Intelligent control method and system for computer hydraulic pipe bending machine
Through intelligent control methods and sensing technology, combined with intelligent decision-making modules and data processing technology, high-precision and high-efficiency control of hydraulic pipe bending machines is achieved, solving the problem that traditional control methods cannot meet the precise needs of modern industries, and improving product quality and productivity.
Patent Information
- Application Number
- CN202411205273.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-30
AI Technical Summary
The traditional hydraulic pipe bending machine control method cannot achieve high-precision and efficient pipe bending operations. It depends on the experience and skills of the operator and cannot meet the modern industry's needs for accuracy and efficiency.
Using intelligent control methods, by reading straight tube specification information and bending standards, calibrating effective point clouds, interactive homologous bending records, mining abnormal defect categories, determining characteristic denaturation thresholds and mapping parameter control thresholds, combining intelligent decision-making modules for hierarchy It can analyze the control of the pre-control strategy, and combine sensing technology and data processing technology to perform real-time control and feedback adjustment to achieve accurate control of the pipe bending process.
Accurate control and optimization of the pipe bending process is achieved, product quality and production efficiency are improved, and the problem that traditional control methods cannot achieve high-precision and high-efficiency pipe bending operations.
Smart Images

Figure CN118719886B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent control technology, and in particular to an intelligent control method and system for a computer hydraulic pipe bending machine. Background Art
[0002] With the continuous progress of modern machinery manufacturing industry, computer hydraulic pipe bending machines are increasingly used in various industrial fields, especially in heavy industries such as automobiles, aerospace, and shipbuilding. However, with the continuous improvement of pipe bending technology and the diversification of product demand, the traditional hydraulic pipe bending machine control method has gradually revealed its limitations. Traditional control methods often rely on the experience and skills of operators and cannot achieve high-precision and high-efficiency pipe bending operations. Summary of the invention
[0003] The present application provides an intelligent control method and system for a computer hydraulic pipe bending machine, which is used to solve the technical problem that the prior art cannot achieve high-precision and high-efficiency pipe bending operations.
[0004] In view of the above problems, the present application provides an intelligent control method and system for a computer hydraulic pipe bending machine.
[0005] The first aspect of the present application provides an intelligent control method for a computer hydraulic pipe bender, the method comprising:
[0006] Read straight pipe specification information and bending standards, calibrate valid point clouds, wherein the valid point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the valid cross-section position; interactively record homologous bending, mine abnormal defect classes, and determine feature degeneration threshold-mapping parameter control threshold, perform matrix conversion to determine rigid constraints; combine the valid point cloud with the rigid constraints, and perform hierarchical control analysis in combination with the intelligent decision-making module to determine the pre-control strategy, wherein the hierarchical standard is the pipe bending device control layer-pump station control layer-circuit control layer; the computer control system responds to the pre-control strategy and performs bending control on the computer hydraulic pipe bending machine, and the intelligent decision-making module is built in the computer control system; synchronously perform control monitoring and return control data to the digital feedback device, perform deviation analysis based on the pipe bending deformation data and the equipment parameter control data, determine the deviation detection result, and use the standard deviation interval as the constraint positioning control point; the computer control system responds to the digital feedback device to perform bending feedback control.
[0007] The second aspect of the present application provides an intelligent control system for a computer hydraulic pipe bender, the system comprising:
[0008] An effective point cloud calibration unit, the effective point cloud calibration unit reads the straight pipe specification information and the bending standard, and calibrates the effective point cloud, wherein the effective point cloud is marked with Gaussian curvature, including stress points and key points, which are distributed at any annular position of the effective section position; a rigid constraint condition determination unit, the rigid constraint condition determination unit interacts with the homologous bending records, mines abnormal defect classes, and determines the feature degeneration threshold-mapping parameter control threshold, and performs matrix conversion to determine the rigid constraint condition; a pre-control strategy determination unit, the pre-control strategy determination unit combines the effective point cloud and the rigid constraint condition, and combines the intelligent decision-making module to perform hierarchical control analysis to determine the pre-control strategy, wherein the hierarchical The standard is the pipe bending device control layer-pump station control layer-circuit control layer; a bending control unit, which controls the computer control system to respond to the pre-control strategy, and performs bending control on the computer hydraulic pipe bending machine, and the intelligent decision-making module is built in the computer control system; a control point positioning unit, which synchronously performs control monitoring and sends back control data to the digital feedback device, executes deviation analysis based on the bending deformation data and the equipment parameter control data, determines the deviation detection result, and locates the control point with the standard deviation interval as the constraint; a bending feedback control unit, which controls the computer control system to respond to the digital feedback device and perform bending feedback control.
[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0010] The present application reads straight pipe specification information and bending standards, calibrates valid point clouds, wherein valid point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the valid cross-section position; interactive homologous bending records are used to mine abnormal defect classes, and characteristic degeneration thresholds-mapping parameter control thresholds are determined, and matrix transformation is performed to determine rigid constraints; hierarchical control analysis is performed in combination with valid point clouds and rigid constraints, and an intelligent decision-making module is used to determine a pre-control strategy, wherein the hierarchical standard is the bending device control layer-pump station control layer-circuit control layer; a computer control system responds to the pre-control strategy and performs bending control on a computer hydraulic pipe bender, and an intelligent decision-making module is built into the computer control system; control monitoring is performed simultaneously and control data is transmitted back to a digital feedback device, deviation analysis is performed based on the bending deformation data and the equipment parameter control data, deviation detection results are determined, and control points are positioned with the standard deviation interval as a constraint; the computer control system responds to the digital feedback device and performs bending feedback control. The present invention solves the technical problem that the prior art cannot achieve high-precision and high-efficiency pipe bending operations. By integrating advanced sensing technology, data processing technology and intelligent control algorithms, the present invention realizes precise control and optimization of the pipe bending process, thereby achieving the technical effect of improving product quality and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0012] Figure 1 A schematic flow chart of an intelligent control method for a computer hydraulic pipe bending machine provided in an embodiment of the present application;
[0013] Figure 2 A schematic diagram of the structure of an intelligent control system for a computer hydraulic pipe bending machine provided in an embodiment of the present application.
