Intelligent machine tool cutter temperature control method

Abstract

The application discloses an intelligent machine tool cutter temperature control method, comprising the following steps: S1, before cutting processing, a cutting temperature prediction model U TY is used to predict a cutting temperature trend curve I(t g , T g ) of the cutter; S2, a cutter temperature control model U TC based on a machine learning algorithm is used to generate a temperature control strategy function f(t g , C g ); S3, the cutter is used to cut a workpiece, and a cooling device is controlled according to the temperature control strategy function f(t g , C g ), so that the cutter is cooled by the cooling device; and S4, in the whole cutting process, actual temperature values T s of the cutter at different moments are detected. The application can accurately control the temperature of the cutter, significantly reduce the wear of the cutter and prolong the service life of the cutter.

Description

Technical Field

[0001] This invention relates to the field of machine tool component technology, and in particular to an intelligent machine tool cutting tool temperature control method. Background Technology

[0002] High cutting temperatures are a major cause of tool wear. However, higher cutting temperatures are beneficial for improving the toughness of carbide tool materials. In heated cutting, there is usually an optimal range for cutting temperature, within which not only do the workpiece material properties meet requirements, but the tool life is also relatively high. Similarly, the tool temperature also has a suitable range; controlling the tool to maintain a suitable temperature can help extend its service life.

[0003] During machining, the tool temperature will rise sharply due to friction. Therefore, the tool temperature control is mainly to control the temperature drop. During the cutting process, certain measures are usually required to cool the tool, such as air cooling or cutting fluid spraying. In cutting machining using cutting fluid cooling, the main factors affecting heat generation and dissipation are cutting parameters and cutting fluid (Yang Zhengle, Chen Yaru, Chen Yuanbo. Automatic control of cutting temperature in machining[J]. Digital World, 2019(4):1.). Cutting parameters are mainly determined by three elements: cutting speed, feed rate (or feed speed vf), and depth of cut. These are the process parameters required to adjust the relative speed and relative position between the tool and the workpiece. They can be expressed by the three elements of cutting parameters and can reflect the overall cutting process. Cutting fluid is mainly used for tool cooling. The amount of cutting fluid sprayed, its temperature, and its pressure will have a significant impact on its cooling effect.

[0004] Traditional tool temperature control schemes are generally based on feedback mechanisms. This involves first detecting the actual tool temperature, and then providing cutting fluid at a specific temperature, flow rate, and pressure based on the detected temperature to bring the tool temperature closer to the target temperature. Examples include a tool temperature detection and cooling device for an automated machine tool disclosed in patent CN112405027A, a cooling device for a turning machine tool and a turning machine tool using the same device disclosed in patent CN116765444A, and an automatic temperature control system for machine tool disclosed in patent CN207642796U. However, this approach suffers from significant lag, poor temperature control accuracy, and slow response (i.e., slow temperature regulation), resulting in ineffective tool temperature control. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent machine tool cutting tool temperature control method to address the shortcomings of the prior art.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an intelligent machine tool cutting tool temperature control method, comprising the following steps: S1. Before cutting, first determine the cutting process characteristic value E of the machine tool. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Through the cutting temperature prediction model U TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g ); The cutting temperature trend curve represents different machining times t. g Predicted cutting temperature T g ; S2. Fixed characteristic value L according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ), through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C g ); The temperature control strategy function includes different processing times t. g Control value C of the cooling process parameters g ; S3. Start the machine tool and perform cutting on the workpiece using the cutting tool, while simultaneously following the temperature control strategy function. f (t) g C g The cooling equipment is controlled to cool the cutting tool. The start time of the cutting process is the same as the start time of the cooling equipment working according to the temperature control strategy function, which is taken as time zero. S4. During the entire cutting process, detect the actual temperature value T of the tool at different times. s And according to T s With temperature control value T b The difference determines whether the temperature control strategy function needs to be adjusted. f (t) g C g Manual optimization is performed until the cutting process is completed; If manual optimization was performed, then the manually optimized temperature control strategy function will be... Replace the original temperature control strategy function f (t) g C g ).

[0007] Preferably, step S4 specifically includes: S4-1. During the entire cutting process, measure the actual tool temperature T at different times. s Real-time statistics of T during the cutting process s >T b The time of temperature exceedance is denoted as t. CX And calculate the temperature exceedance time t. CX The actual temperature value T of the tool at each moment within the time frame s With temperature control value T b The difference between them is ΔT; Calculate the degree of temperature deviation δ. ; S4-2. During the cutting process, when δ>δ T The duration exceeds the threshold t ε1 hour: An alarm message is issued, and the cooling process control parameters are manually adjusted until δ≤δ T And the duration exceeds the threshold t ε2 Record the manually adjusted time period t re The internal cooling process control parameters are replaced, and the t value in the currently used temperature control strategy function is replaced. re The cooling process control parameters within the specified time period are updated to obtain the corrected temperature control strategy function. And serve as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Otherwise, the temperature control strategy function will be used. f (t) g C g This serves as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Where, δ T The threshold for the degree of temperature deviation is set in advance.

