Metal plate cutting control method and system

Abstract

The invention discloses a metal plate cutting control method and system, and relates to the technical field of cutting control, and the method comprises the steps: collecting historical cutting data, employing an improved K-means clustering algorithm to divide working conditions into similar working condition clusters, constructing a parameter optimization model based on an improved particle swarm algorithm, and building a dynamic parameter library; core cutting state indexes are calculated, and a three-level anomaly recognition mechanism is established based on the indexes; current working condition parameters are input into a dynamic parameter library for feature matching, an optimization model of a matching cluster is called to output an initial parameter combination, deviation indexes are calculated based on real-time monitoring data, an improved ant colony algorithm is adopted to construct a self-adaptive correction model, and corrected optimal cutting control parameters are generated; the corrected parameters are converted into control instructions and issued to the cutting equipment, the parameters are monitored and finely adjusted in real time, the cutting path is dynamically corrected through a partition compensation method, the effect is evaluated after cutting is completed, a dynamic parameter library is updated, and the cutting precision, efficiency and system adaptability are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of cutting control technology, specifically to a method and system for cutting metal sheets. Background Technology

[0002] Metal sheet cutting technology is widely used in machinery manufacturing, automobiles, aerospace and other fields. Its core goal is to meet the processing requirements of complex parts through high-precision and high-efficiency cutting processes. Existing cutting control technology mainly relies on preset parameters and human experience. PLC control systems are used to adjust basic parameters such as cutting power and speed. Some advanced systems have introduced sensor monitoring technology to collect data such as temperature and displacement in real time during the cutting process, and use simple feedback mechanisms to fine-tune the parameters to improve cutting quality.

[0003] However, existing technologies still have significant shortcomings in practical applications. Traditional methods classify working conditions in a rather coarse manner, failing to fully consider the comprehensive influence of multiple dimensions such as sheet material properties, equipment status, and environmental factors, resulting in insufficient parameter matching accuracy and difficulty in adapting to complex and ever-changing cutting needs. Secondly, existing anomaly identification mechanisms are mostly triggered by a single threshold, lacking hierarchical early warning and dynamic adjustment capabilities, and cannot effectively cope with sudden anomalies during the cutting process, which can easily cause processing defects or equipment damage.

[0004] Furthermore, the parameter optimization and correction process of the existing system lacks adaptability, making it difficult to achieve closed-loop optimization. The insufficient utilization of historical data makes it difficult for the optimization model to be continuously updated and iterated, and it is impossible to dynamically adjust the optimal parameters with changes in working conditions. These shortcomings limit the further improvement of cutting accuracy and efficiency. Summary of the Invention

[0005] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a metal sheet cutting control method and system. By dividing the working conditions into multiple dimensions, matching the dynamic parameter library, and real-time monitoring and adaptive correction, it solves the problems of insufficient parameter matching accuracy, delayed abnormal response, and inadequate utilization of historical data in traditional cutting technologies, significantly improving cutting accuracy, efficiency, and system adaptability.

[0006] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a method for controlling the cutting of metal sheets, comprising: Metal sheet cutting conditions are divided into multiple dimensions. Historical cutting data are collected and features are extracted. An improved K-means clustering algorithm is used to divide the conditions into clusters of similar conditions. A parameter optimization model based on an improved particle swarm optimization algorithm is constructed, and a dynamic parameter library is established. Deploy multi-source sensors on the cutting equipment to collect real-time monitoring data, calculate core cutting status indicators, and establish a three-level anomaly identification mechanism based on the indicators; Input the current working condition parameters into the dynamic parameter library for feature matching, call the optimization model of the matching cluster to output the initial parameter combination, calculate the deviation index based on real-time monitoring data, construct an adaptive correction model using an improved ant colony algorithm, and generate the corrected optimal cutting control parameters. The correction parameters are converted into control commands and sent to the cutting equipment. The parameters are monitored and fine-tuned in real time. The cutting path is dynamically corrected using a partition compensation method. After the cutting is completed, the effect is evaluated and the dynamic parameter library is updated.

[0007] Furthermore, the multiple dimensions for classifying metal sheet cutting conditions include sheet material properties, cutting target, equipment status, environment, cutting path, and cooling. Among these, sheet material properties include material, thickness, initial stress value, and surface flatness; cutting target includes part shape complexity level, dimensional accuracy requirements, and cut surface roughness requirements; equipment status includes cutting head wear level, laser generator power stability, and auxiliary gas purity; environment includes room temperature, humidity, and air dust concentration; cutting path includes path length, number of inflection points, and closed or non-closed path type; and cooling includes cooling method, cooling air velocity, and cooling area coverage.