[0014] Explanation of the reference numerals: effective point cloud calibration unit 11 , rigid constraint condition determination unit 12 , pre-control strategy determination unit 13 , bending control unit 14 , control point positioning unit 15 , bending feedback control unit 16 . DETAILED DESCRIPTION
[0015] The present application provides an intelligent control method and system for a computer hydraulic pipe bender, aiming to solve the technical problem that the existing technology cannot achieve high-precision and high-efficiency pipe bending operations. By integrating advanced sensing technology, data processing technology and intelligent control algorithms, precise control and optimization of the pipe bending process are achieved, achieving the technical effect of improving product quality and production efficiency.
[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0017] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.
[0018] Embodiment 1
[0019] like Figure 1 As shown, the present application provides an intelligent control method for a computer hydraulic pipe bender, the method comprising:
[0020] Step S100: reading straight pipe specification information and bending standards, and calibrating effective point clouds, wherein the effective point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the effective cross-section position;
[0021] In the embodiment of the present application, the straight pipe specification information, such as material, diameter, wall thickness, etc., is quickly and accurately retrieved and extracted through the digital information retrieval system within the enterprise. The industry standard database is accessed to obtain the specific standards for bending. The bending standards include the target bending radius, bending angle, and quality requirements after bending.
[0022] Then use CAD or CAE software to create a digital model of the straight pipe. On the digital model, simulate the expected bending process to predict possible deformation and stress distribution. Use finite element analysis software to simulate the mechanical behavior of the straight pipe during the bending process. Analyze and determine the stress points and key points of the pipe during the bending process, that is, those areas where stress concentration or deformation is most likely to occur. Based on the results of the finite element analysis, use the algorithm to automatically determine the location of stress points and key points, find areas of stress concentration, identify points with the largest deformation, etc. Once these points are determined by the algorithm, these points are marked as valid point clouds. The valid point cloud is marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the effective section position.
[0023] Step S200: Interacting homologous bending records, mining abnormal defect classes, and determining feature degeneration thresholds-mapping parameter control thresholds, performing matrix conversion to determine rigid constraint conditions;
[0024] In an embodiment of the present application, homologous bending records are obtained from the production management system, that is, multiple record data from the same bending process are obtained. The homologous bending records include various parameters in the bending process, such as pressure, temperature, time, and the shape and size of the product after bending.
[0025] Data analysis techniques, such as cluster analysis and anomaly detection, are used to mine the bending records and identify abnormal defects, which are caused by material, process or equipment problems.
[0026] Then select the features that have a great impact on product quality, such as bending angle, maximum deformation, etc. Obtain these feature values directly from the bending records by calculation or measurement. Perform statistical analysis on the same source bending records, calculate the average value, standard deviation and other statistics of each feature, and set the threshold of feature variation in combination with industry standards and expert experience, that is, when a feature exceeds the normal range, the product is considered to be defective.
[0027] Then, linear or nonlinear regression analysis is performed using statistical software such as SPSS and Excel to establish a mapping relationship between the characteristic degeneration threshold and the mapping parameter control threshold.
[0028] Finally, through physical simulation methods such as finite element analysis, the deformation and stress distribution of the product under different production process parameters are predicted to determine the rigid constraint conditions.
[0029] Step S300: combining the effective point cloud and the rigid constraint conditions, and combining the intelligent decision-making module to perform hierarchical control analysis and determine the pre-control strategy, wherein the hierarchical standard is the pipe bending device control layer - pump station control layer - circuit control layer;
[0030] In the embodiment of the present application, the effective point cloud data is fused with the rigid constraints, including combining the key information in the point cloud data with the rigid constraints, such as the maximum deformation and stress limit of the material, to form a comprehensive data set. The data is preprocessed, including data cleaning, standardization and normalization, to ensure data quality and consistency.
[0031] The intelligent decision-making module is pre-built. When building it, we first collect historical data, including valid point cloud data, rigid constraints, key parameters in the pipe bending process, such as bending angle, pressure, temperature, and corresponding quality control results, such as the quality grade of the pipe bending, whether there are defects, etc. These data are used as training sets and validation sets. Features related to the pipe bending process, such as point density, curvature, shape features, etc., are extracted from the valid point cloud data. At the same time, rigid constraints, such as the maximum deformation of the material, stress limit, etc., are also used as feature inputs.
[0032] Then select the convolutional neural network and use the historical data as the training set for training. During the training process, provide a set of labeled data, i.e., quality control results, as training samples. The model makes predictions by learning the relationship between the features and labels in these samples. During the training process, the parameters of the model are continuously adjusted through optimization algorithms such as gradient descent to minimize the prediction error. Use the validation set to evaluate the performance of the model. Evaluate the prediction ability of the model by calculating indicators such as accuracy, recall, and F1 score. If the performance is poor, adjust the model's hyperparameters, such as learning rate, regularization coefficient, etc. or use a more complex model structure to optimize the performance. In addition, use techniques such as cross-validation to further evaluate the stability and generalization ability of the model. Finally, integrate the trained model into the intelligent decision-making module.
