Online automatic measurement and compensation method for digital manufacturing

Abstract

The invention discloses an on-line automatic measurement and compensation method for digital manufacturing, which comprises the following steps of: scanning the surface of a workpiece to be processed in real time through a high-precision laser scanner to obtain a three-dimensional geometric model of the workpiece; environmental parameters in the laser machining process are monitored in real time through multiple sensors; on the basis of the collected real-time geometrical morphology and environmental parameters, the processing deviation of the workpiece is calculated in real time through a data processing and analyzing module, and a compensation model is generated; according to the calculation result of the compensation model, the central control system adjusts a laser processing path and processing parameters in real time; after each time of processing, a feedback signal of each time of error is used as a basis for updating the compensation model so as to ensure that the compensation model is continuously optimized. The method has the following advantages and effects: dynamic correction and long-term optimization of machining errors can be realized by combining real-time data acquisition and an optimization compensation model technology of a deep learning algorithm, so that the machining precision and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of laser processing intelligent manufacturing, and in particular to an online automatic measurement and compensation method for digital manufacturing. Background Art

[0002] Digital manufacturing is an important production method in modern manufacturing. It uses computer-aided design (CAD), computer-aided manufacturing (CAM) technology, CNC machine tools and automation equipment to achieve high-precision and high-efficiency production goals. In these processes, laser processing is widely used in cutting, marking, engraving and other fields due to its high precision and non-contact characteristics. However, in the actual laser processing process, due to the influence of various factors, such as temperature changes, equipment wear, material non-uniformity, etc., the processing accuracy of the workpiece is often affected, resulting in errors in size, shape, position and other aspects of the final workpiece.

[0003] In order to improve the accuracy and stability of laser processing, the errors in the processing process must be monitored and corrected in real time. Traditional error correction methods usually rely on manual adjustment and empirical judgment, and cannot solve the error problem efficiently and accurately. With the development of intelligent manufacturing technology, error compensation technology based on data feedback has gradually become an effective means to improve processing accuracy. Through real-time data acquisition, intelligent compensation algorithms and dynamic adjustment mechanisms, rapid correction of processing errors can be achieved, improving product accuracy and quality.

[0004] The prior art specifically has the following deficiencies: 1. Insufficient real-time data acquisition and error detection: Most existing technologies rely on traditional measuring equipment and manual inspection, which cannot perform comprehensive and real-time detection of errors in the processing process.

[0005] 2. The accuracy and stability of the compensation effect are not high: Although some intelligent compensation methods have been proposed, the accuracy and stability of the compensation effect are difficult to maintain at a high level due to the lack of real-time feedback and dynamic adjustment.

[0006] 3. Insufficient adaptive ability of the compensation model: Existing compensation models are usually unable to make adaptive adjustments according to different processing conditions, and are often unable to cope with constantly changing processing errors in complex processing environments.

[0007] 4. Lack of feedback mechanism: Most current compensation methods do not have an effective feedback mechanism, resulting in a lack of continuity and stability in the error correction process, and an inability to achieve gradual optimization and improvement after each processing.

[0008] In order to solve the above problems, the present invention proposes an online automatic measurement and compensation method for digital manufacturing, which combines real-time data acquisition and optimization compensation model technology of deep learning algorithm to realize dynamic correction and long-term optimization of processing errors, thereby improving processing accuracy and efficiency. Summary of the invention

[0009] The purpose of the present invention is to provide an online automatic measurement and compensation method for digital manufacturing to solve the problems raised in the background technology.

