An off-line automatic plate number and position calibration system

Abstract

The present application provides a kind of plate piece off-line automatic number of collision and position calibration system, through the innovative architecture of "off-line automatic number of collision + intelligent algorithm calibration + real-time data interaction", effectively solve the pain points such as low efficiency, insufficient precision, poor compatibility in traditional plate processing, provide a set of efficient, high-precision, intelligent position calibration solution for high-end manufacturing field.The system number of collision process is separated from machining center, can simultaneously number of collision to multiple plates, machining center does not need to wait, equipment utilization rate improves;From plate identification, number of collision, parameter calculation to processing execution full-process automation, single plate processing time is shortened to 8-10 minutes, efficiency is improved.And can be based on plate material, size Automatic matching number of collision parameters, based on processing feedback Dynamic optimization servo parameter, reduce the dependence of artificial experience.

Description

Technical Field

[0001] This invention relates to the field of sheet metal processing technology, and in particular to an offline automatic contact counting and position calibration system and method for sheet metal, an electronic device, and a computer-readable storage medium. Background Technology

[0002] In high-end manufacturing fields such as sheet metal and mold making, the positional accuracy of sheet metal processing directly determines the assembly performance and service life of the product. Traditional sheet metal processing relies primarily on two methods for positional calibration: one is manual online calibration, where workers manually operate a calibration instrument on the machining center to collect the sheet metal coordinates and then input them into the CNC system to adjust machining parameters; the other is offline manual measurement, where sheet metal position data is obtained through a coordinate measuring machine and then manually entered into the machining center. Both methods have several unresolved problems:

[0003] 1. Low efficiency: Manual online measurement takes an average of 10-15 minutes per piece, which occupies valuable production time in the machining center and results in equipment utilization rate of only 60%-70%; Offline manual measurement requires transferring the board to the measurement room, and the round trip and measurement time is about 15-20 minutes per piece, which further lengthens the production cycle.

[0004] 2. Insufficient precision: Manual operation is easily affected by experience, fatigue, and environmental interference. The error of the count is usually between ±0.01 and ±0.02 mm, which is difficult to meet the processing requirements of high-precision sheet metal parts (such as titanium alloy sheet metal parts in the aerospace field, which require an accuracy of ±0.003 mm). In addition, manual data recording is prone to typos, which can lead to processing errors.

[0005] 3. Poor compatibility: The data formats of machining centers and data entry equipment from different brands are not uniform, and manual data entry requires format conversion, which increases the probability of errors; moreover, the data cannot be centrally managed, making it difficult to carry out subsequent process optimization and quality traceability.

[0006] 4. Limited capacity: In online batch processing mode, the machining center cannot perform batch processing and machining tasks in parallel. When processing in batches, the equipment waiting time accumulates, making it difficult to increase the overall capacity.

[0007] While some existing offline contact measurement systems have achieved automated data acquisition, they suffer from problems such as weak data preprocessing capabilities, lack of intelligent calibration algorithms, and poor data interaction with machining centers, failing to completely resolve the aforementioned pain points. Therefore, developing a high-precision, high-efficiency, and fully automated offline contact measurement and position calibration system for sheet metal is of significant practical importance. Summary of the Invention

[0008] To address the technical problems existing in the prior art, the present invention provides the following technical solution:

[0009] On the one hand, an offline automatic contact counting and position calibration system for board components is provided, including:

[0010] An offline data acquisition workstation is used to collect three-dimensional coordinate data of multiple feature points on a board and to preprocess the coordinate data.

[0011] A data center server cluster, communicatively connected to the offline coordinate data workstation, is used to receive and store processed coordinate data. It incorporates a weighted least squares position calibration optimization algorithm to calculate the position calibration parameters for the six degrees of freedom of the plate based on the coordinate data and the theoretical design coordinates of the plate.

[0012] The intelligent machining center is connected to the data center server cluster and is used to retrieve the position calibration parameters and adjust the machining coordinate system and tool compensation parameters based on the parameters to perform sheet metal machining.

[0013] Preferably, the offline contact counting workstation includes a high-precision contact counting device, an XYZ motion platform, and an intelligent positioning fixture;

[0014] The high-precision contact-type counter is used to contact feature points on the surface of the plate and generate trigger signals under the drive of the XYZ motion platform.

[0015] The XYZ motion platform is used to drive the contact counter to move along a preset path in three-dimensional space, and the grating ruler on it is used to synchronously collect the actual three-dimensional coordinates of the contact counter tip in the global coordinate system when the contact is triggered.

[0016] The intelligent positioning fixture is used to fix the board to be tested.

[0017] Preferably, the offline data collection workstation further includes:

[0018] The data acquisition and preprocessing module is configured to: receive the raw coordinate data acquired by the grating ruler, and perform median filtering and mean filtering on the raw coordinate data to remove noise signals;

[0019] The board identification module is configured to read board identification information through a barcode reader or RFID reader to associate it with the board's CAD design model and theoretical design coordinates.