[0008] Preferably, the intelligent machine tool cutting tool temperature control method further includes the following steps: S5, Tool Temperature Control Model U TC Optimizations will be made, specifically including: S5-1. In each cutting process, the statistical analysis shows that δ≤δ throughout the entire cutting process. T Time t Y Total processing time (t) z The proportion ηY , , will η Y As a feedback evaluation parameter of the temperature control strategy function; S5-2, Then fix the characteristic value L of the cooling process in the current cutting process. c Temperature control value T b Cutting temperature trend curve I (t) g T g The combination of these data points serves as the temperature control data q. T , temperature control data q T Compared with the actual temperature control strategy function in the current cutting process f s (t) g C g Combine them and match the corresponding η. Y As a practical temperature control strategy function f s (t) g C g The evaluation labels are used to form an optimized training data q. Y Store in the optimized dataset Q Y middle; S5-3. After several cutting processes, the optimized dataset Q is used. Y For the tool temperature control model V TC The optimization is performed using temperature control data q. T Input, actual temperature control strategy function f s (t) g C g ) is the output, η Y The goal is to minimize the temperature of the cutting tool, and the optimized tool temperature control model U' is obtained after training. TC Replace the previous tool temperature control model U TC .

[0009] Preferably, 0.5% < δ T <10%; t ε1 = (0.1-10%)×t z ;t ε2 =(0.5-2)×t ε1 .

[0010] Preferably, the cutting process characteristic value E g = Cutting speed Vc × Feed rate f × Depth of cut αp, where the unit of cutting speed Vc is m / min, the unit of feed rate f is mm / r, and the unit of depth of cut αp is mm; Material characteristic value E of the machined workpiece c = Tool hardness - Workpiece hardness; Environmental factor characteristic value E h The ambient temperature during the processing is expressed in °C.

[0011] Preferably, the cooling device includes a cooling host for providing a cooling medium, a delivery pipe connected to the host, and a nozzle connected to the delivery pipe for spraying the cooling medium onto the cutting tool for cooling. Cooling process fixed characteristic value L c The temperature T of the cooling medium J The unit is ℃; Temperature control value T b This is the upper limit of the working temperature of the cutting tool, which is set in advance according to the type of cutting tool, and the unit is ℃.

[0012] Preferably, the cooling process parameters include the injection flow rate Q and pressure P of the cooling medium injected by the nozzle; the injection flow rate is the volumetric flow rate, with units of L / min; Control value C of cooling process parameters k Including the injection flow rate control value Q of the cooling medium Ck and pressure control value P Ck ; Working value C of cooling process parameters g Including the operating value Q of the cooling medium injection flow rate Cg and pressure working value P Cg .

[0013] Preferably, in step S3, the temperature control strategy function is followed. f (t) g C g The cooling equipment is controlled to ensure that the corresponding processing time t is within acceptable limits. g Below, the operating value C of the cooling process parameters of the cooling equipment. g The value of the injection flow rate should be kept consistent with the cooling process parameter control value Ck: that is, the operating value of the injection flow rate Q. Cg With injection flow control value Q Ck Consistent, pressure control value P Ck With pressure working value P Cg Consistent.

[0014] Preferably, the cutting temperature prediction model U TY This includes a NARX prediction model for predicting tool temperature at different machining times, and a cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model to generate a tool temperature versus machining time optical system. g T g The curve generation module; The NARX prediction model was constructed using the following method: S1-1, Constructing the training dataset Q1: S1-1-1. Without cooling measures, use cutting tools to machine workpieces with different hardness, and record the characteristic value E for each cutting process. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Below, different processing times t g The temperature of the cutting tool T at that time g The cutting process characteristic value E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Processing time t g Temperature T g Combined to form a training data set q1; S1-1-2. Through several experiments, the training data q1 obtained are combined to construct the training dataset Q1. S1-2, Using the training dataset Q1, with cutting process feature values ​​E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h For input, different processing times t g The temperature of the cutting tool T at that time g To produce the output, the NARX model is trained, and the trained NARX prediction model is obtained. In step S1, before the cutting process, the cutting process characteristic value E is first set. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Inputting the data into the NARX prediction model yields different processing times t. g The predicted tool temperature T g The curve generation module automatically generates the cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model. g T g ).

[0015] Preferably, the tool temperature control model U TC It is constructed using the following method: S2-1, Construct the training dataset Q2: S2-1-1. Conduct the experiment according to the following method: (1) Fixed characteristic value L under different cooling processes c Temperature control value T b Cutting temperature trend curve I (t) g T g Under these conditions, different processing times t are obtained manually. g Control value C of the cooling process parametersg Forming a temperature control strategy function fr (t) g C g ); (2) Following the method in step S3, the temperature control strategy function is applied during the cutting process. fr (t) g C g The cooling equipment is controlled to cool the cutting tool until the cutting process is completed. If throughout the entire cutting process: All satisfy δ>δ T The duration does not exceed the threshold t ε1 Then the current temperature control strategy function will be... fr (t) g C g As a qualified function, retain the data from the current experiment and fix the characteristic value L of the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g Temperature control strategy function fr (t) g C g The data are combined to form a training data set q2; Otherwise, the data from the current experiment will not be saved; S2-1-2. Conduct several experiments and combine all the training data q2 obtained to construct the training dataset Q2; S2-2, Using the training dataset Q2, the feature value L is fixed according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ( ) is the input, and the temperature control strategy function is... fr (t) g C g The output is used to train the GAN model, and the trained model is the tool temperature control model U. TC .

