Intelligent adaptive machining method that can automatically identify parts to be processed

Abstract

This invention belongs to the field of intelligent manufacturing technology and discloses an intelligent adaptive machining method that can automatically identify parts to be processed. First, data acquisition and transmission are performed, collecting the spindle power data of the CNC machine tool during part processing and sending the data to an industrial control computer. Then, a constrained derivative dynamic time bending algorithm is used on the industrial control computer to perform similarity matching on the parts to be processed. If the matching fails, the optimal feed rate is calculated based on a fuzzy logic controller and saved to a database for adaptive machining on the CNC machine tool. If the matching succeeds, the optimal feed rate saved in the database is used for adaptive machining on the CNC machine tool. This method can automatically identify the type of part to be processed and automatically match the corresponding optimal feed rate, improving processing efficiency while simplifying the adaptive machining operation process and enhancing the intelligence level of adaptive machining on CNC machine tools.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing technology and relates to an intelligent adaptive manufacturing method that can automatically identify parts to be processed. Background Technology

[0002] In traditional CNC machining, cutting parameters are determined during the pre-machining programming stage, typically by computer-aided manufacturing software or the operator's experience. However, using fixed process parameters is unsuitable for actual machining conditions, leading to several adverse consequences: firstly, setting process parameters too low results in low machining efficiency, making it difficult to fully utilize the CNC equipment's performance; secondly, setting process parameters too high or using an unreasonable combination leads to excessively high cutting forces, cutting temperatures, and cutting power, resulting in significant tool deformation, rapid wear, and decreased workpiece surface quality, and in severe cases, damage to the tool or even the spindle. Adaptive machining on CNC machine tools can adjust cutting parameters in real time according to the machine tool's machining load. This intelligent manufacturing technology enhances the intelligence of CNC machine tools, improves parts machining efficiency, and reduces reliance on experienced operators.

[0003] In the field of adaptive machining for CNC machine tools, domestic scholars have already conducted relevant research. In the patent "Control Method and Device for CNC Machine Tools and Their Adaptive Machining" (application number: CN202010930080.6), a deep learning model is used to predict cutting forces, and the optimal feed rate is autonomously found based on the predicted cutting forces to achieve adaptive machining of CNC machine tools. In the patent "An Adaptive Control Method for Machine Tools Based on GA-BP Neural Network Algorithm" (application number: CN201910732917.3), a neural network algorithm is used to optimize and adaptively adjust the feed rate and spindle speed in real time, which can effectively improve machining efficiency and quality. In the paper "Online Monitoring and Machining Parameters of Composite Material Drilling Surface Roughness,"... In the paper "Adaptive Optimization of Number" (Journal of Mechanical Engineering, 2020, 56(02): 27-34+42.), an online monitoring model for hole wall roughness based on support vector regression was established, and the drilling parameters were optimized under the current monitoring state using the simulated annealing algorithm to ensure drilling quality. In the paper "Research on Constant Power Constraint Adaptive Machining Method with Integrated Chatter Control" (Mechanical Manufacturing and Automation, 2018, 47(05): 41-44+52.), based on fuzzy theory and variable speed suppression of cutting chatter theory, the adjustment principle of machine tool machining parameters was formulated with machining efficiency as the goal, and constant power constraint adaptive machining with integrated chatter control was realized.

[0004] Analysis of existing adaptive machining technologies for CNC machine tools reveals that while current adaptive machining methods can improve machining efficiency by adjusting cutting parameters, they cannot automatically identify the workpiece to be machined. Before applying adaptive machining, operators need to pre-select the type of workpiece, which increases the workload when there are many types of workpieces, making it difficult to promote adaptive machining technology in practical applications. Therefore, an intelligent adaptive machining method that can automatically identify the workpiece is needed. This invention addresses these problems by proposing an automatic workpiece identification method. This method identifies the type of workpiece based on the machine tool spindle power collected during CNC machining and adjusts the CNC machine tool's feed rate in real time using a set constant reference power, thereby improving machining efficiency and enhancing the intelligence level of CNC machining. Summary of the Invention

[0005] To address the issue that common adaptive machining methods for CNC machine tools require manual selection of the type of part to be machined, this invention designs an intelligent adaptive machining method that can automatically identify the part to be machined.