[0010] The above technical objectives of the present invention are achieved through the following technical solutions: An online automatic measurement and compensation method for digital manufacturing includes the following steps: Step S1, scanning the surface of the workpiece to be processed in real time by a high-precision laser scanner to obtain a three-dimensional geometric model of the workpiece, wherein the laser scanner is equipped with a high-resolution sensor, which can accurately measure the surface of the workpiece at different processing stages, and scan the surface of the workpiece by a laser beam to obtain the geometric shape of the workpiece in real time; Step S2, using a variety of sensors to monitor the environmental parameters during the laser processing in real time, each of the sensors can collect environmental data in the processing area online in real time, and transmit the data to the central control system for subsequent data processing and compensation; Step S3, based on the collected real-time geometric shape and environmental parameters, the processing deviation of the workpiece is calculated in real time through the data processing and analysis module to generate a compensation model, which can predict and quantify the errors in the processing process; Step S4: According to the calculation results of the compensation model, the central control system adjusts the laser processing path and processing parameters in real time to correct the errors in the processing process; Step S5: After each processing, the feedback signal of each error will be used as the basis for updating the compensation model to ensure that the compensation model is continuously optimized.

[0011] By adopting the above technical solutions, high-precision laser scanners can obtain real-time three-dimensional geometric models of workpieces at multiple processing stages and accurately capture changes in the surface morphology of workpieces. Laser scanning has high resolution and sensitivity, and can detect tiny errors that traditional measurement methods cannot identify; through a variety of sensors to monitor environmental parameters in real time, such as temperature, humidity, laser power, etc., it can comprehensively collect various environmental data during the processing process and reduce processing errors caused by environmental fluctuations. For example, temperature changes will affect the expansion or contraction of the workpiece, humidity changes will affect the processing accuracy, and laser power fluctuations will directly affect the processing effect. By real-time monitoring of these environmental changes, the compensation model can adjust the compensation strategy in real time to ensure that the errors in the processing process are minimized; through the analysis of real-time geometric morphology and environmental parameters by the compensation model, the laser processing path and parameters can be automatically adjusted in each processing step, so that the processing process can be self-corrected. This automated compensation method not only improves processing accuracy, but also avoids human intervention and improves production efficiency; error correction is performed at each processing stage to minimize the accumulation of errors. The compensation model can be adjusted in real time according to different processing stages, optimize the processing path and parameters, and avoid the decrease in accuracy caused by error expansion during the processing process.

[0012] It is further configured that the step S3 includes the following steps: Step S31, comparing the real-time geometric shape acquired by the high-precision laser scanner with the original design geometric model of the workpiece; Step S32: Temperature changes can cause expansion or contraction of the workpiece, and power fluctuations of the laser equipment can affect the processing effect. It is necessary to calculate the errors caused by temperature changes and laser power fluctuations; Step S33: Use a deep neural network as the core architecture of the compensation model, and use the workpiece geometry and environmental parameters as input to predict possible errors that may occur during the processing.

[0013] By adopting the above technical solution, by comparing the original design geometric model with the actual geometric model obtained by high-precision laser scanning, the geometric errors of the workpiece can be accurately quantified. These errors usually include shape errors, size errors and position errors. By quantifying the errors into mathematical models, data support can be provided for subsequent compensation strategies. Temperature and laser power fluctuations are important factors affecting processing accuracy. The expansion or contraction of the workpiece caused by temperature changes and the fluctuation of laser power will lead to processing errors. In this method, the changes in temperature and laser power will be corrected by the compensation model to ensure the minimization of processing errors. Through corrections based on environmental data, the influence of these external interference factors can be eliminated in real time during the processing process. Using deep neural networks as the core architecture of the compensation model, it is possible to predict possible errors in the processing process based on real-time geometric morphology and environmental parameter inputs. Deep neural networks have strong learning and generalization capabilities and can be trained with a large amount of processing data to improve the accuracy and adaptability of the model. This technical solution significantly improves the model's ability to handle complex error patterns, especially in a dynamically changing production environment.

[0014] It is further arranged that the step S4 consists of a path compensation calculation step and a laser power and focal length adjustment step; The path compensation calculation step includes the following sub-steps: Step S411, error identification and compensation path modeling; Step S412: calculating the path compensation amount; Step S413: application of path optimization algorithm; The laser power and focal length adjustment step includes the following sub-steps: Step S421, dynamic adjustment of laser power; Step S422: focal length compensation.

