A high-precision CNC machine tool machining compensation adjustment method

Abstract

The present invention relates to the field of machine tool detection, and more specifically, to a method for high-precision CNC machine tool processing compensation adjustment. The method comprises: obtaining multi-dimensional machine tool operating parameters at each historical sampling time point; training a preset prediction network model to obtain a machine tool detection model; constructing an operating parameter sequence, and calculating the entropy value of a target historical sampling time point in a target dimension through the parameter value in the operating parameter sequence; presetting a judgment condition, and calculating the learning rate correction weight of the target historical sampling time point; updating the learning rate of the machine tool detection model, and obtaining a machine tool detection correction model, and inputting the machine tool operating parameters at the real-time sampling time point into the machine tool detection correction model to obtain the machine tool operating status at the real-time sampling time point. Through the technical solution of the present invention, the timeliness and accuracy of machine tool operating status detection and tracking can be improved.

Description

Technical Field

[0001] The present invention relates to the field of machine tool detection, and more specifically, to a high-precision CNC machine tool processing compensation adjustment method. Background Art

[0002] In modern manufacturing, high-precision machine tools are the core equipment for precision machining, and their operating status directly affects the machining accuracy and production efficiency of products. However, as the machine tools run for a long time and the working environment changes, various parts of the machine tools may be worn, loose, or have performance degradation. If these problems are not discovered and handled in time, they will lead to reduced machining accuracy, reduced production efficiency, and even cause serious machine tool failures, resulting in huge economic losses.

[0003] At present, the patent application document with the publication number CN114662856A discloses a large-scale rolling mill state tracking method and system, which includes obtaining rolling mill state data and saving it to the server under the premise of ensuring data consistency; reducing the dimension and normalizing the classified and labeled rolling mill state data; extracting physical signs after normalization to obtain a feature matrix, and generating a sample set of data of different dimensions in the feature matrix, and then correlating the sample set through data mining technology, analyzing and obtaining a differentiated state analysis model library of the rolling mill equipment under different working conditions, and evaluating the optimal analysis model; comparing the feature matrix with the optimal analysis model to evaluate the health status of the rolling mill equipment; and sharing resources and managing the whole life cycle of the rolling mill equipment in a healthy state after evaluation. When this method obtains the rolling mill state data, there are phenomena such as data delay and missing, which easily lead to the difficulty of reference of the tracking results of the equipment operation status, and thus it is difficult to adjust its operation parameters in time. Summary of the invention

[0004] In order to solve the above-mentioned technical problem that the state of the equipment cannot be adjusted timely and accurately, the present invention proposes a high-precision CNC machine tool processing compensation adjustment method.

[0005] The present invention discloses a high-precision CNC machine tool processing compensation adjustment method, comprising the following steps: obtaining multi-dimensional machine tool operating parameters at each historical sampling time point, taking any dimension of the machine tool operating parameters as a target dimension; training a preset prediction network model according to the machine tool operating parameters obtained at the historical sampling time point to obtain a machine tool detection model, taking any historical sampling time point as a target historical sampling time point; constructing the machine tool operating parameters of multiple historical sampling time points adjacent to the target historical sampling time point in the target dimension and the machine tool operating parameters at the target historical sampling time point into an operating parameter sequence, and calculating the entropy value of the target historical sampling time point in the target dimension through the parameter values ​​in the operating parameter sequence; in response to the entropy value satisfying a preset judgment condition, accumulating the difference between the entropy value of the target historical sampling time point and the previous historical sampling time point adjacent to the target historical sampling time point in each dimension, and calculating the learning rate correction weight of the target historical sampling time point, and the learning rate correction weight satisfies the relationship:

[0006] ,in, Indicates The learning rate correction weight at each historical sampling time point, Indicates The historical sampling time point is The entropy value in the dimension is Indicates The historical sampling time point is The entropy value in the dimension is Represents the total number of dimensions, Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, represents the learning rate correction weight, Represents the normalization function. In the learning rate correction weight relationship, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension;

[0007] The learning rate of the machine tool detection model is updated according to the learning rate correction weight to obtain the machine tool detection correction model, and the machine tool operation parameters at the real-time sampling time point are input into the machine tool detection correction model to obtain the machine tool operation status at the real-time sampling time point to adjust the machine tool.

[0008] By adopting the above technical solution, the entropy value of each historical sampling time point in different dimensions is calculated, the judgment conditions are set, and the learning rate in the machine tool detection model corresponding to the historical sampling time points that meet the judgment conditions is updated to obtain the machine tool detection correction model. Then, the real-time sampling time points are input into the machine tool detection correction model to obtain the machine tool operating status corresponding to the real-time sampling time points, so that the CNC machine tools can be adjusted in a timely and accurate manner.

