To synchronize and correct the time, For the timestamp collected by the sensor, K is the timestamp of the master control unit, and K is the synchronization coefficient. This is the transmission delay compensation value.

[0017] As a specific embodiment of the technical solution in this application, the training process of the deep belief network process optimization judgment model in the fusion attention mechanism includes:

[0018] A network architecture consisting of an input layer, three restricted Boltzmann machine hidden layers, and an output layer is constructed, with the input layer dimension consistent with the dimension of the high-dimensional processing feature matrix.

[0019] The network is pre-trained using a contrastive divergence algorithm, and the reconstruction error is calculated using the following formula:

[0020] ,in, The reconstruction error is N, where N is the number of samples. For the input sample, To reconstruct the sample;

[0021] An attention weight matrix is ​​introduced to enhance the training of key features until the reconstruction error is less than a preset threshold of 0.001.

[0022] As a specific embodiment of the technical solution of this application, the optimal combination of process parameters includes the optimized spindle speed range, the optimized cutting feed rate, the optimized cutting fluid flow rate, and the optimized tool durability threshold. The optimized tool durability threshold is calculated using the following formula: ,in, The value represents the tool durability, where C is a material constant, m, n, and p are exponential coefficients, and k is the temperature effect coefficient. The average temperature of the main spindle.

[0023] As a specific solution of the technical solution of this application, it also includes constructing a digital twin model of the processing technology, inputting the optimal combination of process parameters into the digital twin model for virtual processing simulation, and obtaining a cloud map of the simulation processing error distribution.

[0024] As a specific embodiment of the technical solution of this application, the virtual machining simulation is achieved through a finite element analysis algorithm, and the simulation error is calculated using the following formula: ,in, The total simulation error is... For geometric error weights, Geometric machining error, This represents the error caused by thermal deformation.

[0025] As a specific solution of the technical solution of this application, it also includes returning to retrain the feature enhancer model if the maximum error value in the simulation processing error distribution cloud map exceeds a preset simulation error threshold, wherein the preset simulation error threshold is 0.0006mm.

[0026] As a specific solution of the technical solution of this application, it also includes sending the optimal combination of process parameters verified by simulation to the machine tool PLC control system, and collecting the dynamic accuracy data of the spindle in real time during the machining process. The dynamic accuracy data of the spindle includes the instantaneous speed fluctuation of the spindle, the thermal drift of the spindle axis, and the time domain peak value of the cutting force.

[0027] As a specific solution to the technical solution of this application, it also includes closed-loop correction based on the spindle dynamic accuracy data and a preset dynamic accuracy standard. The preset dynamic accuracy standard is: rotational speed fluctuation ≤ ±0.5%, axial thermal drift ≤ 0.0004 mm / h, and cutting force peak fluctuation ≤ ±0.3%. The correction coefficient is calculated using the following formula:

[0028] ,in, For correction factor, This is the proportionality coefficient. This is the measurement error. This is for annotation error.

[0029] As a specific solution of the technical solution of this application, the closed-loop correction is achieved by dynamically adjusting the spindle drive current and the cutting fluid injection angle, and the adjustment range is linearly proportional to the correction coefficient.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] The precision machining process optimization method for machine tool spindle systems adopts a timestamp alignment algorithm to achieve spatiotemporal synchronization of multi-source data and uses a high-dimensional feature matrix for feature enhancement, which effectively eliminates data time deviation and redundant information, and improves the pertinence and accuracy of feature extraction.

[0032] Meanwhile, the deep belief network model that integrates the attention mechanism, trained through a 3-layer restricted Boltzmann machine structure and contrastive divergence algorithm, can adaptively focus on key processing features and control the reconstruction error within 0.001. Compared with traditional algorithms, it has stronger high-dimensional feature learning ability and optimization judgment accuracy, and can simultaneously meet the multi-objective optimization requirements of processing accuracy, surface roughness and energy consumption reduction.