[0015] It is further arranged that the path compensation calculation step is specifically as follows: Step S411: The first step of path compensation is error identification and compensation path modeling. Based on the compensation model generated in step S3, the error distribution of the workpiece surface is accurately identified. The error of each processing path will be calibrated and corrected according to the real-time feedback data. In this process, the output of the compensation model is used to analyze the error on each processing path, identify the source of the error, and customize a compensation plan for each error source. Step S412, calculation of path compensation is the core of path compensation calculation. By analyzing the difference between the real-time geometric form and the design model, the path compensation required for each path segment can be calculated. The calculation of path compensation is dynamic and adjusted according to real-time environmental parameters. In this process, the state of the equipment, including the focusing accuracy of the laser beam and the fluctuation factor of the laser power, also needs to be combined. Step S413: Use the optimization algorithm in step S412 to find the optimal compensation path.

[0016] By adopting the above technical solution, the first step of path compensation is to accurately identify the error source of the workpiece surface through the compensation model. The error of each section of the processing path will be calibrated and corrected. With the help of real-time feedback data, the source of the error can be identified in time, and a special compensation strategy can be formulated for each error source. Such precise compensation can avoid unnecessary over-correction and ensure the maximum processing accuracy of each path; the calculation of the path compensation amount not only depends on real-time geometric data, but also needs to take into account various factors such as equipment status and environmental parameters. This dynamic calculation method ensures that the compensation amount of each section of the processing path is updated in real time, so as to cope with any uncertainty in the processing process; the path optimization algorithm can automatically adjust the processing path according to the calculated compensation amount to ensure that the path error is minimized. Through iterative calculation, the optimization algorithm can find the most suitable compensation path, avoid error accumulation, and improve processing accuracy and production efficiency.

[0017] It is further configured that step S413 includes the following sub-steps: Step S4131: firstly generate an initial population according to the design model and real-time data, and these individuals represent different path compensation schemes; Step S4132: for each compensation path, evaluate the fitness of the path by calculating its error size; Step S4133, by selecting a path with higher fitness for crossover, a new compensation path is generated; the crossover process simulates the recombination of genes in nature, and a better path is obtained through crossover; Step S4134: In order to avoid local optimal solutions, randomness is introduced through mutation operations to generate new path solutions, thereby expanding the search space; Step S4135: Through multiple rounds of iterations, the optimization algorithm will eventually find the most suitable compensation path; Through the above optimization algorithm, an accurate compensation solution can be provided for each processing path.

[0018] By adopting the above technical solution, the optimization algorithm can introduce a global search mechanism in the calculation of path compensation, and continuously update the compensation path through multiple rounds of iterations. In the entire path compensation process, it can avoid local optimal solutions and improve the accuracy of the compensation path. Through operations such as natural selection, crossover and mutation, the optimization algorithm can find the most suitable path among multiple possible compensation paths; the optimization algorithm can adapt to different processing error modes and flexibly adjust the compensation path in different processing environments. Especially in complex processing environments, the optimization algorithm can adjust parameters according to feedback signals to ensure that each path can be effectively corrected.

[0019] Further configuration is that the laser power and focal length adjustment steps are: Step S421: The adjustment of the laser power is related to the real-time temperature and humidity data, and the compensation model calculates the compensation amount of the laser power based on the real-time data input.

[0020] By adopting the above technical solution and monitoring the temperature and humidity environmental parameters, the laser power can be adjusted in real time to ensure that the laser equipment works stably under different environmental conditions. Fluctuations in temperature and humidity will affect the laser power. Through power compensation based on real-time data, the processing accuracy can be effectively guaranteed; focal length compensation ensures that the laser beam is always focused on the workpiece surface, and the focal length compensation amount is adjusted through real-time geometric shape and temperature changes. This ensures that the laser beam will not deviate from the target position due to changes in focal length during the processing, thereby maintaining the accuracy of the processing process.

[0021] It is further configured that the step S5 includes the following steps: Step S51, long-term data accumulation and compensation model optimization; Step S52: application of adaptive learning and reinforcement algorithm; Step S53: Closed-loop mechanism of dynamic adjustment and feedback.