[0006] The technical solution of this invention:

[0007] An intelligent adaptive machining method that can automatically identify parts to be processed, the steps of which are as follows:

[0008] S1: Data Acquisition and Transmission: Acquire spindle power data during CNC machine tool machining of parts, and send the acquired spindle power data to the industrial control computer;

[0009] S1.1: Install a power sensor at an appropriate location on the spindle controller of the CNC machine tool, and use the power sensor to collect spindle power data during the machining process;

[0010] S1.2: Send the spindle power data to the industrial computer via the data transmission line;

[0011] S2: Use an industrial control computer to perform a similarity matching comparison between the spindle power data obtained in step S1 and the spindle power data in the database. If the similarity matching comparison fails, proceed to step S3; if the matching is successful, proceed to step S4. The database is stored in the industrial control computer.

[0012] S2.1: Use the Piecewise Aggregate Approximation Algorithm (PAA algorithm) to perform dimensionality reduction on the spindle power data obtained in step S1 to obtain the dimensionality-reduced spindle power data;

[0013] S2.2: Apply the constrained derivative dynamic time bending algorithm (CDDTW algorithm) to perform similarity matching and comparison between the dimension-reduced spindle power data and each spindle power data in the database to determine whether the optimal feed rate of the machined part exists in the database;

[0014] S2.2.1: Calculate the distance matrix D between the dimensionality-reduced spindle power data in step S2.1 and the first spindle power data in the database, and the element d in the i-th row and j-th column of the distance matrix D. i,j The calculation formula is:

[0015]

[0016] Where, d i,j Let a be the element in the i-th row and j-th column of the distance matrix D; i ′ represents the numerical derivative of the i-th data point in the dimensionality-reduced spindle power data, calculated using the following formula: And a′1=a′2, a′ m =a′ m-1 b j ′ is the numerical derivative of the j-th data point of the first spindle power data in the database, calculated using the following formula: And b1′=b2′, b′ n =b′ n-1 ω is the penalty intensity coefficient; m is the number of data points in the dimensionality-reduced spindle power data; n is the number of data points in the first spindle power data in the database;

[0017] S2.2.2: Solve for the alignment path of the distance matrix D. The alignment path consists of path points in the distance matrix D that satisfy the following conditions: (1) The starting path point of the alignment path is d. 1,1 The end path point is d. m,n (2) In the alignment path, the next path point must be one of the three elements adjacent to the previous path point. These three path points are located to the right, below and to the lower right of the previous path point, respectively. (3) The next path point is the one with the smallest value among the three path points mentioned above. If the three path points have the same value, the path point to the lower right is selected first, then the path point below, and finally the path point to the right.

[0018] S2.2.3: Calculate the alignment path smoothness S of the distance matrix D. The formula for the alignment path smoothness S is: Where, N inc N represents the total number of special path points in the alignment path, characterized by being located to the lower right of the previous path point; tot This represents the total number of all path points in the alignment path;

[0019] S2.2.4: Repeat steps S2.2.1 to S2.2.3 to calculate the alignment path smoothness S of the dimensionality-reduced spindle power data and each spindle power data in the database. If there is an alignment path smoothness S greater than the similarity threshold α, it is considered that the optimal feed speed of the machined part exists in the database, and the optimal feed speed corresponding to the spindle power data in the database with the largest alignment path smoothness S is selected as the optimal feed speed of the machined part; if there is no alignment path smoothness S greater than the similarity threshold α, it is considered that the optimal feed speed of the machined part is not in the database.

