Multi-parameter cooperative control method for round bar processing equipment based on artificial intelligence

Abstract

The invention discloses a round bar processing equipment multi-parameter cooperative control method based on artificial intelligence, and particularly relates to the field of round bar processing equipment control, which comprises the steps of data acquisition, data processing, data analysis and result judgment. According to data acquisition, a force sensor, a temperature sensor, a displacement sensor and a vibration sensor of specific models are deployed, data are acquired according to different set frequencies of a high-speed cutting process and a common cutting process, abnormal values of the acquired data are eliminated according to the Pauta criterion in data processing, and Kalman filtering denoising and normalization operation are performed. And in the data analysis step, processed data are substituted into a turning and milling cutting force model, an improved Archard tool wear model and a multiple regression machining precision model to be solved, and according to result judgment, the machining quality is evaluated according to a solving result comparison standard, and the equipment state is judged in combination with a threshold value. According to the method, multi-parameter collaborative optimization is realized, the processing quality stability and efficiency are improved, the dependence on artificial experience is reduced, and the downtime is shortened.

Description

Technical Field

[0001] This invention relates to the field of control technology for round bar processing equipment, and more specifically, to a multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence. Background Technology

[0002] In the field of round bar machining, the process involves multiple parameters such as cutting speed, feed rate, and tool wear. These parameters are interrelated and jointly affect the machining quality and efficiency of round bars. Traditional round bar machining equipment control mainly relies on the operator's experience to manually adjust various parameters. The operator judges the machining situation and adjusts parameters by manually observing phenomena during the machining process, such as cutting sounds and workpiece surface condition.

[0003] However, in practical use, it still has some drawbacks. For example, the existing technology relies on human experience, and the skill levels and judgment standards of different operators vary, resulting in large fluctuations in the processing quality of round bars in the same or different batches, making it difficult to ensure product consistency. In addition, manual parameter adjustment often requires repeated trial and error, which wastes a lot of time in the process of finding the optimal parameter combination. It is also difficult to achieve precise coordinated control of multiple parameters, which cannot give full play to the processing efficiency of the equipment, resulting in low production efficiency. At the same time, the existing technology requires operators to have rich processing experience and high skill levels to complete parameter adjustment and quality judgment, which increases personnel training costs and management difficulty. Furthermore, it is difficult for humans to monitor the operating status of the equipment in real time and accurately, which may lead to problems such as excessive tool wear and equipment failure not being detected and dealt with in a timely manner, increasing equipment maintenance costs and downtime. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence, which solves the problems mentioned in the background art through the following scheme.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence, comprising: S1: data acquisition: deploying force sensors, temperature sensors, displacement sensors and vibration sensors to acquire data according to a set data acquisition frequency;

[0006] S2: Data processing: Based on the data collected in step S1, perform outlier removal, data denoising, and data normalization operations;

[0007] S3: Data Analysis: Substitute the data processed in step S2 into the cutting force model, tool wear model, and machining accuracy model to solve the problem;

[0008] S4: Result Judgment: Based on the solution results of step S3, judge the processing quality and equipment status.

[0009] Preferably, the force sensor is deployed as follows: a high-precision strain gauge force sensor, model HBMU9C, is selected and installed at the part where the tool and workpiece are in direct contact. It is securely connected by a special fixture to ensure accurate acquisition of cutting force data. The sensor has a measurement range of 0-5000N and an accuracy of ±0.1%FS. It can sense changes in cutting force in real time and provide key data for monitoring the stability of the machining process.

[0010] The temperature sensor is deployed as follows: A K-type thermocouple temperature sensor is used, and it is positioned on the cutting tool tip, workpiece surface, and key transmission components of the machine tool. One end of the thermocouple is tightly fitted to the measured area and fixed with high-temperature thermally conductive adhesive to ensure accurate temperature measurement. Its measurement range is -200℃ to 1372℃, with an accuracy of ±1.5℃ or ±0.4%t, effectively monitoring temperature changes caused by friction and cutting heat during machining.

[0011] The deployment method of the displacement sensor is as follows: Select a laser displacement sensor, such as KEYENCELK-G152, and install it on the machine tool worktable. Adjust its position by means of a bracket so that the laser beam emitted by it is perpendicular to the part of the round bar to be processed. The sensor has a measurement range of 0-15mm and a resolution of up to 0.1μm. It can accurately measure the displacement of the round bar during the processing and is used to accurately control the processing size of the round bar.