[0016] The beneficial effects of this invention are: The present invention employs a cutting temperature prediction model U TY Based on the upper cutting process characteristic value E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h It can predict the tool temperature at various moments during the cutting process and generate the tool cutting temperature trend curve I(t). g T gThen, by first predicting the cutting temperature trend curve I(t) of the tool temperature versus machining time optical system, the cutting temperature trend curve is obtained. g T g Then, based on the set temperature control value T b Fixed characteristic value L of cooling process c Through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C g By further controlling the cooling equipment, the actual temperature of the tool at various times can be made closer to the temperature control value T. b It has better synchronization and can avoid the lag between the actual temperature and the target temperature in the feedback control scheme; by accurately controlling the tool temperature, tool wear can be significantly reduced and tool service life can be extended. This invention also enables the development of a tool temperature control model U. TC Continuous optimization can be performed to continuously improve the tool temperature control model U. TC The generated temperature control strategy function f (t) g C g The accuracy of ). Attached Figure Description

[0017] Figure 1 This is a flowchart of the intelligent machine tool cutting tool temperature control method of the present invention; Figure 2 In step S1 of embodiment 1 of the present invention, U TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g ); Figure 3 The wear curves are for Example 1 and the comparative example. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to embodiments, so that those skilled in the art can implement it based on the description.

[0019] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0020] Reference Figure 1 This invention provides an intelligent machine tool cutting tool temperature control method, comprising the following steps: S1. Before cutting, first determine the cutting process characteristic value E of the machine tool. g , Material characteristic value E of the workpiece c Environmental factor characteristic value Eh Through the cutting temperature prediction model U TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g The cutting temperature trend curve represents the temperature changes at different machining times t. g Predicted cutting temperature T g .

[0021] During the cutting process, friction between the cutting tool and the workpiece generates heat, causing the tool temperature to rise sharply. This temperature rise is influenced by various factors, which can be mainly summarized into the following three aspects: (1) Cutting process, including cutting speed Vc, feed rate f, depth of cut αp, etc. In this invention, the cutting process characteristic value E is used. g Characterizing this influencing factor, specifically, the cutting process characteristic value E g = Cutting speed Vc × Feed rate f × Depth of cut αp, where the unit of cutting speed Vc is m / min, the unit of feed rate f is mm / r, and the unit of depth of cut αp is mm. The above process parameters in the cutting process are obtained in advance based on the specific processing requirements of the workpiece and are known quantities; for example, for a certain workpiece to be processed into a product of a certain shape and size, its cutting process will be designed and confirmed in advance, and in this invention, the cutting process can be directly retrieved.

[0022] (2) The characteristics of the workpiece material can be mainly attributed to the hardness difference between the workpiece and the cutting tool. In this invention, the characteristic value E of the workpiece material is used. c Characterization is performed to obtain the material characteristic value E of the machined workpiece. c = Tool hardness - Workpiece hardness; Hardness can be characterized by Rockwell hardness, Vickers hardness, etc., as long as they are consistent.

[0023] (3) The environmental influence of the cutting process is mainly characterized by the ambient temperature during the cutting process, which is denoted as the environmental factor characteristic value E. h (Unit: °C)

[0024] Based on the above factors, the tool temperature at various moments during the cutting process can be predicted. Specifically, this invention employs a cutting temperature prediction model U. TY Based on the above influencing factors (cutting process characteristic value E) g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Predict the tool temperature at various moments during the cutting process and generate the tool cutting temperature trend curve I(t). g T g ).

[0025] In this invention, the cutting temperature prediction model U TY This includes a NARX prediction model for predicting tool temperature at different machining times, and a cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model to generate a tool temperature versus machining time optical system. g T g The curve generation module; The NARX prediction model is constructed using the following method: S1-1, Constructing the training dataset Q1: S1-1-1. Without cooling measures, use cutting tools to machine workpieces with different hardness, and record the characteristic value E for each cutting process. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Below, different processing times t g The temperature of the cutting tool T at that time g The cutting process characteristic value E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Processing time t g Temperature T g Combined to form a training data set q1; S1-1-2. Through several experiments, the training data q1 obtained are combined to construct the training dataset Q1. S1-2, Using the training dataset Q1, with cutting process feature values ​​E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h For input, different processing times t g The temperature of the cutting tool T at that time g To produce the output, the NARX model is trained, and the trained NARX prediction model is obtained. In step S1, before the cutting process, the cutting process characteristic value E is first set. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Inputting the data into the NARX prediction model yields different processing times t. g The predicted tool temperature T g The curve generation module automatically generates the cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model. g T g ).

[0026] The NARX (Nonlinear Autoregressive Exogenous) model is a time series modeling technique used to predict future values ​​based on past values. The NARX model is a feedforward neural network that includes autoregressive (AR) and exogenous (EX) inputs. In this invention, the NARX model is used as the basic network model and trained on a constructed training dataset Q1, thereby enabling accurate prediction of tool temperature at various moments during the cutting process.