[0020] S2.2.5: If the optimal feed rate of the part being processed is not in the database, then set a label for the dimensionality-reduced spindle power data to obtain the dimensionality-reduced spindle power data with the label;

[0021] Furthermore, in step S2.2.1, the penalty intensity coefficient ω is typically set to 0.6 to 0.8;

[0022] Furthermore, in step S2.2.4, the similarity threshold α is typically set to 60%.

[0023] S3: Calculate the optimal feed rate of the workpiece corresponding to the spindle power data obtained in S1 using the industrial control computer, and save the optimal feed rate and spindle power data as the same data in the database;

[0024] S3.1: Set a constant power reference value Pi for the dimensionality-reduced spindle power data with labels from step S2.2.5. ref A dual-input, single-output fuzzy logic controller is selected as the solution model for the feed rate adaptive control model, used to solve for the feed rate optimization value. The inputs to the fuzzy logic controller are the power deviation e and the power deviation change rate ec. The formula for the power deviation e is: e = P o -Pi ref , where P o This is the dimensionality-reduced spindle power data with labels; the formula for the power deviation change rate (ec) is: Where de is the differential of power deviation and dt is the differential of time;

[0025] S3.2: Use negative large NB, negative medium NM, negative small NS, zero ZO, positive small PS, positive medium PM, and positive large PB linguistic variables to describe power deviation e, power deviation change rate ec, and optimal feed rate f. c e, ec and f c Let the fuzzy universe of discourse be [-p, p], and its fuzzy set be denoted as {NB, NM, NS, ZO, PS, PM, PB}. Trigonometric functions are used as e, ec, and f. c Membership function;

[0026] S3.3: Formulate a fuzzy control rule table, use the Mamdani fuzzy inference method to calculate the fuzzy membership output, and use the area centroid method for defuzzification calculation to obtain the optimal feed rate f. c ;

[0027] S3.4: Save the obtained optimal feed rate and the tagged, dimension-reduced spindle power data as the same data entry into the database;

[0028] Furthermore, in step S3.2, e, ec, and f c In the fuzzy universe of discourse [-p, p], p is a positive integer not greater than 7.

[0029] S4: Adaptive control of machine tool feed rate using industrial control computer: Select the optimal feed speed consistent with the part to be processed from the database on the industrial control computer, use the optimal feed speed to solve the feed rate control value, and send the feed rate control value to the CNC system of the CNC machine tool, thereby realizing adaptive machining of the CNC machine tool;

[0030] S4.1: Let the feed rate control value of the CNC machine tool be M. f The calculation formula is: Among them, f c f is the optimal feed rate stored in the database. o The feed rate set for the machining program;

[0031] S4.2: The industrial computer and the CNC system of the CNC machine tool are connected via network cable for real-time communication;

[0032] S4.3: Adjust the feed rate control value M f The data is sent to the CNC system, which then uses the machine tool's built-in PLC module to achieve adaptive control of the machine tool's feed rate.

[0033] The beneficial effects of this invention are:

[0034] (1) The intelligent adaptive machining method of the present invention can automatically identify the type of part to be processed and select the corresponding control parameters for adaptive machining, thereby improving the intelligence level of adaptive machining of CNC machine tools.

[0035] (2) The intelligent adaptive processing method of the present invention can automatically identify the parts to be processed, which improves the processing efficiency of the parts and saves production time and cost.

[0036] (3) The intelligent adaptive machining method of the present invention can automatically identify the workpiece to be processed, and can adaptively adjust the feed rate in the CNC machine tool machining process, thereby improving the intelligence level of the CNC machine tool. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the power sensor arrangement for an embodiment.

[0038] Figure 2 This is a diagram illustrating the adaptive control effect of the feed rate in Example 1;

[0039] Figure 3 This is a diagram illustrating the adaptive control effect of the feed rate in Example 2;

[0040] Figure 4 This is a flowchart of the method of the present invention.

[0041] In the diagram: 1 is a three-axis vertical machining center; 2 is the spindle servo motor; 3 is the worktable; 4 is the power sensor; and 5 is the spindle. Detailed Implementation

[0042] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings and technical solutions.