[0012] The deployment method of the vibration sensor is as follows: a piezoelectric vibration sensor, model PCB352C33, is used and installed in key parts such as the machine tool bed and spindle housing. It is fastened with bolts to ensure that the sensor is tightly coupled with the equipment structure. Its measurement frequency range is 0.5Hz-10kHz and its sensitivity is 100mV / g. It can effectively collect vibration signals during the processing and help judge the operating status of the equipment and the processing quality.

[0013] Preferably, the data acquisition frequency is set as follows: For high-speed cutting processes, the data acquisition frequency of the force sensor and vibration sensor is set to 5000Hz to capture sudden parameter changes during cutting; the temperature sensor acquisition frequency is set to 100Hz to balance the continuity of temperature changes and the pressure of data processing; the displacement sensor acquisition frequency is 500Hz during the rapid change of bar size and reduced to 100Hz during the finishing stage when the size is close to the target value; for ordinary cutting processes, the acquisition frequency of the force sensor and vibration sensor is adjusted to 2000Hz, the temperature sensor to 50Hz, and the displacement sensor to 200Hz during roughing and 50Hz during finishing.

[0014] Preferably, the outlier removal steps are as follows:

[0015] The Raida criterion is used to identify outliers. Let the collected data sequence be x1, x2, ..., x... n First, calculate the mean of the sequence. Then calculate the standard deviation using Bessel's formula. If a data point xj satisfies If xj is identified as an outlier, it is removed from the data sequence. For the missing positions after outlier removal, linear interpolation is used to fill in the gaps, i.e., based on two adjacent valid data points xj. j-1 With x j+1 Calculate supplementary values

[0016] Preferably, the data denoising steps are as follows:

[0017] The Kalman filter algorithm is used to denoise the data. Let the system state equation be x. k =Ax k-1 +Bu k-1 +w k-1 The observation equation is z k =Hx k +v k , where x k Let A be the system state at time k, A be the state transition matrix, B be the control input matrix, and u be the system state at time k. k-1 To control the input, w k-1 For process noise, z k Let v be the observation value at time k, H be the observation matrix, and v be the value at time k. k Initialize the state estimate to observe noise. Given the estimated error covariance P0, at each time k, the prediction step is performed first: P k|k-1 =AP k-1 A T +Q, where Q is the process noise covariance, and then perform the update step:

[0018] K k =P k|k-1 H T HP k|k-1 H T +R) -1 ;

[0019]

[0020] P k =(IK k H)P k|k-1 K kLet R be the Kalman gain, R be the observation noise covariance, and I be the identity matrix. The optimal state estimate after denoising is obtained through iterative calculation. For multi-dimensional data, the above algorithm is extended to a multi-dimensional matrix operation form. The state equations and observation equations of each dimension are constructed independently, while the process noise covariance matrix Q and the observation noise covariance matrix R are shared, and the filtering process is performed in the same way.

[0021] Preferably, the data normalization operation involves the following steps:

[0022] The maximum-minimum normalization method is used to map the data to the interval [0, 1]. Let the original data be x, and the minimum value in the data sequence be x. min The maximum value is x max Then the normalized data For different types of sensor data, determine the x-value of each data sequence. min With x max Perform a normalization operation.

[0023] Preferably, the cutting force model is as follows:

[0024] Based on metal cutting theory, a mathematical model of the relationship between cutting force and multiple factors is constructed. For turning, the empirical formula F is adopted. c =C f ·v a ·f b ·d c ·K, where F c For cutting force, C f is the cutting force coefficient, which is related to the tool material, workpiece material and cutting conditions, and is determined through fitting of a large number of cutting experiments; v is the cutting speed; f is the feed rate; d is the depth of cut; a, b, and c are exponents, which are determined by experiments based on different materials and machining processes; K is a correction coefficient related to the tool and workpiece materials, which is determined by consulting the metal cutting handbook and experimental verification.

[0025] For milling operations, the cutting force model is adjusted to F. c =C f ·v a ·f z b ·a p c ·a e d ·K, where f z a is the feed per tooth. p a is the axial cutting depth. e The radial cutting depth is determined by the specific milling process and material properties.