[0022] Further configuration is that the step S51 is specifically as follows: The data after each processing is completed will be used as training data for the compensation model, and the compensation model will be gradually optimized by continuously accumulating feedback data; based on machine learning and deep learning algorithms, the compensation model can automatically adjust the correction strategy to cope with different processing error patterns and environmental changes.

[0023] By adopting the above technical solution and training the compensation model through deep neural networks, it is possible to handle more complex error patterns. The network can learn deeper error laws and automatically make compensation adjustments in a constantly changing processing environment. By continuously accumulating data and conducting training, the compensation model can automatically adjust the correction strategy to ensure that errors are minimized in each round of processing, thereby improving production efficiency and product quality.

[0024] Further configuration is that the step S52 is specifically as follows: The compensation model will gradually adapt after multiple processing, and further improve the compensation effect by strengthening the algorithm. The compensation strategy will be adjusted based on the feedback signal. Each feedback provides a basis for adjusting future processing, so that the compensation model can maintain high precision in a constantly changing production environment.

[0025] By adopting the above technical solution and continuously adjusting the compensation strategy and parameters, the compensation model can be gradually optimized. Each feedback will provide a basis for future adjustments to the model to ensure that each processing error is minimized.

[0026] Further configuration is that the step S53 is specifically as follows: The feedback of compensation effect and the optimization of compensation model form a closed-loop mechanism. During a processing process, dynamic adjustments can be made according to real-time feedback signals to ensure the optimal setting of processing paths and parameters. In this process, errors occurring during processing can be automatically identified and corrected. Through the closed-loop mechanism, the processing accuracy can be gradually corrected and improved during each processing process.

[0027] By adopting the above technical solution, through real-time feedback signals, the compensation model can be adjusted after each processing to ensure that the processing path and parameters are always kept in the optimal state and the processing accuracy is maximized. Each feedback and adjustment provides a basis for improving future processing and ensures that the processing accuracy continues to improve over time.

[0028] In summary, the present invention has the following beneficial effects: The online automatic measurement and compensation method for digital manufacturing proposed in the present invention effectively solves the problem of processing errors in the digital manufacturing process through the combination of real-time laser scanning, sensor monitoring, deep learning and optimization algorithms, and has made innovations in real-time, accuracy and adaptability. Through long-term data accumulation and feedback mechanism, the present invention can continuously optimize the compensation model to ensure the stability and improvement of processing accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Schematic diagram of the process of the embodiment. DETAILED DESCRIPTION

[0030] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0031] As attached Figure 1 As shown, this embodiment discloses an online automatic measurement and compensation method for digital manufacturing, comprising the following steps: Step S1, scanning the surface of the workpiece to be processed in real time by a high-precision laser scanner to obtain a three-dimensional geometric model of the workpiece, wherein the laser scanner is equipped with a high-resolution sensor, which can accurately measure the surface of the workpiece at different processing stages, and scan the surface of the workpiece by a laser beam to obtain the geometric shape of the workpiece in real time; The geometry of the workpiece, including but not limited to the workpiece surface shape, thickness, and material properties; Step S2, using a variety of sensors to monitor the environmental parameters during the laser processing in real time, each of the sensors can collect environmental data in the processing area online in real time, and transmit the data to the central control system for subsequent data processing and compensation; Step S3, based on the collected real-time geometric shape and environmental parameters, the processing deviation of the workpiece is calculated in real time through the data processing and analysis module to generate a compensation model, which can predict and quantify the errors in the processing process; Step S4: According to the calculation results of the compensation model, the central control system adjusts the laser processing path and processing parameters in real time to correct the errors in the processing process; Step S5: After each processing, the feedback signal of each error will be used as the basis for updating the compensation model to ensure that the compensation model is continuously optimized.