[0009] Preferably, the prediction network model is a BP neural network, which includes an input layer, a hidden layer and an output layer; the input layer is connected to the hidden layer, and the hidden layer is connected to the output layer; after the multi-dimensional machine tool operating parameters are input from the input layer, the input is feature extracted through the hidden layer, and the extracted features are output through the output layer to generate the machine tool operating status.

[0010] Preferably, the training preset prediction network model includes: obtaining multi-dimensional machine tool operating parameters at any historical sampling time point as input information, and taking the machine tool operating status at the next historical sampling time point adjacent to any historical sampling time point as label information to obtain a set of training data; inputting the input information in the training data into the prediction network model to obtain an output result; using a mean square error loss function to calculate the loss value between the output result and the label information, using the gradient descent method to back-propagate the error signal according to the loss value, and updating the weights and biases in the network according to the gradient of the network parameters and the learning rate to reduce the loss value; iteratively updating the network parameters of the prediction network model until the loss value is less than the set loss value or the maximum number of training times is reached, stopping the update, and obtaining a machine tool detection model.

[0011] At the same time, iterative training of the prediction network model can realize intelligent detection of the machine tool status, thereby facilitating real-time adjustment of the machine tool's operating parameters.

[0012] Preferably, the entropy value satisfies the polynomial:

[0013] , ,in, Indicates The historical sampling time point is The parameter values ​​obtained in the dimensions, Indicates The historical sampling time point is The probability of the parameter value obtained in the dimension appearing in the running parameter sequence, Indicates the total number of machine tool operating parameters in the operating parameter sequence. Indicates The historical sampling time point is The entropy value in the dimension is Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, Indicates the parameter value. represents the probability, represents the logarithmic function. In the entropy polynomial, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension.

[0014] In this way, the entropy value of the target historical sampling time point in the target dimension can be calculated by running the parameter values ​​in the parameter sequence, thereby realizing a quantitative evaluation of the target historical sampling time point.

[0015] Preferably, the preset judgment condition includes: comparing the entropy value of the target historical sampling time point in the target dimension with the preset entropy value threshold of the target dimension, and updating the machine tool detection model in response to the entropy value of the target historical sampling time point in the target dimension being greater than the entropy value threshold.

[0016] Preferably, the preset judgment condition also includes: constructing the entropy value of the target historical sampling time point in all dimensions into an entropy value vector; calculating the distance value of the Euclidean distance between the entropy value vector of the target historical sampling time point and the entropy value vector of the previous historical sampling time point adjacent to the target historical sampling time point; comparing the distance value with a preset distance threshold, and updating the machine tool detection model in response to the distance value being greater than the preset distance threshold.

[0017] Preferably, the learning rate of the updated machine tool detection model satisfies the relationship:

[0018] ,in, Indicates The learning rate of the machine tool detection model corresponding to each historical sampling time point, Indicates The learning rate correction weight at each historical sampling time point, Indicates The learning rate of the machine tool detection model corresponding to each historical sampling time point, Indicates the sequence number of the historical sampling time point, represents the learning rate, Represents the learning rate correction weight. In the learning rate relationship of updating the machine tool detection model, The historical sampling time point is the target historical sampling time point.

[0019] By updating the learning rate of the machine tool detection model, the accuracy of machine tool operation status detection can be further improved, and real-time adjustment of the machine tool operation status can be achieved.

[0020] The present invention has the following technical effects:

[0021] 1. The present invention can improve the timeliness and accuracy of tracking and detecting the operating status of a machine tool.

[0022] 2. The present invention obtains the machine tool operation parameters of multiple dimensions at each historical sampling time point, constructs and trains a preset prediction network model through the machine tool operation parameters of multiple dimensions at the historical sampling time point, and obtains the machine tool detection model. The entropy value of each historical sampling time point in different dimensions is calculated, and the judgment condition is set. The learning rate in the machine tool detection model corresponding to the historical sampling time point that meets the judgment condition is updated to obtain the machine tool detection correction model, and then the real-time sampling time point is input into the machine tool detection correction model to obtain the machine tool operation status corresponding to the real-time sampling time point, so that the CNC machine tool can be adjusted in a timely and accurate manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0024] Figure 1 The present invention is a flowchart of a high-precision CNC machine tool processing compensation adjustment method according to an embodiment of the present invention.