[0027] S2. Fixed characteristic value L according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ), through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C g The temperature control strategy function includes different processing times t. g Control value C of the cooling process parameters g .

[0028] S3. Start the machine tool and perform cutting on the workpiece using the cutting tool, while simultaneously following the temperature control strategy function. f (t) g C g The cooling equipment is controlled to cool the cutting tool.

[0029] High cutting temperatures are the main cause of tool wear, but higher cutting temperatures are beneficial to improving the toughness of carbide tool materials. Therefore, controlling the tool to maintain a suitable temperature can help extend the tool's service life.

[0030] Therefore, in this invention, after obtaining the optimal cutting temperature range in advance based on the specific material of the cutting tool, a temperature control value T is set. b This ensures that the tool temperature is controlled near a certain value, thereby guaranteeing a longer tool life and better cutting performance without wasting cutting fluid. Furthermore, in this invention, the primary requirement is that the tool temperature does not exceed the temperature control value T. b That is, when the tool temperature is lower than the temperature control value T b At this time, there is no need for temperature control. This method can not only meet the basic usage requirements, but also simplify the control operation and save cutting fluid.

[0031] Conventional tool temperature control schemes are generally based on feedback mechanisms. This means that the actual tool temperature is first detected, and then relevant cooling measures are provided based on the detected actual tool temperature value to make the tool temperature approach the target temperature. This method has a large lag and poor temperature control accuracy, resulting in poor temperature control effect on the tool.

[0032] In this invention, the cutting temperature trend curve I(t) relating tool temperature to machining time is first predicted. g T g Then, based on the set temperature control value T b Fixed characteristic value L of cooling process c Through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C g By further controlling the cooling equipment, the actual temperature of the tool at various times can be made closer to the temperature control value T. b It has better synchronization and can avoid the lag between the actual temperature and the target temperature in the feedback control scheme.

[0033] In this invention, the cooling equipment includes a cooling host for providing a cooling medium, a delivery pipe connected to the host, and a nozzle connected to the delivery pipe for spraying the cooling medium onto the cutting tool for cooling; the cooling process has a fixed characteristic value L. c The temperature T of the cooling medium J The unit is ℃; temperature control value T b This is the upper limit of the working temperature of the cutting tool, which is set in advance according to the type of cutting tool, and the unit is ℃.

[0034] In this invention, the cooling process parameters include the injection flow rate Q and pressure P of the cooling medium injected by the nozzle; the injection flow rate is the volumetric flow rate, with units of L / min; Control value C of cooling process parameters k Including the injection flow rate control value Q of the cooling medium Ck and pressure control value P Ck ; Working value C of cooling process parameters g Including the operating value Q of the injection flow rate of the cooling medium. Cg and pressure working value P Cg .

[0035] In this invention, according to the temperature control strategy function f (t) g C g By controlling parameters such as the injection flow rate Q and pressure P of the cooling medium, the temperature of the cutting tool can be controlled within the temperature control value T. bThis improves the synchronization and accuracy of temperature control by placing the device nearby.

[0036] In step S3, according to the temperature control strategy function f (t) g C g The cooling equipment is controlled to ensure that the corresponding processing time t is within acceptable limits. g Below, the operating value C of the cooling process parameters of the cooling equipment. g The value of the injection flow rate should be kept consistent with the cooling process parameter control value Ck: that is, the operating value of the injection flow rate Q. Cg With injection flow control value Q Ck Consistent, pressure control value P Ck With pressure working value P Cg Consistent.

[0037] In this invention, the tool temperature control model U TC It is constructed using the following method: S2-1, Construct the training dataset Q2: S2-1-1. Conduct the experiment according to the following method: (1) Fixed characteristic value L under different cooling processes c Temperature control value T b Cutting temperature trend curve I (t) g T g Under these conditions, different processing times t are obtained manually. g Control value C of the cooling process parameters g Forming a temperature control strategy function fr (t) g C g ); (2) Following the method in step S3, the temperature control strategy function is applied during the cutting process. fr (t) g C g The cooling equipment is controlled to cool the cutting tool until the cutting process is completed. If throughout the entire cutting process: All satisfy δ>δ T The duration does not exceed the threshold t ε1 Then the current temperature control strategy function will be... fr (t) g C g As a qualified function, retain the data from the current experiment and fix the characteristic value L of the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g Temperature control strategy function fr (t) g Cg The data are combined to form a training data set q2; Otherwise, the data from the current experiment will not be saved; S2-1-2. Conduct several experiments and combine all the training data q2 obtained to construct the training dataset Q2; S2-2, Using the training dataset Q2, the feature value L is fixed according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ( ) is the input, and the temperature control strategy function is... fr (t) g C g The output is used to train the GAN model, and the trained model is the tool temperature control model U. TC .

[0038] S4. During the entire cutting process, detect the actual temperature value T of the tool at different times. s And according to T s With temperature control value T b The difference determines whether the temperature control strategy function needs to be adjusted. f (t) g C g Manual optimization is performed until the cutting process is completed; if manual optimization is performed, the manually optimized temperature control strategy function is then applied. f’ (t) g C g Replace the original temperature control strategy function. f (t) g C g ).