[0008] Furthermore, an improved K-means clustering algorithm is adopted, introducing cutting effect weights to divide all working conditions into K similar working condition clusters. For each similar working condition cluster, a parameter optimization model based on the improved particle swarm optimization algorithm is constructed with the following optimization objectives: the proportion threshold of dimensional error ≤ design tolerance, the proportion threshold of cutting surface roughness ≤ required Ra value, and the proportion threshold of material consumption ≤ the historical average value within the cluster. A dynamic parameter library is established. Whenever a cutting operation is completed, if the cutting effect meets the optimization objective, the working condition parameters, control parameters, and cutting effect data of that operation are added as new samples to the corresponding similar working condition cluster.

[0009] Furthermore, based on the monitoring data, core cutting status indicators are calculated, including the thermal deformation of the sheet metal, the load status of the cutting head, the thermal status of the cutting area, and the auxiliary gas supply status. Among them, the thermal deformation of the sheet metal is calculated by comparing the real-time position with the initial position before cutting to determine the deformation at each monitoring point; the load status of the cutting head is calculated by comparing the motor current with the rated current to determine the load rate; the thermal status of the cutting area is calculated by using temperature data to determine the temperature gradient of the cutting area; and the auxiliary gas supply status is calculated by comparing the flow rate data with the set flow rate to determine the flow rate deviation rate.

[0010] Furthermore, the three-level anomaly identification mechanism includes: Level 1 anomalies are load rate of 90%~100%, flow rate deviation rate of -5%~5%, local temperature gradient ≤5℃ / mm, and local deformation of the sheet material of 0.02mm~0.05mm, which are only logged; Level 2 anomalies are load rate of 100%~120%, flow rate deviation rate of -10%~-5% or 5%~10%, local temperature gradient of 5℃ / mm~10℃ / mm, and local deformation of the sheet material of 0.05mm~0.1mm, which trigger an early warning and initiate parameter optimization and correction; Level 3 anomalies are load rate >120%, flow rate deviation rate <-10% or >10%, local temperature gradient >10℃ / mm, and local deformation of the sheet material of >0.1mm, which immediately suspend cutting and trigger an early warning.

[0011] Furthermore, the working conditions parameters of the current cutting operation are input into the dynamic parameter library. Through the working condition feature matching algorithm, the feature similarity between the current working condition and each similar working condition cluster is calculated. The cluster with the highest similarity is the matching cluster. The parameter optimization model of the matching cluster is called to output the initial optimized parameter combination.

[0012] Furthermore, by comparing real-time monitoring data with standard operating condition data of matching clusters in the dynamic parameter library, deviation indices, including deformation deviation, temperature deviation, and load deviation, are calculated. With the goals of minimizing dimensional error, minimizing material consumption, and maximizing cutting efficiency, an adaptive correction model is constructed. The initial optimization parameters are used as model inputs, and deformation deviation, temperature deviation, and load deviation are used as model constraints. An improved ant colony algorithm is used to solve the model, introducing deviation weights, where deviation weight = absolute deviation value / maximum allowable deviation value. After solving, the corrected optimal combination of cutting control parameters is obtained.

[0013] Furthermore, after converting the correction parameters into control commands, they are sent to the PLC control system of the cutting equipment. During the cutting process, the deviation index is periodically monitored, and if it exceeds the allowable range, parameter fine-tuning is initiated. When using the partition compensation method to dynamically correct the cutting path, the plate is divided into M equidistant regions along the X-axis. The average deformation of N monitoring points in each region is taken as the compensation benchmark. For the movement path of the cutting head, the path coordinates are dynamically corrected according to the compensation benchmark of the current region.

[0014] Furthermore, after the cutting operation is completed, the actual measured values ​​are compared with the design requirements to calculate the actual cutting effect index. If the actual effect meets the optimization target, the working condition parameters, optimization parameters, and cutting effect data of this operation are used as high-quality cases and added to the corresponding similar working condition cluster in the dynamic parameter library.

[0015] A metal sheet cutting control system, comprising: The working condition analysis module divides the metal sheet cutting working conditions into multiple dimensions, collects historical cutting data and extracts features, uses an improved K-means clustering algorithm to divide the working conditions into similar working condition clusters, constructs a parameter optimization model based on an improved particle swarm algorithm, and establishes a dynamic parameter library. The real-time monitoring module deploys multi-source sensors on the cutting equipment to collect real-time monitoring data, calculate core cutting status indicators, and establish a three-level anomaly identification mechanism based on the indicators. The parameter correction module inputs the current working condition parameters into the dynamic parameter library for feature matching, calls the optimization model of the matching cluster to output the initial parameter combination, calculates the deviation index based on real-time monitoring data, and uses an improved ant colony algorithm to build an adaptive correction model to generate the corrected optimal cutting control parameters. The control and feedback module converts the correction parameters into control commands and sends them to the cutting equipment. It monitors and fine-tunes the parameters in real time, uses a partition compensation method to dynamically correct the cutting path, evaluates the effect after cutting, and updates the dynamic parameter library.

