Program method for automatically slitting and manufacturing steel wire mesh cloth

Abstract

The invention relates to the technical field of advanced manufacturing, and discloses a program method for automatically slitting and manufacturing steel wire mesh cloth, the method comprises a multi-modal data acquisition module, an intelligent decision module, a dynamic execution module and a closed-loop feedback module, a multi-modal sensing fusion mechanism is constructed, and when the steel wire mesh cloth is automatically slit, a multi-modal sensing fusion mechanism is established, so that the automatic slitting of the steel wire mesh cloth is realized. Based on cooperative sensing of a millimeter-wave radar, an infrared thermal imager and a stress sensor, a net cloth comprehensive state matrix is generated in real time, the coupling relation of material thickness distribution, temperature gradient and stress states can be comprehensively captured, the technical bottleneck that a traditional single sensor cannot comprehensively sense material characteristics is solved, and a dynamic compensation decision-making system is established, so that the material quality is improved. When high-speed slitting is carried out, based on real-time calculation of the tension fluctuation coefficient and a fuzzy control strategy, self-adaptive adjustment of the slitting speed is achieved, and the technical defect that response of a traditional system is lagged under the high-speed working condition is overcome.

Description

Technical Field

[0001] This invention relates to the field of advanced manufacturing technology, specifically to a process method for automatically cutting and manufacturing steel wire mesh. Background Technology

[0002] Steel wire mesh is a general term for mesh materials woven and welded from materials such as low-carbon steel wire, medium-carbon steel wire, high-carbon steel wire, and stainless steel wire. As a key material in fields such as building reinforcement and protective isolation, the cutting accuracy of steel wire mesh directly affects the quality and stability of the end product.

[0003] Traditional slitting processes mainly suffer from the following technical bottlenecks:

[0004] 1. High reliance on manual experience: Existing slitting equipment mainly relies on operators to set slitting parameters based on experience, which is difficult to adapt to the differences in material properties of different batches of mesh fabric. When encountering tension fluctuations and uneven thickness, defects such as burrs and cut deformation often occur, leading to an increase in scrap rate.

[0005] 2. Insufficient dynamic response capability: using only a single photoelectric sensor, it is unable to detect changes in the internal stress state of the material. The cutting parameters remain fixed and cannot be dynamically adjusted according to real-time tension fluctuations. There is a lack of quality closed-loop feedback, and defect detection lags behind the production process.

[0006] 3. The lack of intelligent parameter decision-making makes it impossible to handle the cutting characteristics of new composite mesh fabrics. When there are sudden changes in local thickness of the mesh fabric, the cutter trajectory compensation lags behind. The separation of quality inspection and process adjustment leads to repeated defects. Summary of the Invention

[0007] (a) Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a program method for automatically cutting and manufacturing steel wire mesh, thus solving the problems mentioned in the background section.

[0009] (II) Technical Solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a method for automatically cutting and manufacturing steel wire mesh, the method comprising the following steps:

[0011] S1. Collect material property data and dynamic stress state data of steel wire mesh based on multi-source sensing devices;

[0012] S2. Perform real-time fusion processing on the material property data and dynamic stress state data to generate a comprehensive state matrix of the mesh fabric.

[0013] S3. Based on a deep convolutional network, perform segmentation parameter matching processing on the comprehensive state matrix of the mesh to generate a target segmentation process parameter set;

[0014] S4. Dynamically compensate the cutter head trajectory parameters of the slitting equipment according to the target slitting process parameter set, and generate three-dimensional slitting path planning data;

[0015] S5. Real-time monitoring of mesh tension fluctuation data based on laser displacement sensing device; when the fluctuation threshold exceeds the preset range, step S6 is triggered.

[0016] S6. Based on the tension fluctuation data, adaptively adjust the cutting speed to generate a dynamic cutting rate correction command.

[0017] S7. Based on the machine vision system, perform quality defect detection on the edge of the cut mesh, generate a cutting quality assessment report and feed it back to step S3.