[0043] The CNC machine tool used in this embodiment is a three-axis vertical machining center. The control method for other CNC machine tools is the same as in this embodiment.

[0044] Example 1

[0045] An intelligent adaptive machining method that can automatically identify parts to be processed is achieved through the following steps:

[0046] S1: Data acquisition and transmission, specifically, acquiring spindle power data when CNC machine tool is machining part A, and sending the acquired spindle power data to industrial control computer;

[0047] S1 specifically consists of two steps:

[0048] S1.1: As Figure 1 As shown, a power sensor 4 is installed at an appropriate position in the three-axis vertical machining center. In this embodiment, the spindle 5 of the three-axis vertical machining center 1 is located above the worktable 3. A spindle servo motor 2 is connected above the spindle 5. The power sensor 4 is arranged on the spindle servo motor 2. Its three current sensing coils are respectively wrapped around the three power lines U, V and W of the spindle motor. The voltage is connected to the three power lines U, V and W in parallel.

[0049] S1.2: The spindle power data is sent to the industrial control computer via the data transmission line.

[0050] S2: Use an industrial computer to perform a similarity matching comparison between the spindle power data obtained in step S1 and the spindle power data in the database;

[0051] S2 specifically consists of two steps:

[0052] S2.1: Use the Piecewise Aggregate Approximation Algorithm (PAA algorithm) to perform dimensionality reduction on the spindle power data obtained in step S1 to obtain the dimensionality-reduced spindle power data;

[0053] S2.2: Apply the constrained derivative dynamic time bending algorithm (CDDTW algorithm) to perform similarity matching and comparison between the dimension-reduced spindle power data and each spindle power data in the database to determine whether the optimal feed rate of the machined part exists in the database;

[0054] Specifically:

[0055] S2.2.1: Calculate the distance matrix D between the dimensionality-reduced spindle power data in step S2.1 and the first spindle power data in the database using formula (1). In formula (1), ω is 0.8.

[0056] S2.2.2: Solve for the alignment path of the distance matrix D. The alignment path consists of path points in the distance matrix D that satisfy the following conditions: (1) The starting path point of the alignment path is d. 1,1 The end path point is d. m,n (2) In the alignment path, the next path point must be one of the three elements adjacent to the previous path point. These three path points are located to the right, below and to the lower right of the previous path point, respectively. (3) The next path point is the one with the smallest value among the three path points mentioned above. If the three path points have the same value, the path point to the lower right is selected first, then the path point below, and finally the path point to the right.

[0057] S2.2.3: Calculate the alignment path smoothness S of the distance matrix D. The formula for the alignment path smoothness S is: Where, N inc N represents the total number of special path points in the alignment path, characterized by being located to the lower right of the previous path point; tot This represents the total number of all path points in the alignment path;

[0058] S2.2.4: Repeat steps S2.2.1 to S2.2.3 to calculate the alignment path smoothness S between the dimension-reduced spindle power data and each spindle power data in the database. The calculation results are shown in Table 1.

[0059] Table 1A shows the unidirectional path smoothness S of each spindle power data in the part and database.

[0060]

[0061] In this example, the similarity threshold α is set to 60%. The smoothness S of the alignment path in Table 1 is less than 60%. Therefore, it is considered that the optimal feed speed of the processed part A is not in the database.

[0062] S2.2.5: Set the label for the dimensionality-reduced spindle power data to the spindle power data of the 3rd part, and obtain the dimensionality-reduced spindle power data labeled as the spindle power data of the 3rd part;

[0063] S3: Calculate the optimal feed rate of the workpiece corresponding to the spindle power data of the third part in step S2.2.5 using an industrial control computer, and save the optimal feed rate and the spindle power data of the third part as the same data in the database;

[0064] S3 specifically consists of four steps:

[0065] S3.1: Set a constant power reference value Pi for the dimensionality-reduced spindle power data with labels from step S2.2.5. ref A dual-input, single-output fuzzy logic controller is selected as the solution model for the feed rate adaptive control model, used to solve for the feed rate optimization value. The inputs to the fuzzy logic controller are the power deviation e and the power deviation change rate ec. The formula for the power deviation e is: e = P o -Pi ref , where P o This is the dimensionality-reduced spindle power data with labels; the formula for the power deviation change rate (ec) is: Where de is the differential of power deviation and dt is the differential of time;

[0066] S3.2: Use negative large NB, negative medium NM, negative small NS, zero ZO, positive small PS, positive medium PM, and positive large PB linguistic variables to describe power deviation e, power deviation change rate ec, and optimal feed rate f. c e, ec and f c The fuzzy universe of discourse is set to [-6, 6], and its fuzzy set is represented as {NB, NM, NS, ZO, PS, PM, PB}. Trigonometric functions are used as e, ec, and f. c Membership function;

[0067] S3.3: Based on expert experience, a fuzzy control rule table is formulated, as shown in Table 2. The traditional fuzzy inference method (Mamdani fuzzy inference method) is used to calculate the fuzzy membership output, and the area centroid method is used for defuzzification calculation to obtain the optimal feed rate f. c ;

[0068] Table 2. Rule Chart of Fuzzy Control Module

[0069]

[0070] S3.4: Save the obtained optimal feed rate and the tagged, dimension-reduced spindle power data as the same data entry into the database;

[0071] S4: Adaptive control of the feed rate of the machine tool using an industrial control computer: Select the optimal feed speed saved in step S3.4 from the database on the industrial control computer, use the optimal feed speed to solve the feed rate control value, and send the feed rate control value to the CNC system of the CNC machine tool, thereby realizing adaptive machining of the CNC machine tool;

[0072] S4 specifically consists of three steps:

[0073] S4.1: Let the feed rate control value of the CNC machine tool be M. f The calculation formula is: Among them, f c f is the optimal feed rate stored in the database. o The feed rate set for the machining program;

[0074] S4.2: The industrial computer and the CNC system of the CNC machine tool are connected via network cable for real-time communication;

[0075] S4.3: Adjust the feed rate control value M f The data is sent to the CNC system, where the machine tool's built-in PLC module enables adaptive control of the machine tool's feed rate. The adaptive control effect of the feed rate in this embodiment is as follows: Figure 2 As shown.

[0076] The technical effect of this embodiment is as follows: the spindle power data of part A is compared with the data in the database. If the matching fails, the power data of part A and the corresponding optimal feed speed are saved in the database as the spindle power data of the third part. The optimal feed speed corresponding to the spindle power data of the third part is used to adaptively adjust the feed rate of the machine tool, which improves the part processing efficiency by 15.4% and saves production time and costs.

[0077] Example 2

[0078] An intelligent adaptive machining method that can automatically identify parts to be processed is achieved through the following steps:

[0079] S1: Data acquisition and transmission, specifically, acquiring spindle power data when the CNC machine tool is machining part B, and sending the acquired spindle power data to the industrial control computer; S1 is specifically divided into two steps, which are the same as S1 in Embodiment 1;

[0080] S2: The spindle power data obtained in step S1 is compared with the spindle power data in the database using an industrial control computer. S2 specifically consists of two steps: steps S2.1, S2.2.1, S2.2.2, S2.2.3, and S2.2.4 are the same as the corresponding steps in Example 1. The calculation results of the alignment path smoothness S are shown in Table 3.

[0081] Table 3B shows the unipath smoothness S of each spindle power data in the parts and database.

[0082]

[0083] In this example, the similarity threshold α is set to 60%. The alignment path smoothness between the spindle power data of part B after dimensionality reduction in Table 3 and the spindle power data of the third part in the database is greater than 60%. It is considered that the optimal feed rate of the processed part B is the same as the optimal feed rate corresponding to the spindle power data of the third part. Then, step S4 is directly implemented.