[0032] Wherein, step S3 includes the following steps: Step S31: Compare the real-time geometric shape acquired by the high-precision laser scanner with the original design geometric model of the workpiece; assuming that the geometric shape of the original design geometric model is represented by G design , and the real-time geometry obtained by scanning is G actual , then the geometric error ϵ geometry It can be expressed as: ϵ geometry =G actual −G design This error usually contains multiple components, including but not limited to shape error, size error, and position error; Step S32: Temperature changes can cause expansion or contraction of the workpiece, and power fluctuations of the laser equipment can affect the processing effect. It is necessary to calculate the errors caused by temperature changes and laser power fluctuations; Assuming the temperature T env and laser power P laser As a known parameter, the processing error caused by temperature change can be expressed as: ΔL=αL0ΔT Among them, α is the thermal expansion coefficient of the material, L0 is the original size, ΔT is the temperature change, and ΔL is the dimensional error caused by temperature change; The processing error caused by laser power fluctuation can be estimated by the following model: ϵ power =k⋅(P laser −P nominal ) Among them, k is the sensitivity of power fluctuation to machining accuracy, P nominal is the ideal power; Step S33, using a deep neural network as the core framework of the compensation model, using the workpiece geometry and environmental parameters as inputs, to predict possible errors that may occur during the processing; Assuming that the output of the neural network is the correction value Δx of the machining error, the output of the model can be expressed as: Δx=f(G actual ,T env ,P laser ) Here f is the neural network model, and the function learned through training can predict the processing error under given input conditions.

[0033] Wherein, step S4 consists of a path compensation calculation step and a laser power and focal length adjustment step; The path compensation calculation step includes the following sub-steps: Step S411, error identification and compensation path modeling; Step S412: calculating the path compensation amount; Step S413: application of path optimization algorithm; The laser power and focal length adjustment step includes the following sub-steps: Step S421, dynamic adjustment of laser power; Step S422: focal length compensation.

[0034] The path compensation calculation steps are as follows: Step S411: The first step of path compensation is error identification and compensation path modeling. Based on the compensation model generated in step S3, the error distribution of the workpiece surface is accurately identified. The error of each processing path will be calibrated and corrected according to the real-time feedback data. In this process, the output of the compensation model is used to analyze the error on each processing path, identify the source of the error, and customize a compensation plan for each error source. Step S412, calculation of path compensation is the core of path compensation calculation. By analyzing the difference between the real-time geometric form and the design model, the path compensation required for each path segment can be calculated. The calculation of path compensation is dynamic and adjusted according to real-time environmental parameters. In this process, the state of the equipment, including the focusing accuracy of the laser beam and the fluctuation factors of the laser power, also needs to be combined, which will affect the accuracy of the processing path. The goal of path compensation calculation is to change the machining path from the original path P orig Adjust to the compensated path P compensated , and its correction value ΔP can be described by the following relationship: ΔP=P orig −P compensated Among them, P compensated is the calculated corrected path, the goal is to minimize the path error; Step S413: Use the optimization algorithm in step S412 to find the optimal compensation path.

[0035] Wherein, step S413 includes the following sub-steps: Step S4131: firstly generate an initial population according to the design model and real-time data, and these individuals represent different path compensation schemes; Step S4132: for each compensation path, evaluate the fitness of the path by calculating its error size; Step S4133, by selecting a path with higher fitness for crossover, a new compensation path is generated; the crossover process simulates the recombination of genes in nature, and a better path is obtained through crossover; Step S4134: In order to avoid local optimal solutions, randomness is introduced through mutation operations to generate new path solutions, thereby expanding the search space; Step S4135: Through multiple rounds of iterations, the optimization algorithm will eventually find the most suitable compensation path; Through the above optimization algorithm, an accurate compensation solution can be provided for each processing path.

[0036] Among them, the laser power and focal length adjustment steps are: Step S421, the adjustment of laser power is related to the real-time temperature and humidity data, and the compensation model calculates the compensation amount of laser power based on the real-time data input; Assume that the laser power P needs to be adjusted laser , power compensation ΔP laser The following will be calculated based on real-time environmental data and the characteristics of the workpiece: ΔP laser = f laser (Tenv ,H env ) Among them, T env and H env Represent temperature and humidity respectively, f laser is the function generated by the compensation model and is used to calculate the adjustment amount of the laser power; The focal length compensation usually depends on the real-time geometric shape, especially the morphological changes of the workpiece surface. The focal length compensation amount Δf focus Adjustable via real-time geometry and temperature changes: Δf focus =g focus (G actual ,T env ) Among them, G actual is the real-time geometry, T env Indicates temperature, g focus It is a function for calculating the focal length compensation. The focal length compensation will ensure that the laser beam is always focused on the workpiece surface and minimize the processing error.