[0018] Preferably, step S1 includes the following steps:

[0019] S11. Scan the surface of the mesh fabric using a millimeter-wave radar sensor at the 800MHz frequency band to generate thickness distribution cloud map data;

[0020] S12. Use an infrared thermal imager to capture temperature gradient data during the movement of the mesh fabric at a sampling frequency of 10Hz.

[0021] S13. Embed a piezoelectric stress sensor array in the slitting roller shaft to collect real-time data on the warp and weft tension ratios of the mesh fabric.

[0022] Preferably, step S2 includes the following steps:

[0023] S21. Establish a spatiotemporal alignment model and perform pixel-level registration between thickness distribution cloud map data and temperature gradient data;

[0024] S22. Frequency domain features are extracted from the tension ratio data using the wavelet packet decomposition algorithm;

[0025] S23. Construct a three-dimensional tensor data structure to store the registered multi-source data:

[0026]

[0027] Thickness ij is the mesh thickness at grid point (i,j), temperature ij is the surface temperature at grid point (i,j), tension k is the tension value of the k-th sensor channel, i,j are the two-dimensional spatial coordinate indices, m is the grid row resolution, and n is the total number of sensor channels.

[0028] Preferably, step S3 includes the following steps:

[0029] S31. Load the pre-trained ResNet-34 network model, and set the input layer to a 128×128×3 convolutional kernel;

[0030] S32. Embed an attention mechanism module in the fully connected layer of the network to calculate the weight allocation for different material features:

[0031]

[0032] α i W is the attention weight for the i-th feature. i T is a trainable weight vector. i Let be the input feature vector, and exp(·) be the natural exponential function.

[0033] S33. The output layer connects to the decision tree classifier to generate a parameter set containing the tool tip angle, feed rate, and pressure value.

[0034] Preferably, step S4 includes the following steps:

[0035] S41. Generate basic slicing paths based on B-spline curve fitting algorithm;

[0036] S42. Dynamically adjust the Z-axis compensation of the tool head based on the material thickness gradient data:

[0037]

[0038] Where k is the material thickness gradient compensation coefficient. Let b be the gradient of the thickness in the X direction, and b be the basic compensation constant.

[0039] S43. Insert a circular arc transition segment at the turning point of the path, with the radius of curvature R satisfying:

[0040]

[0041] Where v is the linear velocity of the cutting head, μ is the friction coefficient between the mesh and the cutting tool, g is the acceleration due to gravity, and δ 安全 For safety margin.

[0042] Preferably, step S5 includes the following steps:

[0043] S51. Deploy two sets of laser Doppler velocimeters upstream and downstream of the slitting line;

[0044] S52. Calculate the tension fluctuation coefficient:

[0045]

[0046] v 上游 v 下游 To determine the speed of mesh deployment along the online and offline lines;

[0047] S53. When η is greater than 0.15, the dynamic adjustment mechanism is triggered.

[0048] Preferably, step S6 includes the following steps:

[0049] S61. Establish a fuzzy PID controller, with the input variable being the value of η and the output variable being the servo motor speed correction amount;

[0050] S62. Set the membership function:

[0051] Low fluctuation domain: η∈[0,0.05) → trigonometric function;

[0052] In the mid-wave domain: η∈[0.05,0.15) → Gaussian function;

[0053] High fluctuation domain: η∈[0.15,0.3] → S-shaped function.

[0054] Preferably, step S7 includes the following steps:

[0055] S71. A linear CCD camera is used to scan the edge of the mesh fabric at a resolution of 5μm.

[0056] S72. Extract burr feature parameters using a morphological edge detection algorithm;

[0057] S73. A quality alarm signal is generated when the length of the burr exceeds 3 pixels continuously.