[0033] By using digital twin models for virtual processing simulation, error distribution cloud maps can be obtained in advance, allowing potential problems to be identified before actual processing. This avoids material waste and efficiency losses caused by traditional trial-and-error methods, and reduces process optimization costs. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the precision machining process optimization method for the machine tool spindle system of the present invention;

[0035] Figure 2 This is a schematic diagram of the system framework of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0037] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0038] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0039] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0040] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediate medium. Furthermore, "above," "over," and "on top" of the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "above," and "below" of the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0041] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0042] like Figures 1-2 As shown, the present invention provides a technical solution: a method for optimizing the precision machining process of a machine tool spindle system, comprising the following steps:

[0043] S1: Acquire multi-dimensional dynamic data of the entire machining process of the machine tool spindle system. The multi-dimensional dynamic data includes real-time spindle speed, cutting feed rate, cutting depth, spindle radial force, axial force, torque, spindle box temperature field distribution, spindle vibration frequency, and tool flank wear width.

[0044] In one embodiment of this application, to ensure the accuracy and comprehensiveness of the acquired data, multiple high-precision sensors are used for data acquisition. Specifically, the real-time spindle speed is acquired using a non-contact laser speed sensor mounted on the spindle end. This sensor has a measurement range of 0-10000 r / min and a measurement accuracy of ±0.1 r / min. The cutting feed rate and depth of cut are acquired using a grating ruler integrated into the machine tool. The grating ruler has a resolution of 0.1 μm, ensuring high accuracy of the feed rate and depth of cut data. The spindle radial force, axial force, and torque are acquired using a triaxial force sensor mounted on the spindle bearing housing. The sensor's measurement range is 0-5000N, with a measurement accuracy of ±0.5%FS. The spindle box temperature field distribution is measured using a distributed fiber optic temperature sensor with a measurement point spacing of 5mm, a measurement range of -50℃ to 200℃, and a measurement accuracy of ±0.2℃. The spindle vibration frequency is acquired by an accelerometer mounted on the spindle box, with a measurement frequency range of 0-10kHz and a measurement accuracy of ±1%. The tool flank wear width is obtained through online detection using an industrial camera with a resolution of 12 megapixels. Combined with an image recognition algorithm, the measurement accuracy can reach 0.001mm.

[0045] It should be clear that in this application, all data collected by the sensors is transmitted to the data processing center in real time via a data acquisition card. The sampling frequency of the data acquisition card is no less than 1kHz to ensure that dynamic changes during the processing can be captured.

[0046] S2: Perform spatiotemporal synchronization processing and feature enhancement on the multi-dimensional dynamic data. The spatiotemporal synchronization processing is achieved through a timestamp alignment algorithm, and the feature enhancement is calculated using the following formula to obtain the high-dimensional processed feature matrix:

[0047] ,in, This is the high-dimensional machining feature matrix, where n is the spindle speed. For cutting feed rate, For cutting depth, The radial force of the main shaft, T is the axial force of the main spindle, and T is the torque of the main spindle. These are characteristic values ​​of the temperature field distribution. The peak value of the vibration spectrum is represented by VB, where VB is the wear width of the tool's flank face. These are the feature weight coefficients, and ;

[0048] In this embodiment, the timestamp alignment algorithm for spatiotemporal synchronization processing corrects time deviations using the following formula:

[0049] ,in, To synchronize and correct the time, For the timestamp collected by the sensor, The timestamp is the main control unit timestamp, and K is the synchronization coefficient. In this application, the value ranges from 0 to 1 and can be adjusted according to the communication delay characteristics between the sensor and the main control unit. The transmission delay compensation value was calculated by conducting multiple tests on the communication link between the sensor and the data processing center, and taking the average transmission delay. The initial value can be dynamically corrected based on the real-time monitored transmission delay during actual operation.