[0033] When the intelligent decision-making module performs hierarchical control analysis and determines the pre-control strategy, at the pipe bending device control layer, the intelligent decision-making module adjusts the key parameters of the pipe bending device, such as bending angle, bending speed, etc., in real time according to the point cloud data and rigid constraints to ensure that the pipe bending process meets the quality requirements. At the pump station control layer, the pump station is responsible for providing the hydraulic power required in the pipe bending process. The intelligent decision-making module adjusts the working pressure and flow of the pump station in real time according to the needs of the pipe bending device to ensure that the power requirements in the pipe bending process are met. Finally, based on the output of the intelligent decision-making module, the pre-control strategy is determined.
[0034] Step S400: the computer control system responds to the pre-control strategy to perform bending control on the computer hydraulic pipe bender, and the intelligent decision-making module is built into the computer control system;
[0035] In the embodiment of the present application, the computer control system sends a control instruction to the hydraulic pipe bender according to the pre-control strategy. The hydraulic pipe bender starts to perform the pipe bending action according to the received instruction.
[0036] The intelligent decision module is built into the computer control system to monitor various parameters in the pipe bending process in real time. If it is found that the actual pipe bending parameters deviate from the range set by the pre-control strategy, the intelligent decision module will be activated immediately. The intelligent decision module dynamically adjusts the pre-control strategy based on real-time data and algorithm analysis. The adjusted control strategy is resent to the hydraulic pipe bender to restore it to normal bending state.
[0037] Step S500: synchronously perform control monitoring and transmit control data back to the digital feedback device, perform deviation analysis based on the bending tube deformation data and the equipment parameter control data, determine the deviation detection result, and use the standard deviation interval as a constraint to locate the control point;
[0038] In the embodiment of the present application, during the production process of the bent pipe, the deformation data and equipment parameter control data of the bent pipe are collected in real time through sensors. The deformation data of the bent pipe includes the bending angle, deformation amount, etc., and the equipment parameter control data includes pressure, temperature, equipment status, etc. The collected data is then synchronously transmitted to the digital feedback device.
[0039] After receiving the data, the digital feedback device performs a deviation analysis. The deviation analysis mainly compares the real-time bending deformation data with the equipment parameter control data and the preset standard value or historical data. By using statistical methods to analyze the data, such as calculating the mean, standard deviation, coefficient of variation, etc., the deviation of the bending deformation is obtained, that is, whether the deformation deviates from the expected standard range, which is pre-set according to production needs. According to the results of the deviation analysis, it is determined whether the deformation of the bending pipe exceeds the acceptable range, thereby obtaining the deviation detection result.
[0040] When the control point is located with the standard deviation interval as the constraint, the standard deviation interval is set, which is a reasonable deformation deviation interval set according to the requirements of product quality and production process. When the deformation data exceeds this interval, the deformation is returned to the acceptable range by adjusting the process parameters or equipment settings, and these adjusted points are regarded as control points.
[0041] Step S600: The computer control system responds to the digital feedback device to perform bending feedback control.
[0042] In the embodiment of the present application, the computer control system first receives the deflection detection result sent by the digital feedback device. This information includes real-time data of the bending deformation, deviation from the standard value, whether the deformation exceeds the acceptable range, etc. The computer control system then analyzes the received deflection detection results to determine whether the deformation exceeds the preset standard deviation interval, as well as the direction and degree of deviation. Based on the deflection detection results, the computer control system will calculate the required adjustment amount, including fine-tuning of process parameters such as pressure, temperature, and bending speed, to ensure that the bending deformation returns to an acceptable range.
[0043] Finally, the computer control system automatically adjusts the settings of the tube bending machine based on the calculated adjustment amount. For example, if the deformation deviation is too large, the system may reduce the pressure or adjust the bending speed to reduce the deformation.
[0044] Furthermore, step S100 in the method provided in the application embodiment also includes:
[0045] Establishing a curvature calculator, wherein the curvature calculator is trained based on a Gaussian curvature calculation mechanism and is in communication connection with a computer control system;
[0046] Based on the curvature calculator, traverse the valid point cloud to perform Gaussian curvature calculation to determine a point cloud curvature set;
[0047] In combination with the bending standard, a standard deviation interval between the adjacent point cloud and the same-frequency point cloud based on the spatial position is set, wherein the adjacent point cloud is distributed along the axial direction of the tube, and the same-frequency point cloud is distributed along the tangential direction of the tube;
[0048] Based on the point cloud curvature set and the standard deviation interval, mapping association is performed with the valid point cloud.
[0049] In an embodiment of the present application, a curvature calculator is constructed based on a Gaussian curvature calculation mechanism. Gaussian curvature describes the degree of convexity and concavity of a surface at a certain point and is an important indicator for describing the local geometric shape of the surface. When establishing a curvature calculator, a specific mathematical formula is used, such as calculating the surface normal, the principal curvature, and other methods to calculate the Gaussian curvature of each point in the point cloud. In order to train the curvature calculator, a machine learning algorithm is used to optimize the calculation process of the Gaussian curvature to improve accuracy and efficiency. During training, a large amount of point cloud data is collected as a training set for training the curvature calculator. Using the known Gaussian curvature calculation method, combined with machine learning technology, the curvature calculator is trained so that it can automatically and accurately calculate the Gaussian curvature of each point in the point cloud.
[0050] Next, we use the established curvature calculator to traverse each point in the valid point cloud and calculate the Gaussian curvature. Through this step, we get the point cloud curvature set containing the Gaussian curvature of all points.
[0051] When setting the standard deviation range, key quality control indicators are extracted from the press bending standards, such as the allowable range of the bending radius, the accuracy requirements of the bending angle, etc.