[0016] (III) Beneficial Effects This invention provides a method and system for controlling the cutting of metal sheets, which has the following advantages: (1) By dividing the metal sheet cutting conditions into multiple dimensions and using the improved K-means clustering algorithm and particle swarm optimization model, intelligent classification and parameter optimization of the working conditions are realized, improving cutting accuracy and efficiency. The dynamic parameter library enables the effective use of historical data, optimizes resource allocation, reduces material consumption, enhances system adaptability, and can quickly match similar working conditions and output optimal parameters, providing a reliable basis for subsequent real-time correction and significantly improving cutting quality and stability.

[0017] (2) By deploying multi-source sensors to monitor the cutting process in real time and establishing a three-level anomaly identification mechanism, the core indicators such as thermal deformation of the plate, cutting head load, temperature gradient and gas flow are accurately captured, and dynamic hierarchical early warning of abnormal conditions is achieved. The reliability of monitoring is improved by data synchronization and preprocessing, and the cutting parameters are optimized by real-time feedback. This effectively reduces cutting defects caused by equipment overload or environmental fluctuations, ensures processing quality and safety, and reduces downtime and maintenance costs.

[0018] (3) Through feature matching and adaptive correction model of dynamic parameter library, the cutting parameters are intelligently optimized and adjusted in real time, accurately matched similar working condition clusters, quickly output the initial optimized parameter combination, calculate the deviation index based on real-time monitoring data, and dynamically correct parameters by improving ant colony algorithm, effectively suppress abnormalities such as plate deformation and temperature fluctuation, improve cutting accuracy and efficiency, reduce material waste, enhance system adaptability, ensure that the optimal control parameters can be generated under different working conditions, and significantly improve processing quality and stability.

[0019] (4) Through dynamic parameter conversion and real-time control optimization, the precise execution and closed-loop management of the cutting process are realized. The correction parameters are quickly converted into equipment instructions. PLC control ensures accurate response of the actuator. Deviation is monitored in real time and parameters are fine-tuned to avoid cutting interruption. The path is dynamically corrected by the partition compensation method, especially for the pre-compensation of the inflection point of irregular parts, which significantly improves the dimensional accuracy. The effect is evaluated after cutting and the dynamic parameter library is updated to form a continuous optimization closed loop. At the same time, the robustness of the system is improved by the error tracing mechanism, and finally the cutting quality, efficiency and resource utilization are improved simultaneously. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the steps of the metal sheet cutting control method of the present invention; Figure 2 This is a schematic diagram of the metal sheet cutting control method of the present invention; Figure 3 This is a schematic diagram of the metal sheet cutting control system of the present invention. Detailed Implementation

[0021] 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.