[0050] It should be clear that, in one embodiment of this application, the feature weight coefficients... The determination of the weight coefficients adopts the analytic hierarchy process (AHP). By constructing a judgment matrix, the importance of each feature parameter in the processing is quantitatively evaluated, thereby determining its weight coefficient. For feature parameters that have a greater impact on processing accuracy, such as spindle radial force and temperature field distribution characteristics, larger weight coefficients are assigned, while for relatively minor feature parameters, smaller weight coefficients are assigned. In practical applications, the feature weight coefficients can be dynamically adjusted according to different processing materials and processing requirements.

[0051] S3: Based on the aforementioned high-dimensional processing feature matrix, a deep belief network process optimization judgment model with an attention fusion mechanism is constructed. The model performs optimization judgment using the following formula:

[0052] Where O is the optimization judgment result vector. Let W be the activation function and W be the weight matrix. Let b be the attention mechanism function, and b be the bias term.

[0053] In one embodiment of this application, the deep belief network process for optimizing the judgment model training includes:

[0054] A network architecture consisting of an input layer, three restricted Boltzmann machine (RBM) hidden layers, and an output layer is constructed. The dimension of the input layer is consistent with the dimension of the high-dimensional processing feature matrix. The number of neurons in each RBM hidden layer is set according to the dimension of the input layer. The number of neurons in the first hidden layer is 1.5 times the dimension of the input layer, the number of neurons in the second hidden layer is 0.8 times that of the first hidden layer, and the number of neurons in the third hidden layer is 0.6 times that of the second hidden layer. The number of neurons in the output layer is consistent with the number of optimization target parameters.

[0055] The network is pre-trained using a contrastive divergence algorithm, and the reconstruction error is calculated using the following formula:

[0056] ,in, The reconstruction error is N, where N is the number of samples. For the input sample, To reconstruct the sample;

[0057] An attention weight matrix is ​​introduced to enhance the training of key features. The attention weight matrix is ​​dynamically adjusted according to the importance of each feature in the processing, and key features are given higher weights. During the training process, the network parameters and attention weight matrix are continuously optimized through the backpropagation algorithm until the reconstruction error is less than the preset threshold of 0.001.

[0058] S4: Output the optimal combination of process parameters based on the optimization judgment result vector. The optimal combination of process parameters satisfies the optimization objectives of machining accuracy error ≤ 0.0008 mm, surface roughness Ra ≤ 0.4 μm, and machining energy consumption reduction ≥ 15%.

[0059] The optimal combination of process parameters includes the optimized spindle speed range, optimized cutting feed rate, optimized cutting fluid flow rate, and optimized tool durability threshold. The optimized tool durability threshold is calculated using the following formula: ,in, For tool durability, C is a material constant, selected according to the material being machined. For example, when machining 45 steel, C is set to 800. m, n, and p are exponential coefficients, with values ​​of 0.2, 0.3, and 0.1 respectively. k is the temperature influence coefficient, with a value of 1.02. The average temperature of the main spindle.

[0060] In one embodiment of this application, the process further includes constructing a digital twin model of the processing technology, inputting the optimal combination of process parameters into the digital twin model for virtual processing simulation, and obtaining a cloud map of the simulation processing error distribution.

[0061] In one embodiment of this application, the digital twin model of the machining process is constructed based on three-dimensional modeling software to accurately reproduce the structure and machining environment of the machine tool spindle system. The model includes the geometric model and physical property parameters of key components such as spindle, tool, and workpiece, such as the elastic modulus, Poisson's ratio, and coefficient of thermal expansion of the material.

[0062] Virtual machining simulation is achieved through finite element analysis algorithm, and the simulation error is calculated using the following formula: ,in, The total simulation error is... This is the geometric error weight, with a value ranging from 0 to 1. It is adjusted according to the influence of geometric errors and thermal deformation errors during the processing. In this application, The geometric error weight is set to 0.6. Geometric machining errors are calculated by examining the relative movement trajectories of the tool and the workpiece. The thermal deformation error is calculated based on the temperature field distribution data to determine the thermal deformation of the spindle and the workpiece.