[0026] Preferably, the tool wear model is as follows:

[0027] The tool wear process is described using an improved Arcard wear model, where the tool wear volume V = K. v ·∫0F c ·v·dt, where K v The wear coefficient is determined by the tool material, workpiece material, and cutting conditions, and is determined experimentally. t is the cutting time, and F... c Let F be the cutting force and v be the cutting speed. In actual calculations, the integration process is discretized. Let the data acquisition interval be Δt. In the i-th time interval, the cutting force is F. c，i The cutting speed is v i Then the tool wear volume increment ΔV i =K v ·F c，i ·v i ·Δt, cumulative tool wear volume n is the time interval number.

[0028] Preferably, the machining accuracy model is as follows:

[0029] Taking the diameter accuracy of a round bar as an example, a model relating machining accuracy to multiple parameters is constructed. Let the diameter accuracy of the round bar be ΔD = f1(v, f, d, T, w), where T is the tool wear and w is the workpiece material inhomogeneity parameter. A multiple regression analysis method is used to determine the specific form of the function f1. Through numerous machining experiments, the diameter D of the round bar after machining under different machining parameters (v, f, d) is collected and compared with the target diameter D0 to obtain the dimensional deviation ΔD = D - D0. A data sample (v...) is then established. i f i d i T i ΔD i For each number i = 1, 2, ..., m, the coefficients β0, β1, β2, β3, β4, and β5 in the multiple linear regression equation ΔD = β0 + β1v + ​​β2f + β3d + β4T + β5w + ∈ are solved using the least squares method. ∈ represents the random error term. For roundness, a model is constructed with roundness error as the dependent variable and machining parameters and tool wear as independent variables. For surface roughness, an empirical model is established based on cutting theory and experimental data, relating surface roughness to parameters such as cutting speed, feed rate, and tool edge radius. Where Ra is the surface roughness, C is a constant, r is the cutting edge radius of the tool, and a1, b1, and c1 are exponents, determined through experiments.

[0030] Preferably, the method for judging the processing quality is as follows:

[0031] The dimensional accuracy, shape accuracy, and surface roughness of the round bar calculated by the machining accuracy model are compared with the pre-set machining quality standards. For the diameter dimensional accuracy of the round bar, let the target diameter be D0 and the tolerance range be [D0-δ, D0+δ]. If the diameter D calculated by the machining accuracy model satisfies D0-δ≤D≤D0+δ, and the shape accuracy and surface roughness both meet the standards, then the machining quality is judged to be qualified; otherwise, the machining quality is judged to be unqualified. When the machining quality is unqualified, the influence weight of each parameter on the unqualified index is analyzed by the machining accuracy model to determine the key parameters affecting the machining quality.

[0032] The method for determining the device status is as follows:

[0033] The tool wear amount calculated based on the tool wear model is compared with a preset tool wear threshold. Simultaneously, data collected from vibration and temperature sensors is used to determine the equipment's operating status. The tool wear threshold is set to 20% of the initial tool thickness. When the tool wear amount calculated by the tool wear model reaches the threshold, a tool replacement warning is issued. Regarding equipment vibration, the vibration amplitude range during normal operation is set to [A]. min A max When the vibration amplitude collected by the vibration sensor exceeds this range, such as greater than 1.5Amax, the equipment vibration is determined to be abnormal, processing is suspended, and a comprehensive inspection and troubleshooting of the equipment is carried out. For the equipment temperature status, the normal operating temperature range of key components is set. When the temperature sensor measurement value exceeds the corresponding upper limit, the temperature is determined to be abnormal, cooling measures are taken, and the working condition of the cooling system is checked.

[0034] The technical effects and advantages of this invention are as follows:

[0035] 1. This invention collects data in real time by deploying multiple high-precision sensors, and uses mathematical models such as the established cutting force model, tool wear model, and machining accuracy model for data analysis and parameter coordinated control. This avoids the subjective differences of human experience, and can accurately adjust various parameters to ensure the consistency and stability of the round bar machining quality.

[0036] 2. This invention eliminates the need for repeated manual trial and error. The system can automatically optimize processing parameters based on real-time data and model analysis results, achieving coordinated optimization control of multiple parameters, fully utilizing the processing capacity of the equipment, effectively shortening processing time, and improving production efficiency.

[0037] 3. The entire control process of this invention is completed automatically by the system. Operators only need to perform simple monitoring and auxiliary operations, without the need for extensive experience and superb skills, which reduces personnel training costs and management difficulties.