[0058] Preferably, the step S7 is followed by the following step:

[0059] S8. Self-optimization of process parameters based on the slitting quality assessment report:

[0060] S81. Extract the corresponding material property data for the section with excessive burr length;

[0061] S82. Optimize the tool tip tilt angle parameters using the gradient descent algorithm:

[0062]

[0063] Where η is the learning rate coefficient. The burr length is the partial derivative of the burr length with respect to the tilt angle;

[0064] S83. If the burr length still exceeds the standard after three consecutive optimizations, a tool wear warning signal is triggered.

[0065] Preferably, S83 includes:

[0066] S831. Establish a tool tip life prediction model:

[0067]

[0068] Where β is the material wear coefficient and N is the cumulative number of cuts;

[0069] S832. When an early warning signal is triggered, the current working parameters are automatically recorded and a maintenance log is generated.

[0070] S833. Train a random forest regression model based on historical maintenance data to predict the optimal tool change cycle:

[0071]

[0072] Where α is the weighting factor, γ is the safety factor, σ is the periodic standard deviation, and t i For the historical replacement cycle, σ t The periodic standard deviation is given.

[0073] (III) Beneficial Effects

[0074] Compared with the prior art, the present invention provides a program method for automatic cutting and manufacturing of wire mesh, which has the following beneficial effects:

[0075] 1. In this invention, by establishing a dynamic compensation decision system, the cutting speed is adaptively adjusted based on the real-time calculation of the tension fluctuation coefficient and the fuzzy control strategy during high-speed cutting, thus solving the technical defect of the traditional system's response lag under high-speed conditions.

[0076] 2. In this invention, by constructing a multimodal sensing fusion mechanism, when automatically cutting steel wire mesh, a comprehensive state matrix of the mesh is generated in real time based on the collaborative perception of millimeter-wave radar, infrared thermal imager and stress sensor. This can comprehensively capture the coupling relationship between material thickness distribution, temperature gradient and stress state, and solve the technical bottleneck that traditional single sensors cannot fully perceive material properties.

[0077] 3. In this invention, by deploying a closed-loop quality optimization system, after the slitting is completed, the cutting head parameters are optimized in real time based on machine vision burr detection data and intelligent optimization algorithms, thus solving the technical problem of separating quality inspection and process adjustment. Attached Figure Description

[0078] Figure 1 This is a flowchart of the overall method of the present invention. Detailed Implementation

[0079] 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.

[0080] Please see Figure 1 The automatic cutting and manufacturing process of steel wire mesh includes the following steps:

[0081] S1. Collect material property data and dynamic stress state data of steel wire mesh based on multi-source sensing devices;

[0082] S2. Perform real-time fusion processing on the material property data and dynamic stress state data to generate a comprehensive state matrix of the mesh fabric.

[0083] S3. Based on a deep convolutional network, perform segmentation parameter matching processing on the comprehensive state matrix of the mesh to generate a target segmentation process parameter set;

[0084] S4. Dynamically compensate the cutter head trajectory parameters of the slitting equipment according to the target slitting process parameter set, and generate three-dimensional slitting path planning data;

[0085] S5. Real-time monitoring of mesh tension fluctuation data based on laser displacement sensing device; when the fluctuation threshold exceeds the preset range, step S6 is triggered.

[0086] S6. Based on the tension fluctuation data, adaptively adjust the cutting speed to generate a dynamic cutting rate correction command.

[0087] S7. Based on the machine vision system, perform quality defect detection on the edge of the cut mesh, generate a cutting quality assessment report and feed it back to step S3.

[0088] S1 includes the following steps:

[0089] S11. Scan the surface of the mesh fabric using a millimeter-wave radar sensor at the 800MHz frequency band to generate thickness distribution cloud map data;

[0090] S12. Use an infrared thermal imager to capture temperature gradient data during the movement of the mesh fabric at a sampling frequency of 10Hz.

[0091] S13. Embed a piezoelectric stress sensor array in the slitting roller shaft to collect real-time warp and weft tension ratio data of the mesh.