[0052] Adjacent point cloud refers to the point cloud set distributed along the axial direction of the pipe. These points reflect the deformation of the pipe in the length direction and during the bending process. Same frequency point cloud refers to the point cloud set distributed along the tangent direction of the pipe, that is, the circumferential direction of the pipe. These points are used to describe the cross-sectional shape of the pipe and the deformation in the circumferential direction.
[0053] According to the bending standard and the initial shape of the pipe, determine the reference value of each quality control indicator. For example, for the bending radius, the reference value is the ideal bending radius specified in the standard. Based on the bending standard and actual production needs, calculate the maximum and minimum deviation range allowed for each quality control indicator. This usually involves an assessment of process capabilities, considerations of product safety, and customer requirements for product quality. For adjacent point clouds, set the allowable deviation range for the position of each point along the axial direction of the pipe, such as length change, axial displacement, etc. For the same-frequency point cloud, set the allowable deviation range for the position of each point along the tangential direction of the pipe, such as the deformation of the cross-sectional shape, the displacement in the circumferential direction, etc. Through the above steps, set the standard deviation range.
[0054] When mapping and associating with the valid point cloud based on the point cloud curvature set and the standard deviation interval, for each point in the point cloud curvature set, find its corresponding point in the valid point cloud by matching the spatial coordinates. Associate the curvature value of each point with its spatial coordinates in the valid point cloud. In this way, each spatial coordinate point has a corresponding curvature value. For each associated point, according to its curvature value and spatial coordinates, calculate its deviation by comparing the curvature value with the standard deviation interval. According to the deviation, the point is classified as meeting the standard or exceeding the standard. Through the above process, the mapping association is completed.
[0055] Furthermore, step S200 in the method provided in the application embodiment also includes:
[0056] Taking defect type as the first-level clustering standard and visualization feature as the second-level clustering standard, traversing the homologous bending records to perform hierarchical clustering processing, and determining hierarchical clustering clusters;
[0057] Traverse one layer of clusters, perform defect processing, and determine the target defect type;
[0058] Traverse the second-layer clusters, identify the commonality of features and perform lower limit processing as the feature mutation threshold, where the lower limit principle is the minimum eigenvalue;
[0059] The target defect type is mapped to the characteristic variability threshold, and an associated control parameter value based on the characteristic variability threshold is determined as the mapping parameter control threshold, wherein the mapping relationship is one-to-many.
[0060] In the embodiment of the present application, firstly, all data are clustered at the first level according to the defect types in the historical bending records. Defect types include insufficient bending, excessive bending, cracks, etc. Clustering is used to group records with the same or similar defect types together. Based on the first level of clustering, the data within each defect type is clustered at the second level according to visual features such as shape, color, texture, etc. The different manifestations or severity of each defect type are further subdivided through the second level of clustering.
[0061] After two layers of clustering, multiple hierarchical clusters are obtained, each cluster represents a specific defect type and specific visualization features under this type.
[0062] Each cluster obtained by the first-level clustering is traversed, and the data in each cluster is analyzed through statistical analysis and pattern recognition to determine the defect type represented by the cluster.
[0063] After determining the target defect type, traverse each subcluster obtained by the second-level clustering. Analyze the data in each subcluster and identify their common features. These features include shape, size, etc. For each common feature, find the minimum value of the feature in all data as the threshold for the lower limit processing, that is, the feature mutation threshold.
[0064] For each target defect type that has been determined by hierarchical clustering, its key visual features are listed. Through lower-limit processing, the minimum acceptable value of each key feature, i.e., the feature degeneration threshold, is determined. These feature degeneration thresholds are matched with the target defect type to form a mapping relationship.
[0065] When determining the associated control parameter values, for each target defect type, analyze the potential causes that lead to its occurrence, such as improper machine settings, material property problems, or ambient temperature effects. Identify the control parameters directly related to these potential causes from the production process. These parameters can be adjusted during the production process and have a direct impact on product quality. Based on the feature variability threshold, through historical data analysis, determine how to adjust the relevant control parameters to reduce or avoid defects when the feature value is below the threshold.
[0066] The feature degeneration threshold is associated with the control parameter value to form a mapping parameter control threshold. This mapping relationship defines which control parameters should be adjusted and by how much when a key feature value is lower than its degeneration threshold to achieve the purpose of reducing defects. Since the same defect may be caused by multiple different factors, a defect type may correspond to multiple key features and corresponding control parameters.
[0067] Furthermore, the method of performing matrix transformation to determine the rigid constraint condition further includes:
[0068] A three-dimensional matrix is established with defect types as matrix rows, features as matrix columns, and control parameters as matrix verticals;
[0069] The mapping relationship is identified, and a distribution conversion based on the three-dimensional matrix is performed to determine a constraint matrix, and the constraint matrix is used as the rigid constraint condition.
[0070] In the embodiment of the present application, the defect types are the rows of the matrix, and each defect type occupies one row. The features are the columns of the matrix, and each key feature occupies one column. A vertical dimension is added, that is, the depth or height of the matrix, to represent different control parameters. In this way, each control parameter forms a matrix layer. At the intersection of each defect type and feature, fill in the control parameter value associated with the defect type and feature. If a defect type is not directly associated with a feature, the position is left blank. A three-dimensional matrix is established through the above steps.
[0071] Analyze the data using statistical methods to find the correlation between features and control parameters, and between defect types and features. After identifying the mapping relationship, perform distribution transformation, first standardize or normalize the data to ensure that all data are on the same scale. If the data dimension is too high, use methods such as principal component analysis to reduce the data dimension. Finally, select the features and control parameters that have the greatest impact on the defect type.