[0037] Wherein, the step S5 includes the following steps: Step S51, long-term data accumulation and compensation model optimization; Step S52: application of adaptive learning and reinforcement algorithm; Step S53: Closed-loop mechanism of dynamic adjustment and feedback.

[0038] Among them, step S51 is specifically as follows: The data after each processing is completed will be used as training data for the compensation model, and the compensation model will be gradually optimized by continuously accumulating feedback data. Based on machine learning and deep learning algorithms, the compensation model can automatically adjust the correction strategy to cope with different processing error modes and environmental changes. Assuming the compensation model f model The model is constructed through a deep neural network. The input of the model is the tool geometry, environmental parameters and equipment status. The output of the model is the error compensation ΔP compensated , which can be expressed as: ΔP compensated =f model (G actual ,T env ,H env ,P laser ,f focus ) Among them, G actual is the real-time geometry, T env and H env Represent temperature and humidity respectively, P laser is the laser power, ffocus is the focal length; The compensation model is continuously optimized through training data.

[0039] Among them, step S52 is specifically as follows: The compensation model will gradually adapt after multiple processing, and further improve the compensation effect by strengthening the algorithm. The compensation strategy is adjusted based on the feedback signal. Each feedback provides a basis for adjusting future processing, so that the compensation model can maintain high precision in a constantly changing production environment. The objective function J can be expressed as: Among them, ϵ t is the error of the tth step, T is the total processing time; by adjusting the parameters of the compensation model, the error tends to be minimized after each processing.

[0040] Among them, step S53 is specifically as follows: The feedback of compensation effect and the optimization of compensation model form a closed-loop mechanism. During a processing process, dynamic adjustments can be made according to real-time feedback signals to ensure the optimal setting of processing paths and parameters. In this process, errors occurring during processing can be automatically identified and corrected. Through the closed-loop mechanism, the processing accuracy can be gradually corrected and improved during each processing process.

[0041] Application Examples Use a high-precision laser scanner to scan the workpiece surface in real time to obtain the three-dimensional geometric shape of the workpiece. The resolution of the high-precision laser scanner is 0.01mm and the scanning frequency is 100Hz. It can obtain the geometric change data of the workpiece in real time during the processing.

[0042] The setting parameters of the high-precision laser scanner are configured as follows: The scanning accuracy is 0.01 mm, the scanning frequency is 100 Hz, the adjustable range of laser power is 2W to 10W, and the standard working power is 5W; the scanning spot size is 0.2 mm, and the scanning mode is continuous scanning mode, which is suitable for workpieces of different shapes and sizes; the workpiece surface needs to be cleaned to reduce the influence of dust and impurities on the scanning results.

[0043] During the processing, temperature and humidity sensors and laser power sensors are used to collect environmental data in the processing area. These sensors are highly accurate and can provide real-time temperature feedback.

[0044] The configuration of temperature and humidity sensor and laser power sensor is: The temperature sensor has an accuracy of ±0.1°C and an operating temperature range of 0°C to 50°C; The humidity sensor has an accuracy of ±2% RH and an operating humidity range of 10% RH to 90% RH; The laser power sensor has an accuracy of ±0.5W and monitors laser power fluctuations; The above sensors are evenly distributed in the processing area to ensure coverage of the entire processing area. The temperature and humidity sensors are located at the four corners of the workpiece processing area to monitor environmental fluctuations during the processing in real time.

[0045] A deep neural network model is used for compensation learning. The model contains three layers of neurons. The input layer accepts the workpiece geometry and environmental data, and the output layer generates the processing error correction value Δx. At the same time, historical data (including geometry and environmental parameters at different processing stages) are used for supervised learning to train the neural network to generate an accurate error correction function; the training data set includes 1,000 groups of samples, and cross-validation and regularization techniques are used during the training process to prevent overfitting.