[0022] Please see Figures 1-2 This invention provides a method for controlling the cutting of metal sheets, comprising the following steps: Step 1: Divide the metal sheet cutting conditions into multiple dimensions, collect historical cutting data and extract features, use the improved K-means clustering algorithm to divide the conditions into similar condition clusters, construct a parameter optimization model based on the improved particle swarm algorithm, and establish a dynamic parameter library; Step one includes the following: Step 101: Divide the metal sheet cutting process into multiple dimensions, including sheet properties, cutting target, equipment status, environment, cutting path, and cooling. The sheet properties dimension includes material, thickness, initial stress value, and surface flatness. The cutting target dimension includes part shape complexity level, dimensional accuracy requirements, and cut surface roughness requirements (Ra value). The equipment status dimension includes cutting head wear level, laser generator power stability, and auxiliary gas purity. The environment dimension includes room temperature, humidity, and air dust concentration. The cutting path dimension includes path length, number of inflection points, and closed / non-closed path type. The cooling dimension includes cooling method, cooling air velocity, and cooling area coverage. Surface flatness is the maximum difference in convexity or concaveness measured using a laser flatness measuring instrument; the complexity level of part shape is a custom quantitative classification: Level 1: regular polygon, number of inflection points ≤ 4 and edges are straight lines; Level 2: complex polygon, number of inflection points 5-10 or containing 1-2 rounded edges; Level 3: irregular part, number of inflection points > 10 or containing 3 or more free curve edges; the wear level of the cutting head is classified according to the laser energy attenuation rate: Level A: attenuation rate ≤ 5%, Level B: attenuation rate 6%-15%, Level C: attenuation rate 16%-25%, the attenuation rate is measured by a laser power meter; the power stability of the laser generator is the power fluctuation range within 1 hour of continuous operation, in percentage, range 0-3%; Step 102: For each dimension, collect at least 500 sets of historical cutting data. Each set of data includes the specific parameter values, corresponding cutting control parameters, and cutting effect data for that dimension. The cutting control parameters include cutting power, cutting speed, gas pressure, and cutting head height. The cutting effect includes dimensional error, surface roughness, and material consumption. Step 103: Extract features from the collected historical data. Specifically, extract the combination of material, thickness, and initial stress value from the plate properties dimension; extract the combination of dimensional accuracy requirements and part shape complexity level from the cutting target dimension; and extract the combination of cutting head wear level and power stability from the equipment status dimension. An improved K-means clustering algorithm is used, introducing cutting effect weights. Samples with better cutting effects have higher weights. The weight calculation method is: Weight = (1 - Dimensional Error / Design Tolerance) × 0.4 + (1 - Cutting Surface Roughness / Required Ra Value) × 0.3 + Consumable Loss Reduction Rate × 0.3). All working conditions are divided into K similar working condition clusters. For example, "Stainless Steel 304 - 8mm - 100MPa - ±0.05mm - Grade 2 - Grade A" is one cluster, and "Carbon Steel Q235 - 15mm - 300MPa - ±0.1mm - Grade 3 - Grade B" is another cluster. Step 104: For each cluster of similar working conditions, construct a parameter optimization model based on an improved particle swarm optimization algorithm, with optimization targets including the percentage thresholds of dimensional error ≤ design tolerance (e.g., 80%), surface roughness ≤ required Ra value, and material consumption ≤ historical average within the cluster. Use historical cutting data within the cluster as training samples to train the model and output the optimal combination of cutting control parameters. For example, the optimal parameters for a certain cluster are: power 1800W, speed 600mm / min, gas pressure 0.8MPa, and cutting head height 0.3mm. Introduce 5-fold cross-validation during training to ensure that the optimization accuracy of the model meets the target under different sub-working conditions within the cluster, and that the probability of the cutting effect corresponding to the optimized parameters meeting the target requirements is ≥95%. Step 105: Establish a dynamic parameter library, including a mechanism for real-time case entry and periodic model iteration. Whenever a cutting operation is completed, if the cutting effect meets the optimization objective, namely "the proportion threshold of dimensional error ≤ design tolerance, the proportion threshold of cutting surface roughness ≤ the required Ra value, and the proportion threshold of material consumption ≤ the historical average within the cluster", then the working condition parameters, control parameters, and effect data of this operation are added as new samples to the corresponding similar working condition cluster. Every 200 high-quality cases are accumulated, the parameter optimization model of the cluster is retrained and the model parameters are updated. At the same time, if a new working condition combination appears, a similar working condition cluster is automatically added, and at least 50 sets of experimental data are collected to train the optimization model, expanding the working condition coverage of the dynamic library.

[0023] When using this method, refer to steps 101 to 105: By dividing the metal sheet cutting conditions into multiple dimensions and employing an improved K-means clustering algorithm and particle swarm optimization model, intelligent classification and parameter optimization of the conditions are achieved, improving cutting accuracy and efficiency. The dynamic parameter library enables the effective use of historical data, optimizes resource allocation, reduces material consumption, enhances system adaptability, and can quickly match similar conditions and output optimal parameters, providing a reliable foundation for subsequent real-time correction and significantly improving cutting quality and stability.