[0092] S2 includes the following steps:

[0093] S21. Establish a spatiotemporal alignment model and perform pixel-level registration between thickness distribution cloud map data and temperature gradient data;

[0094] S22. Frequency domain features are extracted from the tension ratio data using the wavelet packet decomposition algorithm;

[0095] S23. Construct a three-dimensional tensor data structure to store the registered multi-source data:

[0096]

[0097] Thickness ij is the mesh thickness at grid point (i,j), temperature ij is the surface temperature at grid point (i,j), tension k is the tension value of the k-th sensor channel, i,j are the two-dimensional spatial coordinate indices, m is the grid row resolution, and n is the total number of sensor channels.

[0098] S3 includes the following steps:

[0099] S31. Load the pre-trained ResNet-34 network model, and set the input layer to a 128×128×3 convolutional kernel;

[0100] S32. Embed an attention mechanism module in the fully connected layer of the network to calculate the weight allocation for different material features:

[0101]

[0102] α i W is the attention weight for the i-th feature. i T is a trainable weight vector. i Let be the input feature vector, and exp(·) be the natural exponential function.

[0103] S33. The output layer connects to the decision tree classifier to generate a parameter set containing the tool tip angle, feed rate, and pressure value.

[0104] S4 includes the following steps:

[0105] S41. Generate basic slicing paths based on B-spline curve fitting algorithm;

[0106] S42. Dynamically adjust the Z-axis compensation of the tool head based on the material thickness gradient data:

[0107]

[0108] Where k is the material thickness gradient compensation coefficient. Let b be the gradient of the thickness in the X direction, and b be the basic compensation constant.

[0109] S43. Insert a circular arc transition segment at the turning point of the path, with the radius of curvature R satisfying:

[0110]

[0111] Where v is the linear velocity of the cutting head, μ is the friction coefficient between the mesh and the cutting tool, g is the acceleration due to gravity, and δ 安全 For safety margin.

[0112] S5 includes the following steps:

[0113] S51. Deploy two sets of laser Doppler velocimeters upstream and downstream of the slitting line;

[0114] S52. Calculate the tension fluctuation coefficient:

[0115]

[0116] v 上游 v 下游 To determine the speed of mesh deployment along the online and offline lines;

[0117] S53. When η is greater than 0.15, the dynamic adjustment mechanism is triggered.

[0118] S6 includes the following steps:

[0119] S61. Establish a fuzzy PID controller, with the input variable being the value of η and the output variable being the servo motor speed correction amount;

[0120] S62. Set the membership function:

[0121] Low fluctuation domain: η∈[0,0.05) → trigonometric function;

[0122] In the mid-wave domain: η∈[0.05,0.15) → Gaussian function;

[0123] High fluctuation domain: η∈[0.15,0.3] → S-shaped function.

[0124] include:

[0125] Multimodal data acquisition module: includes millimeter-wave radar array, infrared thermal imager and distributed stress sensor;

[0126] Intelligent decision-making module: Equipped with a GPU-accelerated deep learning inference engine;

[0127] Dynamic execution module: integrates a six-axis robotic arm and an adaptive servo tool holder;

[0128] Closed-loop feedback module: includes machine vision quality inspection unit and industrial IoT gateway.

[0129] The intelligent decision-making module includes a dual-redundant processing architecture:

[0130] The main processing channel runs the TensorRT-optimized ResNet model;

[0131] The backup channel stores an empirical formula library based on the physical properties of materials:

[0132]

[0133] Where K is the material property correction coefficient, σ b R represents the tensile strength limit of the mesh fabric, t represents the actual thickness of the mesh fabric, and R represents the tensile strength limit of the mesh fabric. 刀头 The radius of the arc at the working end of the cutter head;

[0134] When the confidence level of the primary channel is less than 85%, it will automatically switch to the backup channel.

[0135] S7 includes the following steps:

[0136] S71. A linear CCD camera is used to scan the edge of the mesh fabric at a resolution of 5μm.