[0038] 4. This invention monitors the equipment's operating status in real time using vibration sensors, temperature sensors, etc., and accurately predicts tool wear by combining tool wear models. It can issue early warnings in a timely manner and take corresponding measures to reduce equipment failures and downtime, and lower maintenance costs. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] As attached Figure 1 The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence, as shown, includes:

[0042] S1: Data Acquisition: Deploy force sensors, temperature sensors, displacement sensors, and vibration sensors to acquire data according to the set data acquisition frequency;

[0043] The deployment method of the force sensor is as follows: A high-precision strain gauge force sensor, model HBMU9C, is selected and installed at the direct contact point between the tool and the workpiece. It is securely connected by a special fixture to ensure accurate acquisition of cutting force data. The sensor has a measurement range of 0-5000N and an accuracy of ±0.1%FS. It can sense changes in cutting force in real time and provide key data for monitoring the stability of the machining process.

[0044] The temperature sensor is deployed as follows: A K-type thermocouple temperature sensor, such as the OMEGAKMT-010, is used, positioned at the tool tip, workpiece surface, and key machine tool transmission components (such as spindle bearings and lead screw nut pairs). One end of the thermocouple is tightly fitted to the measured area and fixed with high-temperature thermally conductive adhesive to ensure accurate temperature measurement. Its measurement range is -200℃ to 1372℃, with an accuracy of ±1.5℃ or ±0.4%t (whichever is greater), effectively monitoring temperature changes caused by friction and cutting heat during machining.

[0045] The deployment method of the displacement sensor is as follows: Select a laser displacement sensor, such as KEYENCELK-G152, and install it on the machine tool worktable. Adjust its position by means of a bracket so that the laser beam emitted by it is perpendicular to the part of the round bar to be processed. The sensor has a measurement range of 0-15mm and a resolution of up to 0.1μm. It can accurately measure the displacement of the round bar during the processing and is used to accurately control the processing size of the round bar.

[0046] The deployment method of the vibration sensor is as follows: a piezoelectric vibration sensor, model PCB352C33, is used and installed in key parts such as the machine tool bed and spindle housing. It is fastened with bolts to ensure that the sensor is tightly coupled with the equipment structure. Its measurement frequency range is 0.5Hz-10kHz and its sensitivity is 100mV / g. It can effectively collect vibration signals during the processing and help judge the operating status of the equipment and the processing quality.

[0047] The data acquisition frequencies are set as follows: For high-speed cutting processes, the data acquisition frequency of the force sensor and vibration sensor is set to 5000Hz to capture sudden parameter changes during cutting. The temperature sensor acquisition frequency is set to 100Hz to balance the continuity of temperature changes and the data processing pressure. The displacement sensor acquisition frequency is 500Hz during the rapid change of bar size (such as the early stage of roughing) and reduced to 100Hz during the finishing stage when the size is close to the target value. For ordinary cutting processes, the acquisition frequencies of the force sensor and vibration sensor are adjusted to 2000Hz, the temperature sensor to 50Hz, and the displacement sensor to 200Hz during roughing and 50Hz during finishing. The specific frequency values ​​are determined through multiple experiments based on the processing standards and the actual dynamic response characteristics of the equipment to balance data integrity and processing burden.

[0048] S2: Data processing: Based on the data collected in step S1, perform outlier removal, data denoising, and data normalization operations;

[0049] The outlier removal steps are as follows:

[0050] The Raida criterion is used to identify outliers. Let the collected data sequence be x1, x2, ..., x... n First, calculate the mean of the sequence. Then calculate the standard deviation using Bessel's formula. If a data point xj satisfies If xj is identified as an outlier, it is removed from the data sequence. For the missing positions after outlier removal, linear interpolation is used to fill in the gaps, i.e., based on two adjacent valid data points xj. j-1 With x j+1 Calculate supplementary values

[0051] The data denoising steps are as follows:

[0052] The Kalman filter algorithm is used to denoise the data. Taking one-dimensional data as an example, let the system state equation be x. k =Ax k-1 +Bu k-1 +w k-1 The observation equation is z k =Hx k +v k , where x k Let u be the system state at time k, A be the state transition matrix, B be the control input matrix (since there is no external control input during the processing, B = 0), and u be the system state at time k. k-1 To control the input (set to 0), w k-1 For process noise, z k Let v be the observation value at time k, H be the observation matrix, and v be the value at time k. k To observe the noise, initialize the state estimate x0 and the estimation error covariance P0. At each time k, first perform the prediction step: P k|k-1 =AP k-1 A T +Q, where Q is the process noise covariance, and then perform the update step:

[0053] K k =P k|k-1 H T HP k|k-1 H T +R) -1 ;

[0054]

[0055] P k =(IK k H)P k|k-1 Where Kk is the Kalman gain, R is the observation noise covariance, and I is the identity matrix, the denoised optimal state estimate is obtained through iterative calculation. For multi-dimensional data, such as three-dimensional force data collected by force sensors, the above algorithm is extended to a multi-dimensional matrix operation form. The state equations and observation equations of each dimension are constructed independently, while the process noise covariance matrix Q and the observation noise covariance matrix R are shared, and the filtering process is performed in the same way.