[0039] Step S4 is as follows: S4-1. During the entire cutting process, measure the actual tool temperature T at different times. s Real-time statistics of T during the cutting process s >T b The time of temperature exceedance is denoted as t. CX And calculate the temperature exceedance time t. CX The actual temperature value T of the tool at each moment within the time frame s With temperature control value T b The difference between them is ΔT; Calculate the degree of temperature deviation δ. ; S4-2. During the cutting process, when δ>δ T The duration exceeds the threshold t ε1 hour: An alarm message is issued, and the cooling process control parameters are manually adjusted until δ≤δ T And the duration exceeds the threshold t ε2 Record the manually adjusted time period t re The internal cooling process control parameters are replaced, and the t value in the currently used temperature control strategy function is replaced. re The cooling process control parameters within the specified time period are updated to obtain the corrected temperature control strategy function. And serve as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Otherwise, the temperature control strategy function will be used. f (t) g C g This serves as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Where, δ T In a preferred embodiment, the pre-set temperature deviation threshold is 0.5% < δ. T <10%, t ε1 = (0.1-10%)×t z ;t ε2 =(0.5-2)×t ε1 .

[0040] By incorporating manual adjustment of cooling process control parameters, the temperature control strategy function can be modified. f (t) g C g We continuously optimize it to improve its accuracy.

[0041] S5, Tool Temperature Control Model U TC Optimize.

[0042] Specifically, it includes: S5-1. In each cutting process, the statistical analysis shows that δ≤δ throughout the entire cutting process. T Time t Y Total processing time (t) z The proportion η Y , , will η Y As a feedback evaluation parameter of the temperature control strategy function; S5-2, Then fix the characteristic value L of the cooling process in the current cutting process. c Temperature control value T b Cutting temperature trend curve I (t)g T g The combination of these data points serves as the temperature control data q. T , temperature control data q T Compared with the actual temperature control strategy function in the current cutting process f s (t) g C g Combine them and match the corresponding η. Y As a practical temperature control strategy function f s (t) g C g The evaluation labels are used to form an optimized training data q. Y Store in the optimized dataset Q Y middle; S5-3. After several cutting processes, the optimized dataset Q is used. Y For the tool temperature control model V TC The optimization is performed using temperature control data q. T Input, actual temperature control strategy function f s (t) g C g ) is the output, η Y The goal is to minimize the temperature of the cutting tool, and the optimized tool temperature control model U' is obtained after training. TC Replace the previous tool temperature control model U TC .

[0043] By analyzing the tool temperature control model U TC Optimization can continuously improve the tool temperature control model U. TC The generated temperature control strategy function f (t) g C g The accuracy of ). Example

[0044] An intelligent machine tool cutting tool temperature control method includes the following steps: S1. Before cutting, first determine the cutting process characteristic value E of the machine tool. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Through the cutting temperature prediction model U TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g The cutting temperature trend curve represents the temperature changes at different machining times t. g Predicted cutting temperature T g .

[0045] Cutting temperature prediction model U TY This includes a NARX prediction model for predicting tool temperature at different machining times, and a cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model to generate a tool temperature versus machining time optical system. g T g The curve generation module; The NARX prediction model is constructed using the following method: S1-1, Constructing the training dataset Q1: S1-1-1. Without cooling measures, use cutting tools to machine workpieces with different hardness, and record the characteristic value E of each cutting process. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Below, different processing times t g The temperature of the cutting tool T at that time g The cutting process characteristic value E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Processing time t g Temperature T g Combined to form a training data set q1; S1-1-2. Through several experiments, the training data q1 obtained are combined to construct the training dataset Q1. S1-2, Using the training dataset Q1, with cutting process feature values ​​E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h For input, different processing times t g The temperature of the cutting tool T at that time g To produce the output, the NARX model is trained, and the trained NARX prediction model is obtained. In step S1, before the cutting process, the cutting process characteristic value E is first set. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Inputting the data into the NARX prediction model yields different processing times t. g The predicted tool temperature T g The curve generation module automatically generates the cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model. g T g ).

[0046] S2. Fixed characteristic value L according to the cooling process. c Temperature control value T bCutting temperature trend curve I (t) g T g ), through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C g The temperature control strategy function includes different processing times t. g Control value C of the cooling process parameters g .

[0047] S3. Start the machine tool and perform cutting on the workpiece using the cutting tool, while simultaneously following the temperature control strategy function. f (t) g C g The cooling equipment is controlled to cool the cutting tool. The start time of the cutting process is the same as the start time of the cooling equipment operating according to the temperature control strategy function, and is taken as time zero.

[0048] The cooling equipment includes a cooling unit for providing the cooling medium, a delivery pipe connected to the unit, and a nozzle connected to the delivery pipe for spraying the cooling medium onto the cutting tool for cooling; the cooling process has a fixed characteristic value L. c The temperature T of the cooling medium J (20℃); Temperature control value T b This is the upper limit of the working temperature of the tool (665℃) set in advance according to the type of tool.