[0137] S72. Extract burr feature parameters using a morphological edge detection algorithm;

[0138] S73. A quality alarm signal is generated when the length of the burr exceeds 3 pixels continuously.

[0139] S83 includes:

[0140] S831. Establish a tool tip life prediction model:

[0141]

[0142] Where β is the material wear coefficient and N is the cumulative number of cuts;

[0143] S832. When an early warning signal is triggered, the current working parameters are automatically recorded and a maintenance log is generated.

[0144] S833. Train a random forest regression model based on historical maintenance data to predict the optimal tool change cycle:

[0145]

[0146] Where α is the weighting factor, γ is the safety factor, σ is the periodic standard deviation, and t i For the historical replacement cycle, σ t The periodic standard deviation is given.

[0147] Example 1

[0148] Step 1: Multi-source data acquisition and fusion

[0149] At the beginning of the slitting production line, a millimeter-wave radar array continuously scans the surface of the moving mesh fabric to generate a thickness distribution map. Simultaneously activated, an infrared thermal imager captures the temperature field distribution of the mesh fabric. When a local temperature anomaly is detected, the system automatically activates the stress sensor array embedded in the slitting roller to acquire the warp and weft tension data of that area in real time. The three data streams are input into the spatiotemporal alignment model, and a three-dimensional state tensor matrix is ​​generated through pixel-level registration to fully characterize the physical state changes of the mesh fabric.

[0150] Step 2: Intelligent Parameter Decision

[0151] The fused state matrix is ​​input into the deep learning network model, and spatial features are extracted through convolutional kernels. When a region with abrupt changes in mesh thickness is identified, the attention mechanism automatically increases the decision weight of the thickness feature, and the network output module generates a combination of target segmentation parameters. Under special conditions, the system automatically evaluates the model confidence and, if necessary, switches to a backup channel to call empirical formulas in mechanics of materials to calculate the basic parameter values.

[0152] Step 3: Dynamic trajectory compensation

[0153] After generating the basic cutting path based on the slitting parameters, the path is smoothed and optimized using a curve fitting algorithm. In sections where the mesh thickness gradient changes significantly, the system dynamically calculates the vertical compensation of the cutter head and inserts a circular arc transition segment that meets safety constraints at the path turning points. At the same time, based on the real-time tension fluctuation coefficient, the system dynamically adjusts the cutter head feed speed using a fuzzy control strategy to avoid material overload deformation.

[0154] Step 4: Quality Closed-Loop Optimization

[0155] After slitting, the high-resolution vision inspection system scans the edge contour of the mesh and identifies microscopic burr defects through morphological algorithms. When continuous excessive burr features are found, the optimization algorithm reverses the calculation of the correction direction of the cutter head tilt angle and updates the process parameters. The synchronously running tool life prediction model generates preventive replacement suggestions based on the cumulative workload and historical maintenance data.

[0156] Step 5: System Collaboration and Linkage

[0157] The above process forms a closed-loop control circuit:

[0158] Quality inspection results are fed back to the parameter decision module in real time;

[0159] Input tool wear condition data into the life prediction model;

[0160] Abnormal operating conditions automatically trigger a dual-channel switching mechanism;

[0161] Achieve autonomous operation across the entire process of perception, decision-making, execution, and optimization.

[0162] Example 2

[0163] Step 1: Multi-source sensing data fusion

[0164] A millimeter-wave radar array is deployed behind the mesh unwinding device to scan the thickness profile of the high-speed moving mesh surface. Simultaneously activated, an infrared thermal imager detects abnormal temperature rises in localized areas due to friction. The system immediately activates the corresponding piezoelectric strain gauge array in that area to acquire real-time warp and weft tension distribution data. These three heterogeneous data sets are then fused into a matrix using a spatiotemporal registration engine.