[0046] Each original path P orig By calculating the error ΔP, it is corrected to the compensation path P compensated , the goal is to minimize the path error.

[0047] The parameters of the optimization algorithm are configured as follows: The population size is 100 candidate solutions, the crossover probability is 80%, the mutation probability is 5%, and the number of iterations is 500 generations.

[0048] The fitness value is evaluated based on the error of each candidate path. The smaller the error, the higher the fitness value.

[0049] More specifically, in this embodiment, five workpieces of different shapes and sizes were selected for the experiment, the workpiece material was aluminum alloy, and the surface was not treated. A high-precision laser scanner was used for real-time scanning, the ambient temperature was controlled at 22°C ± 2°C, the humidity was controlled at 50% ± 5% RH, and the laser power was set to 5W.

[0050] Step S1: Use a high-precision laser scanner to scan the surface of the workpiece to obtain a real-time three-dimensional geometric model. The scanning data is collected at 100 frames per second with a resolution of 0.01 mm.

[0051] Step S2: The temperature and humidity sensors and the laser power sensors monitor the temperature, humidity and laser power of the processing area in real time, and transmit the data to the central control system in real time.

[0052] Step S3: Based on the scanning data and environmental parameters, the data processing unit calculates the errors in the processing process and generates a compensation model.

[0053] Step S4: adjusting the processing path and laser power in real time according to the compensation model to ensure that the processing error is minimized.

[0054] Step S5: After the processing is completed, the feedback data is used to update the compensation model and optimize the correction strategy.

[0055] The results of this embodiment are: The initial machining error was 0.3 mm. By modifying the compensation model, the error was reduced to 0.05 mm, an improvement of 83%.

[0056] The influence of laser power fluctuation on processing accuracy is reduced to less than 5% through dynamic compensation adjustment, and focal length compensation effectively reduces the error caused by focal length change.

[0057] After processing 10 workpieces continuously, the error correction capability of the compensation model was further optimized, and the final error was kept within 0.02 mm, ensuring the consistency and stability of the product.

[0058] This specific embodiment is merely an explanation of the present invention and is not a limitation of the present invention. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed. However, as long as they are within the scope of the claims of the present invention, they are protected by the patent law.

Claims

1. An online automatic measurement and compensation method for digital manufacturing, characterized in that: The following steps are included: Step S1, scanning the surface of the workpiece to be processed in real time by a high-precision laser scanner to obtain a three-dimensional geometric model of the workpiece, wherein the laser scanner is equipped with a high-resolution sensor, which can accurately measure the surface of the workpiece at different processing stages, and scan the surface of the workpiece by a laser beam to obtain the geometric shape of the workpiece in real time; Step S2, using a variety of sensors to monitor the environmental parameters during the laser processing in real time, each of the sensors can collect environmental data in the processing area online in real time, and transmit the data to the central control system for subsequent data processing and compensation; Step S3, based on the collected real-time geometric shape and environmental parameters, the processing deviation of the workpiece is calculated in real time through the data processing and analysis module to generate a compensation model, which can predict and quantify the errors in the processing process; Step S4: According to the calculation results of the compensation model, the central control system adjusts the laser processing path and processing parameters in real time to correct the errors in the processing process; Step S5: After each processing, the feedback signal of each error will be used as the basis for updating the compensation model to ensure that the compensation model is continuously optimized.

2. The online automatic measurement and compensation method for digital manufacturing according to claim 1, characterized in that: The step S3 includes the following steps: Step S31, comparing the real-time geometric shape acquired by the high-precision laser scanner with the original design geometric model of the workpiece; Step S32: Temperature changes can cause expansion or contraction of the workpiece, and power fluctuations of the laser equipment can affect the processing effect. It is necessary to calculate the errors caused by temperature changes and laser power fluctuations; Step S33: Use a deep neural network as the core architecture of the compensation model, and use the workpiece geometry and environmental parameters as input to predict possible errors that may occur during the processing.