[0072] Based on the above distribution transformation and mapping relationship analysis, upper and lower thresholds are set for features and control parameters according to the mapping relationship and product quality requirements. These thresholds are organized into a two-dimensional constraint matrix, where rows represent different features or control parameters and columns represent their constraints. The constraint matrix is used as a rigid constraint to guide production.
[0073] Furthermore, step S300 in the method provided in the application embodiment also includes:
[0074] Identify the relative spatial position between the pipe bending device and the pre-stressed pipe bending, wherein the relative spatial position is marked with a contact partition; perform a force analysis based on the relative spatial position in combination with the rigid constraint condition, perform a stress extension analysis in combination with the effective point cloud, and determine a layer of control information of the pipe bending device;
[0075] Taking the control information of the first layer as the control standard and based on the inter-layer control operation logic, performing hierarchical calculation to determine the pre-control strategy;
[0076] Among them, the stress epitaxy trend is the epitaxy analysis benchmark, and the stress epitaxy trend is obtained based on the homologous bending record.
[0077] In the embodiment of the present application, a high-precision sensor, such as a laser distance sensor or a position sensor, is first used to measure the relative distance and angle between the pipe bending device and the pre-pressed pipe bending. The image is captured by a high-definition camera, and then the contact zone between the pipe bending device and the pipe is identified through image processing and analysis.
[0078] When conducting a force analysis based on relative spatial position, the force conditions during the pipe bending process are analyzed in detail through finite element analysis. By inputting rigid constraints, such as the maximum deformation and stress limit of the material, into the finite element model, the stress distribution and deformation during the pipe bending process are simulated. Then, the effective point cloud data is used for stress extrapolation analysis. By analyzing the propagation trend and distribution of stress in the pipe, the stress concentration and deformation risks that may occur during the pipe bending process are predicted. Based on the results of the force analysis and stress extrapolation analysis, optimization algorithms such as linear programming and genetic algorithms are used to determine the optimal control parameters of the pipe bending device, including the applied force, bending speed, bending angle, etc., and this information is used as a layer of control information for the pipe bending device.
[0079] Taking the control information of the first layer as the control standard and based on the logic of the inter-layer control operation, when performing hierarchical calculation, the pipe bending device control layer sets the specific control target of the pipe bending device, such as bending angle, speed and applied force, according to the control information of the first layer. At the same time, it receives real-time feedback from the pipe bending device and adjusts the control instructions according to these feedbacks to ensure that the pipe bending process meets the expected standards. The pump station control layer receives and processes instructions from the pipe bending device control layer, which include adjusting the pressure, flow and other parameters of the pump station. These instructions are processed according to the current status and control logic of the pump station to ensure that the pipe bending device receives the required hydraulic power support. The circuit control layer receives instructions from the pump station control layer and converts them into specific circuit control signals, such as adjusting the speed of the motor, the on and off of the switch, etc. At the same time, it monitors the status of the circuit to ensure that the circuit operates within a safe range and takes necessary protection measures. Through the above hierarchical calculation process, the pre-control strategy is determined.
[0080] The stress extrapolation trend is obtained by analyzing the homologous bending records. The stress extrapolation trend is extremely important for predicting and controlling stress concentration during the bending process and preventing excessive deformation or fracture of the material. The stress extrapolation trend serves as the benchmark for extrapolation analysis, that is, when formulating the control strategy, various parameters in the bending process are predicted and adjusted based on this trend. For example, if the stress extrapolation trend shows that the stress will increase sharply at a certain bending angle, then it is necessary to reduce the applied force or adjust the bending speed near this angle to prevent material damage.
[0081] Furthermore, step S500 in the method provided in the application embodiment also includes:
[0082] Identifying the bending deformation data, and determining the actual curvature distribution based on the effective point cloud;
[0083] Based on the actual curvature distribution and the identified Gaussian curvature, performing an overall trend analysis and an adjacent difference trend analysis to determine a first deviation detection result;
[0084] Determine a second deviation detection result based on the actual control response deviation of the parameters of the pre-control strategy;
[0085] Based on the valid point cloud, single-point distortion identification and distortion trend prediction are performed to determine a third deviation detection result;
[0086] The first deflection detection result, the second deflection detection result, and the third deflection detection result are integrated to determine the deflection detection result.
[0087] In an embodiment of the present application, first, necessary preprocessing is performed on the acquired valid point cloud data, including denoising, filling missing data and smoothing, to improve data quality. Using the preprocessed point cloud data, the deformation data of the bend is identified by comparing and analyzing the offset of each point relative to the theoretical position or initial state. For each point in the point cloud, its neighboring point set is selected, and a suitable algorithm, such as quadratic surface fitting, is used to perform local surface fitting. Based on the fitted local surface, the various curvature parameters of each point, such as the principal curvature, the average curvature, etc., are calculated to obtain the actual curvature distribution of the entire bend.
[0088] When performing the analysis of the assembly trend and adjacent difference trend based on the actual curvature distribution and the identified Gaussian curvature, first determine the theoretical curvature distribution and the identified Gaussian curvature of each stage as the reference standard. These standards are based on design requirements and simulation analysis, and represent the ideal curvature state that the bend should achieve at a specific stage. Next, compare the actual curvature distribution with these stage standards. When performing the assembly trend analysis, compare whether the overall trend of the actual curvature distribution is consistent with the theoretical trend. For example, at a certain processing stage, the curvature should theoretically gradually increase or decrease to a certain value, forming a specific overall change trend. If the actual curvature distribution is ahead of or behind the stage standard due to factors such as force response in actual control, but the overall change trend is still consistent with the theoretical trend, the assembly trend is considered normal.