[0063] In one embodiment of this application, if the maximum error value in the simulated processing error distribution cloud map exceeds a preset simulation error threshold, then the feature enhancer model training is restarted, wherein the preset simulation error threshold is 0.0006 mm.

[0064] In one embodiment of this application, when performing feature enhancement again, the feature weight coefficients are adjusted to give greater weight to the feature parameters that cause the error to exceed the standard. During model training, the number of training samples is increased accordingly, and the network structure and training parameters are optimized to improve the prediction accuracy of the model.

[0065] In one embodiment of this application, the optimal combination of process parameters verified through simulation is sent to the machine tool PLC control system to collect the dynamic accuracy data of the spindle during the machining process in real time. The dynamic accuracy data of the spindle includes the instantaneous spindle speed fluctuation, the thermal drift of the spindle axis, and the time-domain peak value of the cutting force. The instantaneous spindle speed fluctuation is collected in real time by a laser speed sensor with a sampling frequency of 10kHz, and is obtained by calculating the speed difference between adjacent sampling times. The thermal drift of the spindle axis is measured by a laser interferometer installed at the end of the spindle with a measurement accuracy of 0.01µm. The time-domain peak value of the cutting force is obtained by analyzing the data collected by a triaxial force sensor, which collects 1000 data points per second and takes the maximum value as the time-domain peak value of the cutting force.

[0066] In one embodiment of this application, a closed-loop correction is further performed based on the spindle dynamic accuracy data and a preset dynamic accuracy standard. The preset dynamic accuracy standard is: rotational speed fluctuation ≤ ±0.5%, axial thermal drift ≤ 0.0004 mm / h, and peak cutting force fluctuation ≤ ±0.3%. The correction coefficient is calculated using the following formula:

[0067] ,in, For correction factor, This is a proportionality coefficient, ranging from 0 to 1, which is adjusted based on the stability of the processing. In this application, The proportionality coefficient is set to 0.5. This is the measurement error. This is for annotation error.

[0068] Closed-loop correction is achieved by dynamically adjusting the spindle drive current and the coolant injection angle, with the adjustment range linearly proportional to the correction coefficient. When the measured error is greater than the standard error, the correction coefficient is increased, correspondingly increasing the adjustment of the spindle drive current and the coolant injection angle. When the measured error is less than the standard error, the correction coefficient is decreased, reducing the adjustment amount to ensure the stability and accuracy of the machining process.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A method for optimizing the precision machining process of a machine tool spindle system, characterized in that, Includes the following steps: Acquire multi-dimensional dynamic data of the entire machining process of the machine tool spindle system. The multi-dimensional dynamic data includes real-time spindle speed, cutting feed rate, cutting depth, spindle radial force, axial force, torque, spindle box temperature field distribution, spindle vibration frequency, and tool flank wear width. The multi-dimensional dynamic data undergoes spatiotemporal synchronization processing and feature enhancement. The spatiotemporal synchronization processing is implemented using a timestamp alignment algorithm, and the feature enhancement is performed to obtain the high-dimensional processed feature matrix using the following formula: ,in, This is the high-dimensional machining feature matrix, where n is the spindle speed. For cutting feed rate, For cutting depth, The radial force of the main shaft, T is the axial force of the main spindle, and T is the torque of the main spindle. These are characteristic values ​​of the temperature field distribution. The peak value of the vibration spectrum is represented by VB, where VB is the wear width of the tool's flank face. These are the feature weight coefficients, and ; Based on the aforementioned high-dimensional processing feature matrix, a deep belief network process optimization judgment model with fused attention mechanism is constructed. The model makes optimization judgments using the following formula: Where O is the optimization judgment result vector. Let W be the activation function and W be the weight matrix. Let b be the attention mechanism function, and b be the bias term. The optimal combination of process parameters is output based on the optimization judgment result vector. The optimal combination of process parameters satisfies the optimization objectives of machining accuracy error ≤ 0.0008 mm, surface roughness Ra ≤ 0.4 μm, and machining energy consumption reduction ≥ 15%.