3. The online automatic measurement and compensation method for digital manufacturing according to claim 1 is characterized in that: The step S4 is composed of a path compensation calculation step and a laser power and focal length adjustment step; The path compensation calculation step includes the following sub-steps: Step S411, error identification and compensation path modeling; Step S412: calculating the path compensation amount; Step S413: application of path optimization algorithm; The laser power and focal length adjustment step includes the following sub-steps: Step S421, dynamic adjustment of laser power; Step S422: focal length compensation.

4. The online automatic measurement and compensation method for digital manufacturing according to claim 3 is characterized in that: The path compensation calculation steps are specifically as follows: Step S411: The first step of path compensation is error identification and compensation path modeling. Based on the compensation model generated in step S3, the error distribution of the workpiece surface is accurately identified. The error of each processing path will be calibrated and corrected according to the real-time feedback data. In this process, the output of the compensation model is used to analyze the error on each processing path, identify the source of the error, and customize a compensation plan for each error source. Step S412, calculation of path compensation is the core of path compensation calculation. By analyzing the difference between the real-time geometric form and the design model, the path compensation required for each path segment can be calculated. The calculation of path compensation is dynamic and adjusted according to real-time environmental parameters. In this process, the state of the equipment, including the focusing accuracy of the laser beam and the fluctuation factor of the laser power, also needs to be combined. Step S413: Use the optimization algorithm in step S412 to find the optimal compensation path.

5. The online automatic measurement and compensation method for digital manufacturing according to claim 4 is characterized in that: The step S413 includes the following sub-steps: Step S4131: firstly generate an initial population according to the design model and real-time data, and these individuals represent different path compensation schemes; Step S4132: for each compensation path, evaluate the fitness of the path by calculating its error size; Step S4133, by selecting a path with higher fitness for crossover, a new compensation path is generated; the crossover process simulates the recombination of genes in nature, and a better path is obtained through crossover; Step S4134: In order to avoid local optimal solutions, randomness is introduced through mutation operations to generate new path solutions, thereby expanding the search space; Step S4135: Through multiple rounds of iterations, the optimization algorithm will eventually find the most suitable compensation path; Through the above optimization algorithm, an accurate compensation solution can be provided for each processing path.

6. The online automatic measurement and compensation method for digital manufacturing according to claim 3 is characterized in that: The laser power and focal length adjustment steps are as follows: Step S421: The adjustment of the laser power is related to the real-time temperature and humidity data, and the compensation model calculates the compensation amount of the laser power based on the real-time data input.

7. The online automatic measurement and compensation method for digital manufacturing according to claim 1, characterized in that: The step S5 includes the following steps: Step S51, long-term data accumulation and compensation model optimization; Step S52: application of adaptive learning and reinforcement algorithm; Step S53: Closed-loop mechanism of dynamic adjustment and feedback.

8. The online automatic measurement and compensation method for digital manufacturing according to claim 7, characterized in that: The step S51 is specifically as follows: The data after each processing is completed will be used as training data for the compensation model, and the compensation model will be gradually optimized by continuously accumulating feedback data; based on machine learning and deep learning algorithms, the compensation model can automatically adjust the correction strategy to cope with different processing error patterns and environmental changes.

9. The online automatic measurement and compensation method for digital manufacturing according to claim 7, characterized in that: The step S52 is specifically as follows: The compensation model will gradually adapt after multiple processing, and further improve the compensation effect by strengthening the algorithm. The compensation strategy will be adjusted based on the feedback signal. Each feedback provides a basis for adjusting future processing, so that the compensation model can maintain high precision in a constantly changing production environment.

10. The online automatic measurement and compensation method for digital manufacturing according to claim 7, characterized in that: The step S53 is specifically as follows: The feedback of compensation effect and the optimization of compensation model form a closed-loop mechanism. During a processing process, dynamic adjustments can be made according to real-time feedback signals to ensure the optimal setting of processing paths and parameters. In this process, errors occurring during processing can be automatically identified and corrected. Through the closed-loop mechanism, the processing accuracy can be gradually corrected and improved during each processing process.

Images