[0049] In this embodiment, the cooling process parameters include the injection flow rate Q and pressure P of the cooling medium injected by the nozzle; the injection flow rate is the volumetric flow rate, with units of L / min; Control value C of cooling process parameters k Including the injection flow rate control value Q of the cooling medium Ck and pressure control value P Ck ; Working value C of cooling process parameters g Including the operating value Q of the cooling medium injection flow rate Cg and pressure working value P Cg .

[0050] In step S3, according to the temperature control strategy function f (t) g C g The cooling equipment is controlled to ensure that the corresponding processing time t is within acceptable limits. g Below, the operating value C of the cooling process parameters of the cooling equipment. g The value of the injection flow rate should be kept consistent with the cooling process parameter control value Ck: that is, the operating value of the injection flow rate Q. Cg With injection flow control value QCk Consistent, pressure control value P Ck With pressure working value P Cg Consistent.

[0051] In this embodiment, the tool temperature control model U TC It is constructed using the following method: S2-1, Construct the training dataset Q2: S2-1-1. Conduct the experiment according to the following method: (1) Fixed characteristic value L under different cooling processes c Temperature control value T b Cutting temperature trend curve I (t) g T g Under these conditions, different processing times t are obtained manually. g Control value C of the cooling process parameters g Forming a temperature control strategy function fr (t) g C g ); (2) Following the method in step S3, the temperature control strategy function is applied during the cutting process. fr (t) g C g The cooling equipment is controlled to cool the cutting tool until the cutting process is completed. If throughout the entire cutting process: All satisfy δ>δ T The duration does not exceed the threshold t ε1 Then the current temperature control strategy function will be... fr (t) g C g As a qualified function, retain the data from the current experiment and fix the characteristic value L of the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g Temperature control strategy function fr (t) g C g The data are combined to form a training data set q2; Otherwise, the data from the current experiment will not be saved; S2-1-2. Conduct several experiments and combine all the training data q2 obtained to construct the training dataset Q2; S2-2, Using the training dataset Q2, the feature value L is fixed according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g( ) is the input, and the temperature control strategy function is... fr (t) g C g The output is used to train the GAN model, and the trained model is the tool temperature control model U. TC .

[0052] S4. During the entire cutting process, detect the actual temperature value T of the tool at different times. s And according to T s With temperature control value T b The difference determines whether the temperature control strategy function needs to be adjusted. f (t) g C g Manual optimization is performed until the cutting process is completed; if manual optimization is performed, the manually optimized temperature control strategy function is then applied. Replace the original temperature control strategy function f (t) g C g )

[0053] Step S4 is as follows: S4-1. During the entire cutting process, measure the actual tool temperature T at different times. s Real-time statistics of T during the cutting process s >T b The time of temperature exceedance is denoted as t. CX And calculate the temperature exceedance time t. CX The actual temperature value T of the tool at each moment within the time frame s With temperature control value T b The difference between them is ΔT; Calculate the degree of temperature deviation δ. ; S4-2. During the cutting process, when δ>δ T The duration exceeds the threshold t ε1 hour: An alarm message is issued, and the cooling process control parameters are manually adjusted until δ≤δ T And the duration exceeds the threshold t ε2 Record the manually adjusted time period t re The internal cooling process control parameters are replaced, and the t value in the currently used temperature control strategy function is replaced. re The cooling process control parameters within the specified time period are updated to obtain the corrected temperature control strategy function. And serve as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Otherwise, the temperature control strategy function will be used. f (t) g C g This serves as the actual temperature control strategy function for current cutting processes. f s (t) g C g ); Where, δ T In this embodiment, δ is a pre-set threshold for temperature deviation. T =1.5%, t ε1 =3%×t z ;t ε2 =1.2×t ε1 .

[0054] S5, Tool Temperature Control Model U TC Optimizations will be made, specifically including: S5-1. In each cutting process, the statistical analysis shows that δ≤δ throughout the entire cutting process. T Time t Y Total processing time (t) z The proportion η Y , , will η Y As a feedback evaluation parameter of the temperature control strategy function; S5-2, Then fix the characteristic value L of the cooling process in the current cutting process. c Temperature control value T b Cutting temperature trend curve I (t) g T g The combination of these data points serves as the temperature control data q. T , temperature control data q T Compared with the actual temperature control strategy function in the current cutting process f s (t) g C g Combine them and match the corresponding η. Y As a practical temperature control strategy function f s (t) g C g The evaluation labels are used to form an optimized training data q. Y Store in the optimized dataset Q Y middle; S5-3. After several cutting processes, the optimized dataset Q is used. Y For the tool temperature control model V TC The optimization is performed using temperature control data q. T Input, actual temperature control strategy function f s(t) g C g ) is the output, η Y The goal is to minimize the temperature of the cutting tool, and the optimized tool temperature control model U' is obtained after training. TC Replace the previous tool temperature control model U TC .

[0055] In this embodiment, the material being processed is 45 tempered steel with a hardness of 56.5 HRC, a cuboid shape, and geometric dimensions of 200mm × 100mm × 40mm.