[0089] The curvature difference between point clouds, i.e. the change of local curvature, is determined during the docking difference trend analysis. During normal processing or deformation, the curvature difference between adjacent point clouds should be kept within a reasonable range, and the change trend should be stable without sudden changes or abnormal fluctuations. If the curvature difference between actual point clouds conforms to this normal trend, it is still judged to be normal even if the actual curvature distribution deviates from the stage standard.
[0090] Based on the above analysis, the first bias detection result is obtained.
[0091] When determining the second deviation detection result, the parameters set in the pre-control strategy are first obtained based on the deviation of the actual control response of the parameters of the pre-control strategy. These parameters include the force, speed, temperature, etc. during the processing. At the same time, the data of the actual control process is collected and compared with the parameters in the pre-control strategy to analyze the deviation of the actual control response. If the actual control response is basically consistent with the parameters of the pre-control strategy and there is no obvious deviation, the second deviation detection result is judged to be normal. If there is a significant deviation in the actual control response, which exceeds the preset tolerance range, the second deviation detection result is judged to be abnormal.
[0092] Based on the effective point cloud, single-point distortion recognition and distortion trend prediction are performed to determine the third deviation detection result. In point distortion recognition, the effective point cloud data is analyzed one by one for the position, shape and other characteristics of each point to identify whether there are distortion points, that is, points with obvious position or shape abnormalities compared with the surrounding points. In the change trend prediction, the identified distortion points are combined with the distribution and morphology of the surrounding points to predict the possible development trend of the distortion. If no obvious distortion points are found in the point cloud, or the number of distortion points is small and the development trend is not obvious, the third deviation detection result is judged to be normal. If there are a large number of distortion points in the point cloud and the distortion trend is obvious, which may have a serious impact on the performance and use of the bent pipe, the third deviation detection result is judged to be abnormal.
[0093] Finally, the first, second and third bias detection results are summarized. Taking the three test results into consideration, if any one or more of them are judged to be abnormal, the overall bias detection result is abnormal; if all items are judged to be normal, the overall bias detection result is normal.
[0094] Furthermore, the method further comprises:
[0095] The digital feedback device includes an error feedback unit and a decision feedback unit;
[0096] In combination with the error feedback unit, deviation feedback management of dynamic execution of the strategy is performed, and temporary storage of feedback management information is performed, wherein there is a temporary storage time limit in a predetermined time zone;
[0097] Based on the feedback management information, the intelligent decision module is updated and learned in combination with the decision feedback unit.
[0098] In the embodiment of the present application, the digital feedback device includes an error feedback unit and a decision feedback unit, and these two units play an important role in the dynamic adjustment and optimization of the system.
[0099] First, the error feedback unit is responsible for monitoring the deviation of the dynamic execution of the strategy. When the system performs operations according to the predetermined strategy, if there is an execution deviation, the error feedback unit will capture this information and perform feedback management. Specifically, the error feedback unit collects and records various deviation data during the strategy execution process, such as delays in execution time, deviations in execution results, etc. These data are temporarily stored in the feedback management information for subsequent analysis and processing. The feedback management information has a temporary storage time limit in a predetermined time zone. For a period of time, the system will retain this information for analysis and learning, but after this time limit, the old data will be overwritten or deleted by the new data to ensure the system's storage efficiency and data timeliness.
[0100] Next, the decision feedback unit will update and learn the intelligent decision module based on these feedback management information. Accurately extract key data such as execution deviation data, relevant environmental variables, and operating parameters from the feedback management information library. Perform data cleaning on the extracted key data to remove duplicate, invalid, or abnormal data to ensure the accuracy and validity of the data. Use statistical methods to analyze key indicators such as the distribution, mean, and variance of execution deviations. Extract meaningful features from the original data, such as the size, frequency, and duration of the deviation. Then obtain the correlation between each feature and between the feature and the execution deviation through correlation analysis.
[0101] Select an appropriate machine learning model, such as linear regression, decision tree, neural network, etc., based on the nature of the problem. Use the cleaned data to train the selected machine learning model. Evaluate the performance and generalization ability of the model through cross-validation techniques. Based on the trained model, evaluate the performance of the current strategy in various situations. Identify deficiencies and areas for improvement in the strategy by comparing the model's predicted results with the actual results. Adjust the strategy parameters or logic based on the model's feedback and strategy evaluation results to optimize the decision-making process.
[0102] Finally, the optimized strategy is integrated into the intelligent decision-making module to replace or modify the original decision-making logic.
[0103] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:
[0104] The present application reads straight pipe specification information and bending standards, calibrates valid point clouds, wherein valid point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the valid cross-section position; interactive homologous bending records are used to mine abnormal defect classes, and characteristic degeneration thresholds-mapping parameter control thresholds are determined, and matrix transformation is performed to determine rigid constraints; hierarchical control analysis is performed in combination with valid point clouds and rigid constraints, and an intelligent decision-making module is used to determine a pre-control strategy, wherein the hierarchical standard is the bending device control layer-pump station control layer-circuit control layer; a computer control system responds to the pre-control strategy and performs bending control on a computer hydraulic pipe bender, and an intelligent decision-making module is built into the computer control system; control monitoring is performed simultaneously and control data is transmitted back to a digital feedback device, deviation analysis is performed based on the bending deformation data and the equipment parameter control data, deviation detection results are determined, and control points are positioned with the standard deviation interval as a constraint; the computer control system responds to the digital feedback device and performs bending feedback control. The present invention solves the technical problem that the prior art cannot achieve high-precision and high-efficiency pipe bending operations. By integrating advanced sensing technology, data processing technology and intelligent control algorithms, the present invention realizes precise control and optimization of the pipe bending process, thereby achieving the technical effect of improving product quality and production efficiency.