[0084] S4: Adaptive control of the machine tool's feed rate using an industrial control computer: The optimal feed speed corresponding to the spindle power data of the third part is selected from the database on the industrial control computer. The optimal feed speed is then used to calculate the feed rate control value, which is sent to the CNC system of the CNC machine tool, thereby achieving adaptive machining of the CNC machine tool. S4 specifically consists of three steps, identical to S4 in Example 1. The adaptive control effect of the feed rate in this embodiment is as follows: Figure 3 As shown.

[0085] The flowchart of the method of the present invention is as follows Figure 4 As shown.

[0086] The technical effect of this embodiment is as follows: by comparing the spindle power data of part B with the data in the database, and after successful matching, the optimal feed rate corresponding to the spindle power data of the third part is used to adaptively adjust the feed rate of the machine tool, thereby improving the part processing efficiency by 14.9% and saving production time and costs.

[0087] It should be noted that the specific embodiments described above are merely illustrative of the principles and processes of the present invention and do not constitute a limitation thereof. Therefore, any modifications and equivalent substitutions made without departing from the spirit and scope of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent adaptive machining method capable of automatically identifying parts to be processed, characterized in that, Here are the steps: S1. Data Acquisition and Transmission: Acquire spindle power data during CNC machine tool machining and send the acquired spindle power data to the industrial control computer; the specific steps of step S1 are as follows: S1: Data Acquisition and Transmission: Acquire spindle power data during CNC machine tool machining of parts, and send the acquired spindle power data to the industrial control computer; S1.1: Install a power sensor at an appropriate location on the spindle controller of the CNC machine tool, and use the power sensor to collect spindle power data during the machining process; S1.2: Send the spindle power data to the industrial computer via the data transmission line; S2: Use the industrial control computer to perform a similarity matching comparison between the spindle power data obtained in step S1 and the spindle power data in the database. If the similarity matching comparison fails, proceed to step S3; if the matching is successful, proceed to step S4. The database is stored in the industrial control computer. The specific steps of step S2 are as follows: S2.1: Use the segmented aggregation approximation algorithm to perform dimensionality reduction on the spindle power data obtained in step S1 to obtain the dimensionality-reduced spindle power data; S2.2: Apply the constrained derivative dynamic time bending algorithm to perform similarity matching and comparison between the dimension-reduced spindle power data and each spindle power data in the database to determine whether the optimal feed rate of the machined part exists in the database; S2.2.1: Calculate the distance matrix D between the dimensionality-reduced spindle power data in step S2.1 and the first spindle power data in the database, and the element d in the i-th row and j-th column of the distance matrix D. i,j The calculation formula is: Where, d i,j Let a be the element in the i-th row and j-th column of the distance matrix D; i ′ The numerical derivative of the i-th data point in the dimensionality-reduced spindle power data is calculated using the following formula: And a ′ 1 = a ′ 2, a ′ m =a ′ m-1 b j ′ The numerical derivative of the j-th data point of the first spindle power data in the database is calculated using the following formula: And b1 ′ =b2 ′ b ′ n =b ′ n-1 ω is the penalty intensity coefficient; m is the number of data points in the dimensionality-reduced spindle power data; n is the number of data points in the first spindle power data in the database; S2.2.2: Solve for the alignment path of the distance matrix D. The alignment path consists of path points in the distance matrix D that satisfy the following conditions: (1) The starting path point of the alignment path is d. 1,1 The end path point is d. m,n (2) In the alignment path, the next path point must be one of the three elements adjacent to the previous path point. These three path points are located to the right, below and to the lower right of the previous path point, respectively. (3) The next path point is the one with the smallest value among the three path points mentioned above. If the three path points have the same value, the path point to the lower right is selected first, then the path point below, and finally the path point to the right. S2.2.3: Calculate the alignment path smoothness S of the distance matrix D. The formula for the alignment path smoothness S is: Where, N inc N represents the total number of special path points in the alignment path, characterized by being located to the lower right of the previous path point; tot This represents the total number of all path points in the alignment path; S2.2.4: Repeat steps S2.2.1 to S2.2.3 to calculate the alignment path smoothness S of the dimensionality-reduced spindle power data and each spindle power data in the database. If there is an alignment path smoothness S greater than the similarity threshold α, it is considered that the optimal feed speed of the machined part exists in the database, and the optimal feed speed corresponding to the spindle power data in the database with the largest alignment path smoothness S is selected as the optimal feed speed of the machined part; if there is no alignment path smoothness S greater than the similarity threshold α, it is considered that the optimal feed speed of the machined part is not in the database. S2.2.5: If the optimal feed rate for the machined part is not in the database, then a label is set for the dimensionality-reduced spindle power data to obtain the dimensionality-reduced spindle power data with the label. S3: Use the industrial control computer to calculate the optimal feed rate of the workpiece corresponding to the spindle power data obtained in step S1, and save the optimal feed rate and spindle power data as the same data in the database; S4: Adaptive control of machine tool feed rate using industrial control computer: Select the optimal feed speed consistent with the part to be processed from the database on the industrial control computer, use the optimal feed speed to solve the feed rate control value, and send the feed rate control value to the CNC system of the CNC machine tool, thereby realizing adaptive machining of the CNC machine tool.