[0056] The system uses a VMC850 three-axis vertical CNC milling machine with a maximum speed of 8000 r / min and a maximum power of 22 kW. The cooling medium is Blasor cutting fluid, and the operating temperature is 20℃.

[0057] The cutting tool uses SG4 ceramic, with a hardness of 95.2 HRC, a bending strength of 1.05 GPa, and an impact toughness ≥14.6 kJ / m. 2 Tool geometry parameters: rake angle g0 = -6°, clearance angle a0 = 6°, principal cutting edge angle kr = 45°, cutting edge inclination angle ls = -6°.

[0058] Cutting process characteristic value E g In the middle section: cutting speed V = 175 m / min, feed rate f = 0.10 mm / r, depth of cut αp = 2.4 mm.

[0059] Temperature control value T b =665℃.

[0060] Reference Figure 2 U is in step S1 of this embodiment 1. TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g ).

[0061] Wear performance test: The wear morphology of the cutting tool was observed using a JSM-6380LA scanning electron microscope; the wear on the flank face of the cutting tool was measured using a tool microscope with an accuracy of 0.01 mm. After each cutting stroke, the wear morphology of the flank face was observed and the wear amount was measured; the cutting time of the cutting tool was calculated based on the cutting path, and the flank face wear amount under different cutting times was calculated and statistically analyzed to plot the wear curve.

[0062] As a comparison, a conventional feedback temperature control scheme is used, which monitors the real-time temperature of the cutting tool and the temperature control value T. b The difference is used to adjust the cooling process control parameters (i.e., the flow rate and pressure of the cooling medium Blasor cutting fluid), and the wear curve is plotted using the same method as above.

[0063] Reference Figure 3 The figures show the wear curves of Example 1 and the comparative example. It can be seen that, compared with the comparative example 1, the tool temperature control scheme of Example 1 can significantly reduce tool wear, thereby extending its service life.

[0064] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details.

Claims

1. An intelligent machine tool cutting tool temperature control method, characterized in that, Includes the following steps: S1. Before cutting, first determine the cutting process characteristic value E of the machine tool. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Through the cutting temperature prediction model U TY The predicted cutting temperature trend curve I(t) of the tool was obtained. g T g ); The cutting temperature trend curve characterizes different machining times t. g Predicted tool temperature T g ; S2. Fixed characteristic value L according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ), through a tool temperature control model U based on machine learning algorithms TC Generate temperature control strategy function f (t) g C k ); The temperature control strategy function includes different processing times t. g Control value C of the cooling process parameters k Control value C of cooling process parameters k Including the injection flow rate control value Q of the cooling medium Ck and pressure control value P Ck ; Cooling process fixed characteristic value L c The temperature T of the cooling medium J The unit is ℃; Temperature control value T b This is the upper limit of the working temperature of the cutting tool, set in advance according to the type of cutting tool, in °C. S3. Start the machine tool and perform cutting on the workpiece using the cutting tool, while simultaneously following the temperature control strategy function. f (t) g C k The cooling equipment is controlled to cool the cutting tool. The start time of the cutting process is the same as the start time of the cooling equipment working according to the temperature control strategy function, which is taken as time zero. S4. During the entire cutting process, detect the actual temperature value T of the tool at different times. s And according to T s With temperature control value T b The difference determines whether the temperature control strategy function needs to be adjusted. f (t) g C k Manual optimization is performed until the cutting process is completed; if manual optimization is performed, the manually optimized temperature control strategy function is then applied. Replace the original temperature control strategy function f (t) g C k Step S4 specifically includes: S4-1. During the entire cutting process, measure the actual tool temperature T at different times. s Real-time statistics of T during the cutting process s >T b The time of temperature exceedance is denoted as t. CX And calculate the time t for temperature exceeding the limit. CX The actual temperature value T of the tool at each moment within the time frame s With temperature control value T b The difference between them is ΔT; Calculate the degree of temperature deviation δ. ; S4-2. During the cutting process, when δ>δ T The duration exceeds the threshold t ε1 hour: An alarm message is issued, and the cooling process control parameters are manually adjusted until δ≤δ T And the duration exceeds the threshold t ε2 Record the manually adjusted time period t re The internal cooling process control parameters are adjusted, and the t parameter in the currently used temperature control strategy function is replaced. re The cooling process control parameters within a given time period are updated to obtain a manually optimized temperature control strategy function. And serve as the actual temperature control strategy function for current cutting processes. f s (t) g C k ); Otherwise, the temperature control strategy function will be used. f (t) g C k This serves as the actual temperature control strategy function for current cutting processes. f s (t) g C k ); Where, δ T A pre-set threshold for the degree of temperature deviation; S5, Tool Temperature Control Model U TC Optimizations will be made, specifically including: S5-1. In each cutting process, the statistical analysis shows that δ≤δ throughout the entire cutting process. T Time t Y accounting for t of the total processing time z The proportion η Y , , will η Y As a feedback evaluation parameter of the temperature control strategy function; S5-2, Then fix the characteristic value L of the cooling process in the current cutting process. c Temperature control value T b Cutting temperature trend curve I (t) g T g The combination of these data points serves as the temperature control data q. T , temperature control data q T Compared with the actual temperature control strategy function in the current cutting process f s (t) g C k Combine them and match the corresponding η. Y As a practical temperature control strategy function f s (t) g C k The evaluation labels are used to form an optimized training data q. Y Store in the optimized dataset Q Y middle; S5-3. After several cutting processes, the optimized dataset Q is used. Y Tool temperature control model U TC The optimization is performed using temperature control data q. T Input, actual temperature control strategy function f s (t) g C k ) is the output, η Y With the goal of maximizing the tool temperature control, the model is trained to obtain an optimized tool temperature control model U'. TC Replace the previous tool temperature control model U TC ; Among them, the cutting process characteristic value E g = Cutting speed Vc × Feed rate f × Depth of cut αp, where the unit of cutting speed Vc is m / min, the unit of feed rate f is mm / r, and the unit of depth of cut αp is mm; Material characteristic value E of the machined workpiece c = Tool hardness - Workpiece hardness; Environmental factor characteristic value E h The ambient temperature during the processing is expressed in °C. Cutting temperature prediction model U TY This includes a NARX prediction model for predicting tool temperature at different machining times, and a cutting temperature trend curve I(t) for generating the relationship between tool temperature and machining time based on the prediction results of the NARX prediction model. g T g The curve generation module.