[0056] The steps of the data normalization operation are as follows:

[0057] The maximum-minimum normalization method is used to map the data to the interval [0, 1]. Let the original data be x, and the minimum value in the data sequence be x. min The maximum value is x max Then the normalized data For different types of sensor data, determine the x-value of each data sequence. min With x max Normalization is performed. For example, if the force sensor data has a minimum value of 100N and a maximum value of 3000N over a certain period of time, and the collected force data is 1500N, the normalized value is approximately 1400 / 2900≈0.517.

[0058] S3: Data Analysis: Substitute the data processed in step S2 into the cutting force model, tool wear model, and machining accuracy model to solve the problem;

[0059] The cutting force model is as follows:

[0060] Based on metal cutting theory, a mathematical model of the relationship between cutting force and multiple factors is constructed. For turning, the empirical formula F is adopted. c =C f ·v a ·f b ·d c ·K, where F c C represents the cutting force (unit: N). f The cutting force coefficient is related to the tool material, workpiece material, and cutting conditions. It is determined through fitting of numerous cutting experiments. v is the cutting speed (unit: m / s), f is the feed rate (unit: mm / r), d is the depth of cut (unit: mm), and a, b, and c are exponents determined experimentally based on different materials and machining processes. For example, when machining 45# steel using carbide tools, a = 0.2, b = 0.75, and c = 1. K is a correction coefficient related to the tool and workpiece materials, determined by consulting metal cutting handbooks and experimental verification.

[0061] It should be further noted that, for milling operations, the cutting force model is adjusted to F. c =C f ·v a ·f z b ·a p c ·a e d ·K, where f z a is the feed per tooth (unit: mm / z). p a is the axial cutting depth (unit: mm). e The radial cutting depth (unit: mm) is given, and the values ​​of each coefficient and exponent are determined based on the specific milling process and material properties.

[0062] The tool wear model is as follows:

[0063] The tool wear process is described using an improved Arcard wear model, where the tool wear volume V = K. v ·∫0F c ·v·dt, where K v The wear coefficient is determined by the tool material, workpiece material, and cutting conditions, and is determined experimentally. t is the cutting time (in seconds). F c Let F be the cutting force and v be the cutting speed. In actual calculations, the integration process is discretized. Let the data acquisition interval be Δt. In the i-th time interval, the cutting force is F. c，i The cutting speed is v i Then the tool wear volume increment ΔV i =K v ·F c，i ·v i ·Δt, cumulative tool wear volume n is the time interval number;

[0064] The machining accuracy model is as follows:

[0065] Taking the diameter accuracy of a round bar as an example, a multi-parameter relationship model between machining accuracy and parameters is constructed. Let the diameter accuracy of the round bar be ΔD = f1(v, f, d, T, w), where T is the tool wear amount (calculated using a tool wear model), and w is the workpiece material inhomogeneity parameter (quantified by measuring the hardness distribution of the workpiece material, etc., and set as a fixed correction value). The specific form of the function f1 is determined using multiple regression analysis. Through numerous machining experiments, the diameter D of the round bar after machining under different machining parameters (v, f, d) is collected and compared with the target diameter D0 to obtain the dimensional deviation ΔD = D - D0. A data sample (v...) is then established. i f i d i T i ΔD i Given a set of numbers i = 1, 2, ..., m (where m is the number of experimental samples), the coefficients β0, β1, p2, β3, β4, and β5 in the multiple linear regression equation ΔD = β0 + β1v + ​​β2f + β3d + β4T + β5w + ∈ are solved using the least squares method. These coefficients represent random error terms. For roundness, a similar method is used, with roundness error as the dependent variable and machining parameters and tool wear as independent variables to construct a model. For surface roughness, based on cutting theory and experimental data, an empirical model is established that relates surface roughness to parameters such as cutting speed, feed rate, and tool edge radius. Where Ra is the surface roughness (unit: μm), C is a constant, r is the cutting edge radius of the tool (unit: mm), and a1, b1, and c1 are exponents, determined through experiments;