[0165] 1. Thickness contour maps are used to locate material accumulation areas;

[0166] 2. Temperature field reveals friction hotspots in the equipment;

[0167] 3. Tension data reflects stress concentration points;

[0168] Step 2: Dual-channel collaborative decision-making

[0169] The fusion matrix is ​​input into the deep learning model of the main processing channel to identify abrupt changes in thickness gradient in region B1. When the model lacks confidence in the cutting parameters of the new stainless steel alloy mesh, the system automatically switches to the backup channel.

[0170] 1. Use the material physical property library to match the strength coefficient of stainless steel;

[0171] 2. Calculate the basic cutting pressure parameters using mechanical formulas;

[0172] 3. Dynamically adjust the pressure value based on real-time tension data;

[0173] Step 3: Dynamic Trajectory Optimization

[0174] After generating the initial slicing path based on the decision parameters, the system targets the detected stress concentration areas:

[0175] 1. A smooth cutting trajectory is reconstructed using the B-spline algorithm;

[0176] 2. Insert curvature-constrained transition arcs at the thickness abrupt change boundary;

[0177] 3. Automatically reduce the cutter head travel speed based on temperature rise data;

[0178] Step 4: Closed-loop quality optimization

[0179] After slitting, the linear scan camera scans the edge of the mesh fabric:

[0180] 1. Morphological processing identified microscopic burrs in the E1 region;

[0181] 2. Optimize the tool tip tilt angle parameters using the gradient backpropagation algorithm;

[0182] 3. Synchronously update the tool life decay model parameters;

[0183] Example of abnormal operating condition response

[0184] When the vision system detects a continuous burr defect:

[0185] 1. The quality assessment module triggered a level three alarm;

[0186] 2. Review historical optimization records in the process parameter library;

[0187] 3. The dynamic compensation system corrects the vibration frequency of the cutter head;

[0188] 4. Maintain the predictive model to shorten the detection cycle;

[0189] Technological advantages are reflected

[0190] 1. Multi-source sensing systems accurately capture material-device interactions;

[0191] 2. A dual-channel decision-making mechanism ensures stable output under special operating conditions;

[0192] 3. The closed-loop optimization process enables autonomous source tracing and correction of defects;

[0193] 4. The anomaly response system significantly reduces unplanned downtime.

[0194] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0195] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for automatically making a program of slitting a steel wire cloth, characterized by: The method comprises the following steps: S1, collecting material characteristic data and dynamic stress state data of steel wire mesh cloth based on a multi-source sensing device; S2, performing real-time fusion processing on the material characteristic data and dynamic stress state data to generate a mesh cloth comprehensive state matrix; S3, performing slitting parameter matching processing on the mesh cloth comprehensive state matrix based on a deep convolutional network to generate a target slitting process parameter set; S4, performing dynamic compensation calculation on the cutter head trajectory parameters of a slitting device according to the target slitting process parameter set to generate three-dimensional slitting path planning data; S5, monitoring mesh cloth tension fluctuation data in real time based on a laser displacement sensing device, and triggering step S6 when the fluctuation threshold exceeds a preset range; S6, performing slitting speed adaptive adjustment processing according to the tension fluctuation data to generate a dynamic slitting speed correction instruction; S7, performing quality defect detection on the edges of the mesh cloth after slitting based on a machine vision system to generate a slitting quality evaluation report and feed back to step S3.

2. A method for automatically cutting and making a program of wire mesh cloth according to claim 1, characterized in that: The S1 comprises the following steps: S11, scanning the surface of the mesh cloth at a frequency of 800 MHz by a millimeter wave radar sensor to generate thickness distribution cloud map data; S12, capturing temperature gradient data of the mesh cloth in motion at a sampling frequency of 10 Hz by an infrared thermal imager; S13, embedding a piezoelectric stress sensor array in the slitting roller shaft to collect real-time mesh cloth warp and weft tension ratio data.