2. The intelligent machine tool cutting tool temperature control method according to claim 1, characterized in that, in, 0.5%<δ T <10%; t ε1 =(0.1%-10%)×t z ;t ε2 =(0.5-2)×t ε1 。 3. The intelligent machine tool cutting tool temperature control method according to claim 1, characterized in that, The cooling device includes a cooling host for providing a cooling medium, a delivery pipe connected to the host, and a nozzle connected to the delivery pipe for spraying the cooling medium onto the cutting tool for cooling.

4. The intelligent machine tool cutting tool temperature control method according to claim 3, characterized in that, The cooling process parameters include the injection flow rate Q and pressure P of the cooling medium injected by the nozzle; The jet flow rate is a volumetric flow rate, with units of L / min; Working value C of cooling process parameters g Including the operating value Q of the cooling medium injection flow rate Cg and pressure working value P Cg .

5. The intelligent machine tool cutting tool temperature control method according to claim 4, characterized in that, In step S3, according to the temperature control strategy function f (t) g C k The cooling equipment is controlled to ensure that the corresponding processing time t is within acceptable limits. g Below, the operating value C of the cooling process parameters of the cooling equipment. g Control value C of cooling process parameters k Maintain consistency: that is, keep the injection flow rate working value Q Cg With injection flow control value Q Ck Consistent, pressure control value P Ck With pressure working value P Cg Consistent.

6. The intelligent machine tool cutting tool temperature control method according to claim 5, characterized in that, The NARX prediction model was constructed using the following method: S1-1, Construct the first training dataset Q1: S1-1-1. Without cooling measures, use cutting tools to machine workpieces with different hardness, and record the characteristic value E for each cutting process. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Below, different processing times t g The temperature of the cutting tool T at that time g The cutting process characteristic value E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Processing time t g The temperature T of the cutting tool g Combined to form a first training data set q1; S1-1-2. Through several experiments, the first training data q1 obtained are combined to construct the first training dataset Q1; S1-2, Using the first training dataset Q1, with cutting process feature values ​​E g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h For input, different processing times t g The temperature of the cutting tool T at that time g To produce the output, the NARX model is trained, and the trained NARX prediction model is obtained. In step S1, before the cutting process, the cutting process characteristic value E is first set. g , Material characteristic value E of the workpiece c Environmental factor characteristic value E h Inputting the data into the NARX prediction model yields different processing times t. g The predicted tool temperature T g The curve generation module automatically generates the cutting temperature trend curve I(t) based on the prediction results of the NARX prediction model. g T g ).

7. The intelligent machine tool cutting tool temperature control method according to claim 6, characterized in that, Tool temperature control model U TC It is constructed using the following method: S2-1, Construct the second training dataset Q2: S2-1-1. Conduct the experiment according to the following method: (1) Fixed characteristic value L under different cooling processes c Temperature control value T b Cutting temperature trend curve I (t) g T g Under these conditions, different processing times t are obtained manually. g Control value C of the cooling process parameters k Forming a temperature control strategy function f (t) g C k ); (2) In the cutting process, according to the temperature control strategy function f (t) g C k The cooling equipment is controlled to cool the cutting tool until the cutting process is completed. If throughout the entire cutting process: All satisfy δ>δ T The duration does not exceed the threshold t ε1 Then the current temperature control strategy function will be... f (t) g C k As a qualified function, retain the data from the current experiment and fix the characteristic value L of the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g Temperature control strategy function f (t) g C k This combination forms a second training data set q2; Otherwise, the data from the current experiment will not be saved; S2-1-2. Conduct several experiments and combine all the obtained second training data q2 to construct the second training dataset Q2. S2-2, Using the second training dataset Q2, the feature value L is fixed according to the cooling process. c Temperature control value T b Cutting temperature trend curve I (t) g T g ( ) is the input, and the temperature control strategy function is... f (t) g C k The output is used to train the GAN model, and the trained model is the tool temperature control model U. TC .

Images