2. The intelligent adaptive machining method for automatically identifying parts to be processed according to claim 1, characterized in that, In step S2.2.1, the penalty intensity coefficient ω is usually set to 0.6 to 0.

8.

3. The intelligent adaptive machining method for automatically identifying parts to be processed according to claim 1, characterized in that, In step S2.2.4, the similarity threshold α is usually set to 60%.

4. The intelligent adaptive machining method for automatically identifying parts to be processed according to claim 1, characterized in that, The specific steps of step S3 are as follows: S3.1: Set a constant power reference value Pi for the dimensionality-reduced spindle power data with labels from step S2.2.

5. ref A dual-input, single-output fuzzy logic controller is selected as the solution model for the feed rate adaptive control model, used to solve for the feed rate optimization value. The inputs to the fuzzy logic controller are the power deviation e and the power deviation change rate ec. The formula for the power deviation e is: e = P o -Pi ref , where P o This is the dimensionality-reduced spindle power data with labels; the formula for the power deviation change rate (ec) is: Where de is the differential of power deviation and dt is the differential of time; S3.2: Use negative large NB, negative medium NM, negative small NS, zero ZO, positive small PS, positive medium PM, and positive large PB linguistic variables to describe power deviation e, power deviation change rate ec, and optimal feed rate f. c e, ec and f c Let the fuzzy universe of discourse be [-p, p], and its fuzzy set be denoted as {NB, NM, NS, ZO, PS, PM, PB}. Trigonometric functions are used as e, ec, and f. c Membership function; S3.3: Formulate a fuzzy control rule table, use the Mamdani fuzzy inference method to calculate the fuzzy membership output, and use the area centroid method for defuzzification calculation to obtain the optimal feed rate f. c ; S3.4: Save the obtained optimal feed rate and the tagged, dimension-reduced spindle power data as the same data entry into the database.

5. The intelligent adaptive machining method for automatically identifying parts to be processed according to claim 4, characterized in that, In step S3.2, e, ec, and f c In the fuzzy universe of discourse [-p, p], p is a positive integer not greater than 7.

6. The intelligent adaptive machining method for automatically identifying parts to be processed according to claim 1, characterized in that, The specific steps of step S4 are as follows: S4.1: Let the feed rate control value of the CNC machine tool be M. f The calculation formula is: Among them, f c f is the optimal feed rate stored in the database. o The feed rate set for the machining program; S4.2: The industrial computer and the CNC system of the CNC machine tool are connected via network cable for real-time communication; S4.3: Adjust the feed rate control value M f The data is sent to the CNC system, which then uses the machine tool's built-in PLC module to achieve adaptive control of the machine tool's feed rate.

Images