3. A method for automatically cutting and making a program of steel wire mesh cloth according to claim 1, characterized in that: The S2 comprises the following steps: S21, establishing a space-time alignment model to perform pixel-level registration on the thickness distribution cloud map data and the temperature gradient data; S22, performing frequency domain feature extraction on the tension ratio data by a wavelet packet decomposition algorithm; S23, constructing a three-dimensional tensor data structure to store the registered multi-source data: The thickness ij is the mesh cloth thickness at the grid point (i, j), the temperature ij is the surface temperature at the grid point (i, j), the tension k is the tension value of the kth sensor channel, i and j are two-dimensional space coordinate indexes, m is the grid row resolution, and n is the total number of sensor channels.

4. The method of claim 1, wherein the method is characterized by: The S3 comprises the following steps: S31, loading a pre-trained ResNet-34 network model, and setting the input layer to a convolution kernel of 128×128×3; S32, embedding an attention mechanism module in the network fully connected layer to calculate the weight distribution of different material characteristics: alpha i is the attention weight for the i-th feature, W i is a trainable weight vector, T i is the input feature vector, and exp(·) is the natural exponential function. S33, connecting a decision tree classifier to the output layer to generate a parameter set containing the cutter head inclination angle, feed speed, and pressure value.

5. The method of claim 1, wherein the method is characterized by: The S4 comprises the following steps: S41, generating a basic slitting path based on a B-spline curve fitting algorithm; S42, dynamically adjusting the cutter Z-axis compensation amount according to the material thickness gradient data: wherein k is a material thickness gradient compensation factor, is the gradient of the thickness in the X direction, and b is a base compensation constant; S43, inserting a circular arc transition section at the path turning point, and the curvature radius R satisfies: where v is the linear speed of the tool tip, μ is the friction coefficient between the cloth and the tool, g is the acceleration of gravity, δ 安全 is the safety margin.

6. The method of claim 1, wherein the method is characterized by: The S5 comprises the following steps: S51, deploying two groups of laser Doppler velocimeters upstream and downstream of the slitting line; S52, calculating the tension fluctuation coefficient: v 上游 、v 下游 Vd is the speed of the downstream web on the slitting line; S53, triggering the dynamic adjustment mechanism when η is greater than 0.

15.

7. A method for automatically cutting and making a program of wire cloth according to claim 1, characterized in that: The S6 comprises the following steps: S61, establishing a fuzzy PID controller, with η value as the input variable and servo motor speed correction amount as the output variable; S62, setting the membership function: Low fluctuation domain: η ∈ [0, 0.05) → triangular function; Medium volatility domain: η ∈ [0.05, 0.15) -> Gaussian function; High volatility domain: η ∈ [0.15, 0.3] -> S-shaped function.

8. The method of claim 1, wherein the method is characterized by: The S7 comprises the following steps: S71, using a linear array CCD camera to scan the edge of the mesh at a resolution of 5μm; S72, extracting burr feature parameters through a morphological edge detection algorithm; S73, generating a quality alarm signal when the burr length continuously exceeds 3 pixels.

9. A method for automatically cutting and making a program of a steel wire mesh cloth according to claim 1, characterized in that: After the S7, the following steps are further included: S8, performing process parameter self-optimization processing based on the slitting quality evaluation report: S81, extracting the corresponding material property data of the burr length exceeding section; S82, optimizing the tool head inclination angle parameter through a gradient descent algorithm: where η is a learning rate coefficient, The length of burr is the partial derivative of the length of the spur with respect to the angle of inclination. S83, when the burr length still exceeds the standard after continuous optimization for 3 times, triggering a tool head wear early warning signal.

10. A method of programming automatic slitting of a wire cloth according to claim 9, characterized in that: The S83 comprises: S831, establishing a tool head life prediction model: Where β is the material wear coefficient, and N is the cumulative cutting frequency; S832, when the early warning signal is triggered, automatically recording the current working parameters and generating a maintenance log; S833, training a random forest regression model based on historical maintenance data to predict the optimal tool replacement cycle: where a is a weighting factor, g is a safety factor, s is the period standard deviation, t i is the historical replacement period, s t is the period standard deviation.

Images