[0025] Figure 2 It is a schematic diagram of a high-precision CNC machine tool machining compensation adjustment system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0027] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0028] The present invention provides a high-precision CNC machine tool processing compensation adjustment method. Figure 1As shown, a high-precision CNC machine tool processing compensation adjustment method includes steps S1 to S5, which are described in detail below.

[0029] S1, obtain the multi-dimensional machine tool operation parameters at each historical sampling time point.

[0030] In one embodiment, the multi-dimensional machine tool operating parameters at each historical sampling time point are obtained according to a preset time interval. Exemplarily, the preset time interval can be set by a person skilled in the art. The present invention sets the preset time interval to 1 second.

[0031] It should be noted that the multi-dimensional machine tool operating parameters are machine parameters of the machine tool when it is working, including but not limited to rolling force, rolling torque, rolling speed, rolling temperature, and machine tool vibration value. The method for obtaining machine tool operating parameters includes: installing a pressure sensor on the machine tool roller to obtain the rolling force at a preset time interval; installing a torque sensor at the machine tool bearing to obtain the rolling torque at a preset time interval; installing a speed sensor on the machine tool transmission device to obtain the rolling speed at a preset time interval; installing an infrared thermometer on the surface of the machine tool roller to obtain the rolling temperature at a preset time interval; installing a vibration sensor on the machine tool base to obtain the machine tool vibration value at a preset time interval. The machine tool operating parameters of different dimensions obtained by the present invention are to detect the data differences at the same historical sampling time points in different dimensions and the differences at different historical sampling time points in the same dimension. Normalization will be performed during the calculation process to eliminate the differences between different dimensions.

[0032] S2, train the preset prediction network model to obtain the machine tool detection model.

[0033] In one embodiment, a preset prediction network model is trained based on the machine tool operation parameters obtained at the historical sampling time point. The prediction network model is a BP neural network. The BP neural network includes an input layer, a hidden layer and an output layer; the input layer is connected to the hidden layer, and the hidden layer is connected to the output layer; after the multi-dimensional machine tool operation parameters are input from the input layer, the input is subjected to feature extraction through the hidden layer, and the extracted features are output through the output layer to generate the machine tool operation status. In the actual application process, the output is the probability of normal operation of the machine tool and the probability of abnormal operation of the machine tool. Generally speaking, when the probability of normal operation of the machine tool is less than the probability of abnormal operation of the machine tool, it means that the machine tool is abnormal. However, usually due to the aging or error range of the machine itself, the probability of normal operation of the machine tool output is usually less than 0.5. For example, the present invention is set to be less than 0.4 when the probability of normal operation of the machine tool is less than 0.4, which means that the machine tool is abnormal.

[0034] Among them, the training process of the preset prediction network model includes: obtaining multi-dimensional machine tool operating parameters at any historical sampling time point as input information, and taking the machine tool operating status at the next historical sampling time point adjacent to any historical sampling time point as label information to obtain a set of training data; inputting the input information in the training data into the prediction network model to obtain the output result; using the mean square error loss function to calculate the loss value between the output result and the label information, using the gradient descent method to backpropagate the error signal according to the loss value, and updating the weights and biases in the network according to the gradient of the network parameters and the learning rate to reduce the loss value; iteratively updating the network parameters of the prediction network model until the loss value is less than the set loss value or the maximum number of training times is reached, stopping the update, and obtaining the machine tool detection model.

[0035] At this point, the machine tool detection model is obtained.

[0036] S3, construct an operation parameter sequence, and calculate the entropy value of the target historical sampling time point in the target dimension through the parameter values ​​in the operation parameter sequence.

[0037] In one embodiment, the machine tool operating parameters of multiple historical sampling time points adjacent to the target historical sampling time point in the target dimension and the machine tool operating parameters of the target historical sampling time point are constructed into an operating parameter sequence. For example, the machine tool operating parameters of 19 historical sampling time points are obtained in the target dimension, and the target historical sampling time point is recorded as , then the 19 historical sampling time points are recorded as , the constructed running parameter sequence is The entropy value of the target historical sampling time point in the target dimension is calculated by running the parameter values ​​in the parameter sequence. The entropy value satisfies the polynomial:

[0038] , ,in, Indicates The historical sampling time point is The parameter values ​​obtained in the dimensions, Indicates The historical sampling time point is The probability of the parameter value obtained in the dimension appearing in the running parameter sequence, Indicates the total number of machine tool operating parameters in the operating parameter sequence. Indicates The historical sampling time point is The entropy value in the dimension is Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, Indicates the parameter value. represents the probability, represents the logarithmic function. In the entropy polynomial, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension.