[0066] It should be further explained that the solution process of the model is as follows: The model calculation process is implemented using computer programming, such as writing a program in Python and calling mathematical computing libraries such as NumPy and SciPy to perform matrix operations and solve equations. Taking the cutting force model as an example, the inputs are the real-time collected and processed cutting speed v, feed rate f, depth of cut d, and the determined cutting force coefficient C. f The current cutting force F is calculated using indices a, b, c and correction factor K. c By analyzing the calculation results of each model, the influence of machining parameters on machining quality and efficiency is obtained. For example, by fixing the feed rate and depth of cut, and gradually changing the cutting speed, the trends of cutting force, tool wear, and surface roughness are analyzed as the cutting speed increases from 10 m / s to 30 m / s. The cutting speed range that maximizes machining efficiency while ensuring machining quality is identified. By plotting parameter-result relationship curves (such as cutting speed-cutting force curve, cutting speed-tool wear curve, and cutting speed-surface roughness curve), the influence patterns are visually displayed, providing a basis for subsequent adjustment of machining parameters.

[0067] S4: Result Judgment: Judge the processing quality and equipment status based on the solution results of step S3;

[0068] It should be noted that the method for judging the processing quality is as follows:

[0069] The dimensional accuracy, shape accuracy, and surface roughness of the round bar calculated by the machining accuracy model are compared with the pre-set machining quality standards. For the diameter dimensional accuracy of the round bar, let the target diameter be D0 and the tolerance range be [D0-δ, D0+δ] (δ is the tolerance value, determined according to the machining accuracy level, such as δ=0.016mm for a round bar with a diameter of 50mm and an IT6 accuracy level). If the diameter D calculated by the machining accuracy model satisfies D0-δ≤D≤D0+δ, and the shape accuracy (roundness error is less than the specified value, such as 0.005mm) and surface roughness (Ra value is less than the specified value, such as 1.6I^1 / 4m) both meet the standards, then the machining quality is judged to be qualified; otherwise, the machining quality is judged to be unqualified. When the machining quality is unqualified, the influence weight of each parameter on the unqualified index is analyzed by the machining accuracy model to determine the key parameters affecting the machining quality. For example, if the diameter is too large, the model analysis finds that the feed rate has the greatest influence on the dimensional deviation, and the current feed rate is outside the reasonable range, then the feed rate is determined to be the key adjustment parameter.

[0070] The method for determining the device status is as follows:

[0071] The tool wear amount calculated based on the tool wear model is compared with a preset tool wear threshold. Simultaneously, data collected from vibration and temperature sensors is used to determine the equipment's operating status. The tool wear threshold is set to 20% of the initial tool thickness. When the tool wear amount calculated by the tool wear model reaches this threshold, a tool replacement warning is issued. Regarding equipment vibration, the vibration amplitude range during normal operation is set to [A]. min A max (Determined through equipment factory standards and extensive test data during normal operation, such as Amin = 0.05g, Amax = 0.5g, where g is the acceleration due to gravity). When the vibration amplitude collected by the vibration sensor exceeds this range, such as greater than 1.5Amax, the equipment vibration is judged to be abnormal, processing is suspended, and a comprehensive inspection and troubleshooting of the equipment is carried out, including checking whether the tool installation is loose, whether the spindle bearing is damaged, and whether there is resonance in the machine tool bed. For the equipment temperature status, the normal operating temperature range of key components is set, such as the upper limit of the tool operating temperature is 600℃ and the upper limit of the spindle bearing operating temperature is 800℃. When the temperature sensor measurement value exceeds the corresponding upper limit, the temperature is judged to be abnormal, and cooling measures are taken (such as turning on the cooling system, suspending processing and waiting for the temperature to drop), and checking whether the cooling system is working properly.

[0072] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0073] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence, characterized in that, include: S1: Data Acquisition: Deploy force sensors, temperature sensors, displacement sensors, and vibration sensors to acquire data according to the set data acquisition frequency; S2: Data processing: Based on the data collected in step S1, perform outlier removal, data denoising, and data normalization operations; S3: Data Analysis: Substitute the data processed in step S2 into the cutting force model, tool wear model, and machining accuracy model to solve the problem; S4: Result Judgment: Based on the solution results of step S3, judge the processing quality and equipment status.

2. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The deployment method of the force sensor is as follows: A high-precision strain gauge force sensor, model HBMU9C, is selected and installed at the direct contact point between the tool and the workpiece. It is securely connected by a special fixture to ensure accurate acquisition of cutting force data. The sensor has a measurement range of 0-5000N and an accuracy of ±0.1%FS. It can sense changes in cutting force in real time and provide key data for monitoring the stability of the machining process. The temperature sensor is deployed as follows: K-type thermocouple temperature sensors are used and arranged on the tool tip, workpiece surface, and key transmission components of the machine tool. One end of the thermocouple is tightly attached to the part being measured and fixed with high-temperature thermally conductive adhesive to ensure accurate temperature measurement. Its measurement range is -200℃ to 1372℃, and its accuracy is ±1.5℃ or ±0.4%t. It can effectively monitor temperature changes caused by friction and cutting heat during the processing. The deployment method of the displacement sensor is as follows: Select a laser displacement sensor, such as KEYENCELK-G152, and install it on the machine tool worktable. Adjust its position by means of a bracket so that the laser beam emitted by it is perpendicular to the part of the round bar to be processed. The sensor has a measurement range of 0-15mm and a resolution of up to 0.1μm. It can accurately measure the displacement of the round bar during the processing and is used to accurately control the processing size of the round bar. The deployment method of the vibration sensor is as follows: a piezoelectric vibration sensor, model PCB352C33, is used and installed in key parts such as the machine tool bed and spindle housing. It is fastened with bolts to ensure that the sensor is tightly coupled with the equipment structure. Its measurement frequency range is 0.5Hz-10kHz and its sensitivity is 100mV / g. It can effectively collect vibration signals during the processing and help judge the operating status of the equipment and the processing quality.

3. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The data acquisition frequencies are set as follows: For high-speed cutting processes, the data acquisition frequency of the force sensor and vibration sensor is set to 5000Hz to capture sudden parameter changes during cutting. The temperature sensor acquisition frequency is set to 100Hz to balance the continuity of temperature changes and the pressure of data processing. The displacement sensor acquisition frequency is 500Hz during the rapid change of the bar size and is reduced to 100Hz during the finishing stage when the size is close to the target value. For ordinary cutting processes, the acquisition frequencies of the force sensor and vibration sensor are adjusted to 2000Hz, the temperature sensor to 50Hz, and the displacement sensor to 200Hz during roughing and 50Hz during finishing.

4. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The outlier removal steps are as follows: Outlier identification is performed using the Laida criterion. Let the collected data sequence be x1, x2, ..., x... n First, calculate the mean of the sequence. Then calculate the standard deviation using Bessel's formula. If a data point xj satisfies If xj is identified as an outlier, it is removed from the data sequence. For the missing positions after outlier removal, linear interpolation is used to fill in the gaps, i.e., based on two adjacent valid data points xj. j-1 With x j+1 Calculate supplementary values 5. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The data denoising steps are as follows: The Kalman filter algorithm is used to denoise the data. Let the system state equation be x. k =Ax k-1 +Bu k-1 +w k-1 The observation equation is z k =Hx k +v k , where x k Let A be the system state at time k, A be the state transition matrix, B be the control input matrix, and u be the system state at time k. k-1 To control the input, w k-1 For process noise, z k Let v be the observation value at time k, H be the observation matrix, and v be the value at time k. k Initialize the state estimate to observe noise. Given the estimated error covariance P0, at each time k, the prediction step is performed first: P k|k-1 =AP k-1 A T +Q, where Q is the process noise covariance, and then perform the update step: K k =P k|k-1 H T (HP k|k-1 H T +R) -1 ; P k =(IK k H)P k|k-1 K k Let R be the Kalman gain, R be the observation noise covariance, and I be the identity matrix. The optimal state estimate after denoising is obtained through iterative calculation. For multi-dimensional data, the above algorithm is extended to a multi-dimensional matrix operation form. The state equations and observation equations of each dimension are constructed independently, while the process noise covariance matrix Q and the observation noise covariance matrix R are shared, and the filtering process is performed in the same way.

6. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The steps of the data normalization operation are as follows: The maximum-minimum normalization method is used to map the data to the interval [0, 1]. Let the original data be x, and the minimum value in the data sequence be x. min The maximum value is x max Then the normalized data For different types of sensor data, determine the x-value of each data sequence. min With x max Perform a normalization operation.

7. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The cutting force model is as follows: Based on metal cutting theory, a mathematical model of the relationship between cutting force and multiple factors is constructed. For turning, the empirical formula F is adopted. c =C f ·v a ·f b ·d c ·K, where F c For cutting force, C f is the cutting force coefficient, which is related to the tool material, workpiece material and cutting conditions, and is determined through fitting of a large number of cutting experiments; v is the cutting speed; f is the feed rate; d is the depth of cut; a, b, and c are exponents, which are determined by experiments based on different materials and machining processes; K is a correction coefficient related to the tool and workpiece materials, which is determined by consulting the metal cutting handbook and experimental verification. For milling operations, the cutting force model is adjusted to F. c =C f ·v a ·f z b ·a p c ·a e d ·K, where f z a is the feed per tooth. p a is the axial cutting depth. e The radial cutting depth is determined by the specific milling process and material properties.

8. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The tool wear model is as follows: The tool wear process is described using an improved Archard wear model, and the tool wear volume is... Where K v The wear coefficient is determined by the tool material, workpiece material, and cutting conditions, and is determined experimentally. t is the cutting time, and F... c Let F be the cutting force and v be the cutting speed. In actual calculations, the integration process is discretized. Let the data acquisition interval be Δt. In the i-th time interval, the cutting force is F. c，i The cutting speed is v i Then the tool wear volume increment ΔV i =K v ·F c，i ·v i ·Δt, cumulative tool wear volume n is the time interval number.

9. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The machining accuracy model is as follows: Taking the diameter accuracy of a round bar as an example, a model relating machining accuracy to multiple parameters is constructed. Let the diameter accuracy of the round bar be ΔD = f1(v, f, d, T, w), where T is the tool wear and w is the workpiece material inhomogeneity parameter. A multiple regression analysis method is used to determine the specific form of the function f1. Through numerous machining experiments, the diameter D of the round bar after machining under different machining parameters (v, f, d) is collected and compared with the target diameter D0 to obtain the dimensional deviation ΔD = D - D0. A data sample (v...) is then established. i f i d i T i , ΔD i For each number i = 1, 2, ..., m, the coefficients β0, β1, β2, β3, β4, and β5 in the multiple linear regression equation ΔD = β0 + β1v + ​​β2f + β3d + β4T + β5w + ∈ are solved using the least squares method. ∈ represents the random error term. For roundness, a model is constructed with roundness error as the dependent variable and machining parameters and tool wear as independent variables. For surface roughness, an empirical model is established based on cutting theory and experimental data, relating surface roughness to parameters such as cutting speed, feed rate, and tool edge radius. Where Ra is the surface roughness, C is a constant, r is the cutting edge radius of the tool, and a1, b1, and c1 are exponents, determined through experiments.

10. The multi-parameter collaborative control method for round bar processing equipment based on artificial intelligence according to claim 1, characterized in that: The method for judging the processing quality is as follows: The dimensional accuracy, shape accuracy, and surface roughness of the round bar calculated by the machining accuracy model are compared with the pre-set machining quality standards. For the diameter dimensional accuracy of the round bar, let the target diameter be D0 and the tolerance range be [D0-δ, D0+δ]. If the diameter D calculated by the machining accuracy model satisfies D0-δ≤D≤D0+δ, and the shape accuracy and surface roughness both meet the standards, then the machining quality is judged to be qualified; otherwise, the machining quality is judged to be unqualified. When the machining quality is unqualified, the influence weight of each parameter on the unqualified index is analyzed by the machining accuracy model to determine the key parameters affecting the machining quality. The method for determining the device status is as follows: The tool wear amount calculated based on the tool wear model is compared with a preset tool wear threshold. Simultaneously, data collected from vibration and temperature sensors is used to determine the equipment's operating status. The tool wear threshold is set to 20% of the initial tool thickness. When the tool wear amount calculated by the tool wear model reaches the threshold, a tool replacement warning is issued. Regarding equipment vibration, the vibration amplitude range during normal operation is set to [A]. min A max When the vibration amplitude collected by the vibration sensor exceeds this range, such as greater than 1.5Amax, the equipment vibration is determined to be abnormal, processing is suspended, and a comprehensive inspection and troubleshooting of the equipment is carried out. For the equipment temperature status, the normal operating temperature range of key components is set. When the temperature sensor measurement value exceeds the corresponding upper limit, the temperature is determined to be abnormal, cooling measures are taken, and the working condition of the cooling system is checked.

Images