[0024] Step 2: Deploy multi-source sensors on the cutting equipment to collect real-time monitoring data, calculate core cutting status indicators, and establish a three-level anomaly identification mechanism based on the indicators; Step two includes the following: Step 201: Deploy sensors on the cutting equipment. Specifically, laser displacement sensors are installed on both sides of the cutting head to collect real-time positions of various monitoring points on the surface of the sheet material; infrared temperature sensors are installed 50mm in front of the cutting head to collect real-time temperature of the cutting area; current sensors are connected in series with the cutting head drive motor circuit to collect real-time motor current; vibration sensors are attached to the cutting head bracket to collect the vibration acceleration of the cutting head; gas flow sensors are installed in the auxiliary gas pipeline to collect real-time gas flow; and environmental sensors are installed inside the cutting workshop to collect room temperature and humidity. Step 202: Using timestamp synchronization technology, the data collected by different sensors are unified to the same time axis. Outliers are removed from the preprocessed data. The 3σ criterion is used to remove abnormal data caused by sensor failure or electromagnetic interference. That is, if the deviation of a data point from the mean of the sensor data exceeds 3 times the standard deviation, it is judged as an outlier. The moving average method is used to smooth the surface position data of the plate collected by the laser displacement sensor. Step 203: Based on the preprocessed monitoring data, calculate the core cutting status indicators, including the thermal deformation of the sheet metal, the load status of the cutting head, the thermal status of the cutting area, and the auxiliary gas supply status. The thermal deformation of the sheet metal is calculated by comparing the real-time position collected by the laser displacement sensor with the initial position before cutting, with expansion being positive and contraction negative. A cubic spline interpolation curve is used to fit the deformation gradient, with the x-axis representing the X-coordinate of the monitoring point and the y-axis representing the deformation, reflecting the deformation differences at different X-positions. The load status of the cutting head is calculated by comparing the motor current collected by the current sensor with the rated current, using the load rate = real-time current / rated current × 100%, to determine if the cutting head is jammed. The thermal status of the cutting area is calculated by using temperature data collected by the infrared temperature sensor, using the temperature gradient = temperature difference between two adjacent points / distance between the two points, to determine if there is localized overheating. The auxiliary gas supply status is calculated by comparing the flow rate data collected by the gas flow sensor with the set flow rate, using the deviation rate = (real-time flow rate - set flow rate) / set flow rate × 100%, to determine if the gas supply is stable. Step 204: Based on the above core status indicators, establish a three-level anomaly identification mechanism: Level 1 anomaly: Load rate 90%~100%, flow deviation rate -5%~5%, local temperature gradient ≤5℃ / mm, local deformation of the plate 0.02mm~0.05mm. At this time, only internal system log recording is triggered, recording the time and abnormal indicator values, without affecting the cutting execution; Level 2 anomaly: Load rate 100%~120%, flow deviation rate -10%~-5% or 5%~10%, local temperature gradient 5℃ / mm~10℃ / mm, local deformation of the plate 0.05mm~0.1mm. At this time, an early warning is triggered, and subsequent parameter optimization and correction are initiated; Level 3 anomaly: Load rate >120%, flow deviation rate <-10% or >10%, local temperature gradient >10℃ / mm, local deformation of the plate >0.1mm. At this time, cutting is immediately suspended, the cutting head is raised to a safe height ≥50mm, confirmed by the Z-axis limit sensor of the cutting head, and an early warning is triggered, awaiting manual confirmation and processing.

[0025] When using this method, refer to steps 201 to 204: By deploying multi-source sensors to monitor the cutting process in real time and establishing a three-level anomaly identification mechanism, the system accurately captures core indicators such as thermal deformation of the sheet material, cutting head load, temperature gradient, and gas flow rate, enabling dynamic graded early warning of abnormal conditions. Through data synchronization and preprocessing, the system improves monitoring reliability and optimizes cutting parameters in conjunction with real-time feedback. This effectively reduces cutting defects caused by equipment overload or environmental fluctuations, ensuring processing quality and safety while reducing downtime and maintenance costs.

[0026] Step 3: Input the current operating condition parameters into the dynamic parameter library for feature matching, call the optimization model of the matching cluster to output the initial parameter combination, calculate the deviation index based on real-time monitoring data, construct an adaptive correction model using the improved ant colony algorithm, and generate the corrected optimal cutting control parameters. Step three includes the following: Step 301: Input the working condition parameters of the current cutting operation into the dynamic parameter library. Calculate the feature similarity between the current working condition and each similar working condition cluster using the working condition feature matching algorithm. The cluster with the highest similarity is the matching cluster. Call the parameter optimization model of the matching cluster to output the initial optimized parameter combination. Step 302: Compare the real-time monitoring data with the standard operating condition data of the matching cluster in the dynamic parameter library. The standard operating condition data is the average monitoring data of historical high-quality cases within the cluster, such as standard deformation of 0.03mm, standard temperature of 800℃, and standard load rate of 90%. Calculate the deviation indicators, including deformation deviation, temperature deviation, and load deviation. Among them, deformation deviation ΔD = current plate deformation - standard deformation, temperature deviation ΔT = current cutting area temperature - standard temperature, and load deviation ΔL = current cutting head load rate - standard load rate. Based on the direction and magnitude of the deviation, determine the type of difference between the current operating condition and the standard operating condition. For example, if ΔD is positive and the absolute value is >0.03mm, it means that the current plate expansion is greater than the standard operating condition, and the parameters need to be adjusted to suppress deformation. Step 303: Construct an adaptive correction model with the objectives of minimizing dimensional error, minimizing material consumption, and maximizing cutting efficiency; use the initial optimization parameters as model inputs and ΔD, ΔT, and ΔL as model constraints; solve the model using an improved ant colony algorithm, which introduces deviation weights. Deviation weight = absolute deviation / maximum allowable deviation. For example, if the maximum allowable deviation for ΔD is 0.1 mm and the current ΔD = 0.04 mm, then the deviation weight = 0.4. The larger the deviation weight, the higher the priority of the corresponding parameter correction; after solving, the optimal combination of cutting control parameters after correction is obtained.