[0105] Embodiment 2
[0106] Based on the same inventive concept as the intelligent control method of a computer hydraulic pipe bender in the aforementioned embodiment, Figure 2 As shown, the present application provides an intelligent control system for a computer hydraulic pipe bending machine, and the system and method embodiments in the present application are based on the same inventive concept. The system includes:
[0107] An effective point cloud calibration unit 11, which reads straight pipe specification information and bending standards, and calibrates effective point clouds, wherein the effective point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the effective cross-section position;
[0108] A rigid constraint condition determination unit 12, wherein the rigid constraint condition determination unit 12 interacts with homologous bending records, mines abnormal defect classes, determines a characteristic degeneration threshold-mapping parameter control threshold, and performs matrix conversion to determine rigid constraint conditions;
[0109] A pre-control strategy determination unit 13, wherein the pre-control strategy determination unit 13 combines the effective point cloud and the rigid constraint condition with the intelligent decision-making module to perform hierarchical control analysis to determine the pre-control strategy, wherein the hierarchical standard is the bending device control layer - the pump station control layer - the circuit control layer;
[0110] A bending control unit 14, wherein the bending control unit 14 controls the computer control system to respond to the pre-control strategy and perform bending control on the computer hydraulic pipe bender, and the intelligent decision-making module is built into the computer control system;
[0111] A control point positioning unit 15, which performs control monitoring and returns control data to a digital feedback device, performs a deviation analysis based on the bending deformation data and the equipment parameter control data, determines the deviation detection result, and locates the control point with the standard deviation interval as a constraint;
[0112] The bending feedback control unit 16 controls the computer control system to respond to the digital feedback device to perform bending feedback control.
[0113] Furthermore, the system is also used to implement the following functions:
[0114] Establishing a curvature calculator, wherein the curvature calculator is trained based on a Gaussian curvature calculation mechanism and is in communication connection with a computer control system;
[0115] Based on the curvature calculator, traverse the valid point cloud to perform Gaussian curvature calculation to determine a point cloud curvature set;
[0116] In combination with the bending standard, a standard deviation interval between the adjacent point cloud and the same-frequency point cloud based on the spatial position is set, wherein the adjacent point cloud is distributed along the axial direction of the tube, and the same-frequency point cloud is distributed along the tangential direction of the tube;
[0117] Based on the point cloud curvature set and the standard deviation interval, mapping association is performed with the valid point cloud.
[0118] Furthermore, the system is also used to implement the following functions:
[0119] Taking defect type as the first-level clustering standard and visualization feature as the second-level clustering standard, traversing the homologous bending records to perform hierarchical clustering processing, and determining hierarchical clustering clusters;
[0120] Traverse one layer of clusters, perform defect processing, and determine the target defect type;
[0121] Traverse the second-layer clusters, identify the commonality of features and perform lower limit processing as the feature mutation threshold, where the lower limit principle is the minimum eigenvalue;
[0122] The target defect type is mapped to the characteristic variability threshold, and an associated control parameter value based on the characteristic variability threshold is determined as the mapping parameter control threshold, wherein the mapping relationship is one-to-many.
[0123] Furthermore, the system is also used to implement the following functions:
[0124] A three-dimensional matrix is established with defect types as matrix rows, features as matrix columns, and control parameters as matrix verticals;
[0125] The mapping relationship is identified, and a distribution conversion based on the three-dimensional matrix is performed to determine a constraint matrix, and the constraint matrix is used as the rigid constraint condition.
[0126] Furthermore, the system is also used to implement the following functions:
[0127] Identify the relative spatial position between the pipe bending device and the pre-stressed pipe bending, wherein the relative spatial position is marked with a contact partition; perform a force analysis based on the relative spatial position in combination with the rigid constraint condition, perform a stress extension analysis in combination with the effective point cloud, and determine a layer of control information of the pipe bending device;
[0128] Taking the control information of the first layer as the control standard and based on the inter-layer control operation logic, performing hierarchical calculation to determine the pre-control strategy;
[0129] Among them, the stress epitaxy trend is the epitaxy analysis benchmark, and the stress epitaxy trend is obtained based on the homologous bending record.
[0130] Furthermore, the system is also used to implement the following functions:
[0131] Identifying the bending deformation data, and determining the actual curvature distribution based on the effective point cloud;
[0132] Based on the actual curvature distribution and the identified Gaussian curvature, performing an overall trend analysis and an adjacent difference trend analysis to determine a first deviation detection result;
[0133] Determine a second deviation detection result based on the actual control response deviation of the parameters of the pre-control strategy;
[0134] Based on the valid point cloud, single-point distortion identification and distortion trend prediction are performed to determine a third deviation detection result;
[0135] The first deflection detection result, the second deflection detection result, and the third deflection detection result are integrated to determine the deflection detection result.
[0136] Furthermore, the system is also used to implement the following functions:
[0137] The digital feedback device includes an error feedback unit and a decision feedback unit;
[0138] In combination with the error feedback unit, deviation feedback management of dynamic execution of the strategy is performed, and temporary storage of feedback management information is performed, wherein there is a temporary storage time limit in a predetermined time zone;
[0139] Based on the feedback management information, the intelligent decision module is updated and learned in combination with the decision feedback unit.