2. The method for optimizing the precision machining process of a machine tool spindle system according to claim 1, characterized in that: The timestamp alignment algorithm for the spatiotemporal synchronization process corrects for time deviations using the following formula: ,in, To synchronize and correct the time, For the timestamp collected by the sensor, K is the timestamp of the master control unit, and K is the synchronization coefficient. This is the transmission delay compensation value.

3. The method for optimizing the precision machining process of a machine tool spindle system according to claim 1, characterized in that: The deep belief network process for optimizing the judgment model training in the fusion attention mechanism includes: A network architecture consisting of an input layer, three restricted Boltzmann machine hidden layers, and an output layer is constructed, with the input layer dimension consistent with the dimension of the high-dimensional processing feature matrix. The network is pre-trained using a contrastive divergence algorithm, and the reconstruction error is calculated using the following formula: ,in, The reconstruction error is N, where N is the number of samples. For the input sample, To reconstruct the sample; An attention weight matrix is ​​introduced to enhance the training of key features until the reconstruction error is less than a preset threshold of 0.

001.

4. The method for optimizing the precision machining process of a machine tool spindle system according to claim 1, characterized in that: The optimal combination of process parameters includes the optimized spindle speed range, optimized cutting feed rate, optimized cutting fluid flow rate, and optimized tool durability threshold. The optimized tool durability threshold is calculated using the following formula: ,in, The value represents the tool durability, where C is a material constant, m, n, and p are exponential coefficients, and k is the temperature effect coefficient. The average temperature of the main spindle.

5. The method for optimizing the precision machining process of a machine tool spindle system according to claim 1, characterized in that: It also includes constructing a digital twin model of the processing technology, inputting the optimal combination of process parameters into the digital twin model for virtual processing simulation, and obtaining a cloud map of the simulation processing error distribution.

6. The method for optimizing the precision machining process of a machine tool spindle system according to claim 5, characterized in that: The virtual machining simulation is achieved through finite element analysis algorithm, and the simulation error is calculated using the following formula: ,in, The total simulation error is... For geometric error weights, Geometric machining error, This represents the error caused by thermal deformation.

7. The method for optimizing the precision machining process of a machine tool spindle system according to claim 5, characterized in that: It also includes returning to retrain the feature enhancer model if the maximum error value in the simulated processing error distribution cloud map exceeds a preset simulation error threshold, wherein the preset simulation error threshold is 0.0006 mm.

8. The method for optimizing the precision machining process of a machine tool spindle system according to claim 5, characterized in that: It also includes sending the optimal combination of process parameters verified by simulation to the machine tool PLC control system, and collecting the dynamic accuracy data of the spindle in real time during the machining process. The dynamic accuracy data of the spindle includes the instantaneous speed fluctuation of the spindle, the thermal drift of the spindle axis, and the time domain peak value of the cutting force.

9. The method for optimizing the precision machining process of a machine tool spindle system according to claim 1, characterized in that: It also includes closed-loop correction based on the spindle dynamic accuracy data and a preset dynamic accuracy standard. The preset dynamic accuracy standard is: rotational speed fluctuation ≤ ±0.5%, axis thermal drift ≤ 0.0004 mm / h, and peak cutting force fluctuation ≤ ±0.3%. The correction coefficient is calculated using the following formula: ,in, For correction factor, This is the proportionality coefficient. This is the measurement error. This is for annotation error.

10. The method for optimizing the precision machining process of a machine tool spindle system according to claim 9, characterized in that: The closed-loop correction is achieved by dynamically adjusting the spindle drive current and the cutting fluid injection angle, and the adjustment range is linearly proportional to the correction coefficient.

Images