[0027] When using this method, refer to the content of steps 301 to 303: Through feature matching and adaptive correction models based on a dynamic parameter library, intelligent optimization and real-time adjustment of cutting parameters are achieved. Similar working condition clusters are accurately matched, and initial optimized parameter combinations are quickly output. Deviation indicators are calculated based on real-time monitoring data. Parameters are dynamically corrected through an improved ant colony algorithm, effectively suppressing abnormalities such as sheet deformation and temperature fluctuations, improving cutting accuracy and efficiency, reducing material waste, enhancing system adaptability, ensuring the generation of optimal control parameters under different working conditions, and significantly improving processing quality and stability.

[0028] Step 4: Convert the correction parameters into control commands and send them to the cutting equipment. Monitor and fine-tune the parameters in real time. Use the partition compensation method to dynamically correct the cutting path. After the cutting is completed, evaluate the effect and update the dynamic parameter library.

[0029] Step four includes the following: Step 401: The modified optimal cutting control parameter combination is converted into control commands for the cutting equipment, such as power commands, motor speed commands, and gas valve opening commands, and sent to the PLC control system of the cutting equipment via industrial Ethernet; After receiving the commands, the PLC control system first presets the parameters of each actuator, such as the cutting head drive motor, gas valve, and laser generator, and after confirming that the status of each mechanism is normal, if the motor is fault-free and the valve opening meets the standard, the cutting operation is started; Step 402: During the cutting process, continuously collect monitoring data from multiple sensors and calculate the current deformation deviation every 100ms. Temperature deviation Load deviation If any deviation indicator exceeds the allowable range, such as If the design tolerance is greater than 60%, then parameter fine-tuning is initiated: Based on the deviation type, the optimal cutting control parameters are slightly modified, with the modification magnitude ≤ 10% of the initial parameters, to avoid sudden parameter changes causing cutting interruptions. The modification formula is: Fine-tuned parameter = Control parameter × (1 + Deviation weight × Correction coefficient), where the correction coefficient is determined based on the deviation direction. If positive, the power correction factor is -0.02; for example, =0.03mm, design tolerance ±0.05mm, allowable range 0.03mm, deviation weight = 0.03 / 0.1 = 0.3, power after fine adjustment = 1700W × (1 + 0.3 × (-0.02)) = 1689.8W; the fine-tuned parameters are updated in real time through the PLC control system to ensure that the parameters always match the current working conditions during the cutting process; Step 403: Based on the real-time deformation data of the sheet metal collected by the laser displacement sensor, the compensation amount of the cutting path is calculated using the partition compensation method. The sheet metal is divided into M (e.g., 10) equidistant regions along the X-axis. The width of each region is equal to the sheet metal length / M. The average deformation of N (e.g., 3) monitoring points in each region is taken as the compensation benchmark for that region. For the X / Y axis movement path of the cutting head, the path coordinates are dynamically corrected according to the compensation benchmark of the current region. For example, if the compensation benchmark of a certain region is +0.06mm, then the X-axis coordinate of the cutting path in that region is the design coordinate -0.06mm. After the compensation amount is calculated, it is converted into the displacement command of the cutting head drive motor to adjust the position of the cutting head in real time. For the inflection point position in the cutting of irregular parts, and the point in the path where the angle change is ≥30°, an early pre-compensation strategy is adopted. The compensation amount is gradually applied starting 5mm before reaching the inflection point to avoid dimensional deviations caused by deformation at the inflection point. Step 404: After the cutting operation is completed, a high-precision testing device is used to evaluate the effect. The actual measured values ​​are compared with the design requirements to calculate the actual cutting effect indicators. If the actual effect meets the thresholds for dimensional error ≤ design tolerance, surface roughness ≤ required Ra value, and material consumption ≤ historical average within the cluster, then the working parameters, optimization parameters, and effect data of this operation are added as high-quality cases to the corresponding similar working condition cluster in the dynamic parameter library. The case tags include operation time, equipment number, and operator. If the actual effect does not meet the standards, such as dimensional error > 80% of design tolerance, then the error tracing process is initiated: the deviation change curve of the monitoring data is compared, such as during the cutting process. For sudden increases in time points, combined with equipment logs such as motor current fluctuation records, the causes of errors are analyzed, such as insufficient parameter fine-tuning range or sensor calibration deviation. Improvement measures are formulated, such as adjusting the parameter fine-tuning range and recalibrating the laser displacement sensor. This case is stored in the dynamic library as a case to be optimized and set as a key sample in the next training, with the sample weight increased to 2 times. At the same time, the material consumption and cutting time of this cutting operation are statistically analyzed and compared with historical similar operations to calculate the optimization gain (optimization gain = (historical average - current value) / historical average × 100%), which provides data support for the weight adjustment of the parameter optimization model.