[0039] At this point, the entropy value of the target historical sampling time point in the target dimension is obtained.

[0040] S4, preset judgment conditions, calculate the learning rate correction weight of the target historical sampling time point.

[0041] In one embodiment, according to the entropy value calculated in step S3, the preset judgment condition includes: comparing the entropy value of the target historical sampling time point in the target dimension with the preset entropy value threshold of the target dimension, in response to the entropy value of the target historical sampling time point in the target dimension being greater than the entropy value threshold, updating the machine tool detection model, it should be noted that each dimension will have a unique entropy value threshold. The preset judgment condition also includes: constructing the entropy value of the target historical sampling time point in all dimensions as an entropy value vector; calculating the distance value of the Euclidean distance between the entropy value vector of the target historical sampling time point and the entropy value vector of the previous historical sampling time point adjacent to the target historical sampling time point; comparing the distance value with the preset distance threshold, in response to the distance value being greater than the preset distance threshold, updating the machine tool detection model.

[0042] In response to the entropy value satisfying the preset judgment condition, the learning rate correction weight of the target historical sampling time point is calculated by accumulating the difference between the entropy value of the target historical sampling time point and the previous historical sampling time point adjacent to the target historical sampling time point in each dimension, and the learning rate correction weight satisfies the relationship:

[0043] ,in, Indicates The learning rate correction weight at each historical sampling time point, Indicates The historical sampling time point is The entropy value in the dimension is Indicates The historical sampling time point is The entropy value in the dimension is Represents the total number of dimensions, Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, Represents the learning rate correction weight. In the learning rate correction weight relationship, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension. Represents the normalization function.

[0044] At this point, the learning rate correction weight at the target historical sampling time point is obtained.

[0045] S5, updating the learning rate of the machine tool detection model, obtaining the machine tool detection correction model, and inputting the machine tool operation parameters at the real-time sampling time point into the machine tool detection correction model to obtain the machine tool operation status at the real-time sampling time point.

[0046] In one embodiment, the learning rate correction weight obtained in step S4 is used to update the learning rate of the machine tool detection model, and the learning rate of the updated machine tool detection model satisfies the relationship:

[0047] ,in, Indicates The learning rate of the machine tool detection model corresponding to each historical sampling time point, Indicates The learning rate correction weight at each historical sampling time point, Indicates The learning rate of the machine tool detection model corresponding to each historical sampling time point, Indicates the sequence number of the historical sampling time point, represents the learning rate, Represents the learning rate correction weight. In the learning rate relationship of updating the machine tool detection model, The historical sampling time point is the target historical sampling time point.

[0048] The updated learning rate of the machine tool detection model is used as the learning rate of the machine tool detection correction model, and the machine tool operation parameters at the real-time sampling time point are input into the machine tool detection correction model to obtain the machine tool operation status at the real-time sampling time point.

[0049] Exemplarily, based on the multiple continuous historical sampling time points described in step S3, the first The machine tool detection model at the historical sampling time point and the The entropy value of each dimension at each historical sampling time point is The entropy value of each dimension at each historical sampling time point meets the preset judgment condition, and the calculation The learning rate correction weight of the historical sampling time point is The learning rate in the machine tool detection model at the historical sampling time point is updated to obtain the machine tool detection correction model, and the machine tool detection correction model is used as the first The machine tool detection model at each historical sampling time point will collect the The multi-dimensional machine tool operation parameters at the historical sampling time point are input into the The machine tool detection model at the historical sampling time point is obtained The machine tool operating status at a historical sampling time point can be tracked and detected in real time.

[0050] It should be noted that there are two situations for adjusting the learning rate. In the first situation, the cumulative value of the entropy value of two adjacent historical sampling time points in each dimension is small. At this time, the learning rate is reduced to ensure that the machine tool detection model can adjust the parameters in detail so that it can converge more stably and accurately when approaching the optimal solution. In the second situation, the cumulative value of the entropy value of two adjacent historical sampling time points in each dimension is large, and because the machine tool state has changed significantly, the original machine tool detection model is no longer applicable. At this time, the machine tool detection model needs to be able to quickly adjust itself to match the new machine tool operating parameters. At this time, increasing the learning rate can enable the model to respond quickly to state changes, so that the model can take bigger steps in the process of gradient descent, accelerate the convergence process, and help the model find a new optimal solution faster, thereby realizing intelligent compensation adjustment for CNC machine tool processing.

[0051] In an optional embodiment, the present invention further discloses a high-precision CNC machine tool processing compensation adjustment system, referring to Figure 2 , including a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a high-precision CNC machine tool processing compensation adjustment method according to the present invention is implemented.