[0140] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0141] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
[0142] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. An intelligent control method for a computer hydraulic pipe bender, characterized in that: The method comprises: Read straight pipe specification information and bending standards, and calibrate effective point clouds, wherein the effective point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the effective cross-section position; Interactive homologous bending records, mining abnormal defect types, and determining feature degeneration thresholds - mapping parameter control thresholds, performing matrix transformation to determine rigid constraints; Combining the effective point cloud with the rigid constraint condition and the intelligent decision-making module, hierarchical control analysis is performed to determine the pre-control strategy, wherein the hierarchical standard is the pipe bending device control layer - pump station control layer - circuit control layer; The computer control system responds to the pre-control strategy to perform bending control on the computer hydraulic pipe bender, and the intelligent decision-making module is built into the computer control system; Simultaneously carry out control monitoring and send back control data to the digital feedback device, perform deviation analysis based on the bending deformation data and equipment parameter control data, determine the deviation detection results, and use the standard deviation interval as a constraint to locate the control point; The computer control system responds to the digital feedback device to perform bending feedback control; The valid point cloud identifier has Gaussian curvature, including: Establishing a curvature calculator, wherein the curvature calculator is trained based on a Gaussian curvature calculation mechanism and is in communication connection with a computer control system; Based on the curvature calculator, traverse the valid point cloud to perform Gaussian curvature calculation to determine a point cloud curvature set; In combination with the bending standard, a standard deviation interval between the adjacent point cloud and the same-frequency point cloud based on the spatial position is set, wherein the adjacent point cloud is distributed along the axial direction of the tube, and the same-frequency point cloud is distributed along the tangential direction of the tube; Based on the point cloud curvature set and the standard deviation interval, mapping and associating with the valid point cloud; Among them, mining abnormal defect classes and determining feature mutation thresholds - mapping parameter control thresholds include: Taking defect type as the first-level clustering standard and visualization feature as the second-level clustering standard, traversing the homologous bending records to perform hierarchical clustering processing, and determining hierarchical clustering clusters; Traverse one layer of clusters, perform defect processing, and determine the target defect type; Traverse the second-layer clusters, identify the commonality of features and perform lower limit processing as the feature mutation threshold, where the lower limit principle is the minimum eigenvalue; Mapping the target defect type and the characteristic variability threshold, and determining an associated control parameter value based on the characteristic variability threshold as the mapping parameter control threshold, wherein the mapping relationship is one-to-many; The performing matrix transformation to determine the rigid constraint condition includes: A three-dimensional matrix is established with defect types as matrix rows, features as matrix columns, and control parameters as matrix verticals; The mapping relationship is identified, and a distribution conversion based on the three-dimensional matrix is performed to determine a constraint matrix, and the constraint matrix is used as the rigid constraint condition.
2. The method according to claim 1, characterized in that Combined with intelligent decision-making modules, hierarchical control analysis is performed, including: Identify the relative spatial position between the pipe bending device and the pre-stressed pipe bending, wherein the relative spatial position is marked with a contact partition; perform a force analysis based on the relative spatial position in combination with the rigid constraint condition, perform a stress extension analysis in combination with the effective point cloud, and determine a layer of control information of the pipe bending device; Taking the control information of the first layer as the control standard and based on the inter-layer control operation logic, performing hierarchical calculation to determine the pre-control strategy; Among them, the stress epitaxy trend is the epitaxy analysis benchmark, and the stress epitaxy trend is obtained based on the homologous bending record.
3. The method according to claim 1, characterized in that The determining of the bias detection result includes: Identifying the bending deformation data, and determining the actual curvature distribution based on the effective point cloud; Based on the actual curvature distribution and the identified Gaussian curvature, performing an overall trend analysis and an adjacent difference trend analysis to determine a first deviation detection result; Determine a second deviation detection result based on the actual control response deviation of the parameters of the pre-control strategy; Based on the valid point cloud, single-point distortion identification and distortion trend prediction are performed to determine a third deviation detection result; The first deflection detection result, the second deflection detection result, and the third deflection detection result are integrated to determine the deflection detection result.
4. The method according to claim 1, characterized in that The method further comprises: The digital feedback device includes an error feedback unit and a decision feedback unit; In combination with the error feedback unit, deviation feedback management of dynamic execution of the strategy is performed, and temporary storage of feedback management information is performed, wherein there is a temporary storage time limit in a predetermined time zone; Based on the feedback management information, the intelligent decision module is updated and learned in combination with the decision feedback unit.
5. An intelligent control system for a computer hydraulic pipe bender, characterized in that: For executing the method according to any one of claims 1 to 4, the system comprises: An effective point cloud calibration unit, which reads straight pipe specification information and bending standards and calibrates effective point clouds, wherein the effective point clouds are marked with Gaussian curvature, including stress points and key points, distributed at any annular position of the effective cross-section position; A rigid constraint condition determination unit, wherein the rigid constraint condition determination unit interacts with homologous bending records, mines abnormal defect classes, determines a characteristic degeneration threshold-mapping parameter control threshold, and performs matrix conversion to determine rigid constraint conditions; A pre-control strategy determination unit, wherein the pre-control strategy determination unit combines the effective point cloud and the rigid constraint condition with an intelligent decision-making module to perform hierarchical control analysis to determine a pre-control strategy, wherein the hierarchical standard is a pipe bending device control layer - a pump station control layer - a circuit control layer; A bending control unit, wherein the bending control unit controls the computer control system to respond to the pre-control strategy and perform bending control on the computer hydraulic pipe bender, and the intelligent decision-making module is built into the computer control system; A control point positioning unit, which performs control monitoring and returns control data to a digital feedback device simultaneously, performs a deviation analysis based on the bending deformation data and the equipment parameter control data, determines the deviation detection result, and locates the control point with the standard deviation interval as a constraint; A bending feedback control unit controls the computer control system to respond to the digital feedback device to perform bending feedback control.