[0030] When using this method, refer to steps 401 to 404: Through dynamic parameter conversion and real-time control optimization, precise execution and closed-loop management of the cutting process are achieved. Correction parameters are quickly converted into equipment instructions, and PLC control ensures accurate response of the actuator. Deviations are monitored in real time and parameters are fine-tuned to avoid cutting interruptions. The path is dynamically corrected using a partition compensation method, especially for pre-compensation of inflection points of irregularly shaped parts, which significantly improves dimensional accuracy. The effect is evaluated after cutting and the dynamic parameter library is updated to form a continuous optimization closed loop. At the same time, the robustness of the system is improved through an error tracing mechanism, ultimately achieving simultaneous improvement in cutting quality, efficiency and resource utilization.

[0031] Please see Figure 3 The present invention also provides a metal sheet cutting control system, comprising: a working condition analysis module, a real-time monitoring module, a parameter correction module, and a control and feedback module, wherein: The working condition analysis module divides the metal sheet cutting working conditions into multiple dimensions, collects historical cutting data and extracts features, uses an improved K-means clustering algorithm to divide the working conditions into similar working condition clusters, constructs a parameter optimization model based on an improved particle swarm algorithm, and establishes a dynamic parameter library. The real-time monitoring module deploys multi-source sensors on the cutting equipment to collect real-time monitoring data, calculate core cutting status indicators, and establish a three-level anomaly identification mechanism based on the indicators. The parameter correction module inputs the current working condition parameters into the dynamic parameter library for feature matching, calls the optimization model of the matching cluster to output the initial parameter combination, calculates the deviation index based on real-time monitoring data, and uses an improved ant colony algorithm to build an adaptive correction model to generate the corrected optimal cutting control parameters. The control and feedback module converts the correction parameters into control commands and sends them to the cutting equipment. It monitors and fine-tunes the parameters in real time, uses a partition compensation method to dynamically correct the cutting path, evaluates the effect after cutting, and updates the dynamic parameter library.

[0032] In the application, the various formulas mentioned are all calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The coefficients in the formulas are set by those skilled in the art according to the actual situation.

[0033] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, and combinations thereof. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0034] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0035] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method of controlling cutting of a metal sheet material, characterized by: Comprise: Divide the metal plate cutting working condition into multiple dimensions, collect historical cutting data and extract features, use improved K-means clustering algorithm to divide the working condition into similar working condition clusters, construct parameter optimization model based on improved particle swarm algorithm, and establish dynamic parameter library; Deploy multiple source sensors on the cutting equipment, collect real-time monitoring data, calculate the core cutting state indicators, and establish a three-level abnormality identification mechanism based on the indicators; Input the current working condition parameters into the dynamic parameter library for feature matching, call the optimization model of the matching cluster to output the initial parameter combination, calculate the deviation index based on the real-time monitoring data, use the improved ant colony algorithm to construct the adaptive correction model, and generate the corrected optimal cutting control parameters; Convert the corrected parameters into control commands and send them to the cutting equipment, monitor and fine-tune the parameters in real time, dynamically correct the cutting path using the partition compensation method, and evaluate the effect after cutting and update the dynamic parameter library.

2. The metal plate cutting control method according to claim 1, characterized by: The multiple dimensions of the metal plate cutting working condition include plate attribute dimension, cutting target dimension, equipment state dimension, environment dimension, cutting path dimension and cooling dimension; among them, the plate attribute dimension includes material, thickness, initial stress value and surface flatness, the cutting target dimension includes part shape complexity level, size accuracy requirement and cutting surface roughness requirement, the equipment state dimension includes cutting head wear level, laser generator power stability and auxiliary gas purity, the environment dimension includes room temperature, humidity and air dust concentration, the cutting path dimension includes path length, number of inflection points and closed or non-closed path type, and the cooling dimension includes cooling method, cooling air speed and cooling area coverage.

3. The metal plate cutting control method according to claim 2, wherein: Using the improved K-means clustering algorithm, the cutting effect weight is introduced, all working conditions are divided into K similar working condition clusters, for each similar working condition cluster, the size error ≤ design tolerance proportion threshold, cutting surface roughness ≤ requirement Ra value proportion threshold, consumable consumption ≤ cluster historical average proportion threshold are used as optimization objectives, and the parameter optimization model based on improved particle swarm algorithm is constructed; a dynamic parameter library is established, and every time a cutting job is completed, if the cutting effect meets the optimization objective, the working condition parameters, control parameters and cutting effect data of this job are added as new samples to the corresponding similar working condition cluster.