[0052] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0053] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory (RRAM), a dynamic random access memory (DRAM), a static random access memory (SRAM), a high-bandwidth memory (HBM), a hybrid memory cube (HMC), etc., or any other medium that may be used to store the required information and may be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or may be accessible or connectable to a device.

[0054] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0055] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A high-precision CNC machine tool processing compensation adjustment method, characterized in that: The following steps are involved: Obtain multi-dimensional machine tool operating parameters at each historical sampling time point, and use any dimension of the machine tool operating parameters as the target dimension; The preset prediction network model is trained according to the machine tool operation parameters obtained at the historical sampling time points to obtain the machine tool detection model, and any historical sampling time point is used as the target historical sampling time point; The machine tool operating parameters of multiple historical sampling time points adjacent to the target historical sampling time point in the target dimension and the machine tool operating parameters of the target historical sampling time point are constructed into an operating parameter sequence, and the entropy value of the target historical sampling time point in the target dimension is calculated by the parameter values ​​in the operating parameter sequence; Preset judgment conditions include: Construct the entropy values ​​of the target historical sampling time points in all dimensions into an entropy value vector; Calculate the Euclidean distance between the entropy value vector of the target historical sampling time point and the entropy value vector of the previous historical sampling time point adjacent to the target historical sampling time point; comparing the distance value with a preset distance threshold, and in response to the distance value being greater than the preset distance threshold, updating the machine tool detection model; Calculate the learning rate correction weight at the target historical sampling time point. The learning rate correction weight satisfies the relationship: ,in, Indicates The learning rate correction weight at each historical sampling time point, , Respectively represent , The historical sampling time point is The entropy value in the dimension is Represents the total number of dimensions, Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, represents the learning rate correction weight, Represents the normalization function. In the learning rate correction weight relationship, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension; Update the learning rate of the machine tool detection model to satisfy the relationship: ,in, , Respectively represent and The learning rate of the machine tool detection model corresponding to each historical sampling time point, represents the learning rate; A machine tool detection and correction model is obtained, and the machine tool operating parameters at the real-time sampling time point are input into the machine tool detection and correction model to obtain the machine tool operating status at the real-time sampling time point to adjust the machine tool.

2. A high-precision CNC machine tool processing compensation adjustment method according to claim 1, characterized in that: The prediction network model is a BP neural network, which includes an input layer, a hidden layer and an output layer; the input layer is connected to the hidden layer, and the hidden layer is connected to the output layer; after the multi-dimensional machine tool operation parameters are input from the input layer, the input is feature extracted through the hidden layer, and the extracted features are output through the output layer to generate the machine tool operation status.

3. A high-precision CNC machine tool processing compensation adjustment method according to claim 1, characterized in that: The training preset prediction network model includes: The multi-dimensional machine tool operation parameters at any historical sampling time point are obtained as input information, and the machine tool operation status at the next historical sampling time point adjacent to any historical sampling time point is used as label information to obtain a set of training data; Input the input information in the training data into the prediction network model to obtain the output result; The mean square error loss function is used to calculate the loss value between the output result and the label information. The error signal is back-propagated according to the loss value using the gradient descent method. The weights and biases in the network are updated according to the gradient of the network parameters and the learning rate to reduce the loss value. Iteratively update the network parameters of the prediction network model until the loss value is less than the set loss value or the maximum number of training times is reached, then stop updating and obtain the machine tool detection model.

4. A high-precision CNC machine tool processing compensation adjustment method according to claim 1, characterized in that: The entropy value satisfies the polynomial: , ,in, Indicates The historical sampling time point is The parameter values ​​obtained in the dimensions, Indicates The historical sampling time point is The probability of the parameter value obtained in the dimension appearing in the running parameter sequence, Indicates the total number of machine tool operating parameters in the operating parameter sequence. Indicates The historical sampling time point is The entropy value in the dimension is Indicates the sequence number of the historical sampling time point, Indicates the dimension number, represents the entropy value, Indicates the parameter value. represents the probability, represents the logarithmic function. In the entropy polynomial, The historical sampling time point is the target historical sampling time point, The dimension is the target dimension.

5. The high-precision CNC machine tool processing compensation adjustment method according to claim 1, characterized in that: The preset judgment conditions include: Compare the entropy value of the target historical sampling time point in the target dimension with a preset entropy value threshold of the target dimension, and in response to the entropy value of the target historical sampling time point in the target dimension being greater than the entropy value threshold, update the machine tool detection model.

Images