4. The metal plate cutting control method of claim 1, wherein: Based on the monitoring data, the core cutting state indicators are calculated, including plate thermal deformation, cutting head load state, cutting area thermal state and auxiliary gas supply state, among them, the plate thermal deformation is calculated by comparing the real-time position with the initial position before cutting; the cutting head load state is calculated by comparing the motor current with the rated current; the cutting area thermal state is calculated by the temperature data; the auxiliary gas supply state is calculated by comparing the flow data with the set flow.

5. The metal plate cutting control method according to claim 4, wherein: The three-level abnormality identification mechanism includes: the first-level abnormality is a load rate of 90% to 100%, a flow deviation rate of -5% to 5%, a local temperature gradient of ≤5 ℃ / mm, and a local deformation amount of the plate of 0.02 mm to 0.05 mm, and only logs are recorded; the second-level abnormality is a load rate of 100% to 120%, a flow deviation rate of -10% to -5% or 5% to 10%, a local temperature gradient of 5 ℃ / mm to 10 ℃ / mm, and a local deformation amount of the plate of 0.05 mm to 0.1 mm, and a pre-warning is triggered and parameter optimization correction is started; the third-level abnormality is a load rate of >120%, a flow deviation rate of <-10% or >10%, a local temperature gradient of >10 ℃ / mm, and a local deformation amount of the plate of >0.1 mm, and cutting is immediately suspended and a pre-warning is triggered.

6. The metal plate cutting control method according to claim 1, wherein: The working condition parameters of the current cutting operation are input into the dynamic parameter library, the feature similarity of the current working condition and each similar working condition cluster is calculated through a working condition feature matching algorithm, the cluster with the highest similarity is the matching cluster, and an initial optimization parameter combination is output by calling the parameter optimization model of the matching cluster.

7. The metal plate cutting control method according to claim 6, wherein: The real-time monitoring data are compared with the standard working condition data of the matching cluster in the dynamic parameter library, a deviation index is calculated, including a deformation deviation, a temperature deviation and a load deviation, an adaptive correction model is constructed with the minimization of size error, the minimization of consumable loss and the maximization of cutting efficiency as the target, the initial optimization parameters are taken as the model input, and the deformation deviation, the temperature deviation and the load deviation are taken as the constraint conditions of the model; An improved ant colony algorithm is used to solve the model, a deviation weight is introduced, the deviation weight = deviation absolute value / maximum allowed deviation, and an optimal cutting control parameter combination after correction is obtained.

8. The metal plate cutting control method according to claim 1, wherein: After the corrected parameters are converted into control instructions, the control instructions are sent to the PLC control system of the cutting equipment, the deviation index is periodically monitored during the cutting process, and parameter fine tuning is started if the deviation index exceeds the allowed range; When the partition compensation method is used to dynamically correct the cutting path, the plate is divided into M equidistant regions along the X-axis direction, the deformation amount average of N monitoring points in each region is taken as the compensation reference, and for the movement path of the cutting head, the path coordinates are dynamically corrected according to the compensation reference of the current region.

9. The metal plate cutting control method according to claim 8, wherein: After the cutting operation is completed, the actual measurement value is compared with the design requirement, and an actual cutting effect index is calculated; if the actual effect meets the optimization target, the working condition parameters, the optimization parameters and the cutting effect data of this operation are taken as high-quality cases, and are added to the corresponding similar working condition cluster of the dynamic parameter library.

10. A metal sheet cutting control system for implementing the method of any one of claims 1 to 9, characterized in that: It comprises: a working condition analysis module that divides the metal plate cutting working condition into multiple dimensions, collects historical cutting data and extracts features, divides the working condition into similar working condition clusters by using an improved K-means clustering algorithm, constructs a parameter optimization model based on an improved particle swarm algorithm, and establishes a dynamic parameter library; a real-time monitoring module that deploys multiple source sensors on the cutting equipment, collects real-time monitoring data, calculates core cutting state indexes, and establishes a three-level abnormality identification mechanism based on the indexes. The parameter correction module inputs the current working condition parameters into the dynamic parameter library for feature matching, calls an optimization model of a matched cluster to output an initial parameter combination, calculates a deviation index based on real-time monitoring data, constructs an adaptive correction model using an improved ant colony algorithm, and generates corrected optimal cutting control parameters. The control and feedback module converts the corrected parameters into control instructions and sends them to the cutting equipment, monitors and fine-tunes the parameters in real time, dynamically corrects the cutting path using a partition compensation method, evaluates the effect after cutting is completed, and updates the dynamic parameter library.

Images