[0020] Preferably, the weighted least squares position calibration optimization algorithm is configured as follows:

[0021] Based on the actual 3D coordinates, theoretical design coordinates, and dynamically allocated weight matrix, a weighted objective function is constructed; and

[0022] By iteratively solving for the minimum value of the weighted objective function, the translation parameters (X0, Y0, Z0) and attitude Euler angles (α, β, γ) of the origin of the plate's own coordinate system in the global collision number coordinate system are calculated, thereby realizing the accurate mapping of the plate from the design coordinate system to the machining coordinate system.

[0023] Preferably, the coordinate data is transmitted between the data center server cluster and the offline data synchronization workstation via the Modbus TCP protocol, and the position calibration parameters are transmitted between the data center server cluster and the intelligent machining center via the Profinet protocol.

[0024] Preferably, the intelligent machining center includes:

[0025] The data retrieval and verification module is configured to request and obtain the location calibration parameters from the data center server cluster based on the board identification information.

[0026] The machining parameter pre-adjustment module is configured to calculate the initial working position of the machining center and the tool compensation value based on the position calibration parameters (X0, Y0, Z0, α, β, γ).

[0027] The motion control and machining execution module is configured to control the machining center to perform machining based on pre-tuned parameters.

[0028] Preferably, the intelligent machining center further includes a real-time feedback correction module, the module being configured as follows:

[0029] During the processing, the actual position data of the worktable is collected in real time;

[0030] The actual location data is compared with the theoretical processing trajectory generated based on the location calibration parameters;

[0031] When the deviation exceeds the preset threshold, the PID compensation algorithm is automatically triggered to dynamically adjust the servo motor output in order to correct the motion trajectory.

[0032] Preferably, the system further includes edge computing nodes deployed on the offline data collection workstations and / or in the data center. These edge computing nodes are configured to perform localized real-time preprocessing on the collected coordinate data, including filtering and outlier removal.

[0033] Preferably, the offline data processing workstation, the data center server cluster, and the intelligent processing center are connected via a layered distributed communication architecture built on industrial Ethernet, with each layer exchanging data through encrypted communication.

[0034] On the other hand, a method for offline automatic contact counting and position calibration of board components is provided, applied to the system described above, including the following steps:

[0035] Panel loading and information identification: The panel is fixed on the offline contact measurement workstation, and the identification module is used to obtain the panel identification information to associate it with its CAD model and theoretical coordinates;

[0036] Automatic collision counting and data acquisition: The offline collision counting workstation plans and executes the collision counting path, drives the collision counting device to contact multiple feature points on the surface of the board, synchronously collects the actual three-dimensional coordinate data of each point, and performs filtering preprocessing;

[0037] Calibration parameter calculation: Input the actual three-dimensional coordinate data and its corresponding theoretical design coordinates into the weighted least squares position calibration optimization algorithm to calculate the translation parameters (X0, Y0, Z0) and attitude Euler angles (α, β, γ) of the plate.

[0038] Parameter transmission and machining pre-adjustment: The calculated position calibration parameters are transmitted to the intelligent machining center, which performs machining coordinate system transformation and tool compensation pre-adjustment based on the parameters;

[0039] Machining execution and real-time correction: The machining center executes machining according to the preset parameters and provides real-time feedback on the actual position during the machining process, compares it with the theoretical trajectory, and dynamically corrects the deviation.

[0040] On the other hand, an electronic device is provided, comprising: a processor; and a memory storing computer-readable instructions, which, when executed by the processor, implement the method described above.

[0041] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored therein, the at least one instruction being loaded and executed by a processor to implement the above method.

[0042] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0043] Compared with traditional board position calibration methods, this system has the following significant advantages:

[0044] 1. Efficiency Improvement

[0045] Offline parallel processing: The numbering process is separated from the machining center, and multiple boards can be numbered simultaneously. The machining center does not need to wait, and the equipment utilization rate is increased from 60%-70% to over 85%.

[0046] Automated process: The entire process from board identification, contact counting, parameter calculation to processing execution is automated, reducing the processing time for a single board from the traditional 25-35 minutes to 8-10 minutes, improving efficiency by more than 60%.

[0047] 2 Accuracy guarantee

[0048] High-precision hardware selection: the contact-type counter has a repeatability of ±0.002mm, and the XYZ platform has a positioning accuracy of ±0.001mm, providing a hardware foundation for data acquisition;

[0049] Intelligent algorithm optimization: The weighted least squares method effectively suppresses the influence of noise and outliers, with calibration parameter accuracy ≤ ±0.0005mm and machining position accuracy controlled within ±0.003mm, meeting the needs of high-end manufacturing.

[0050] 3. Data Collaboration and Traceability

[0051] End-to-end data closed loop: From contact data and calibration parameters to processing feedback, all data is uploaded to the data center in real time, enabling traceability of the production process;

[0052] Cross-device compatibility: Supports multiple communication protocols such as Modbus TCP and Profinet, and can interface with machining centers and measuring equipment from different brands, solving the problem of poor compatibility of traditional systems.

[0053] 4. Cost reduction

[0054] Labor cost savings: By reducing manual counting and parameter entry, a single production line can reduce the number of operators by 2-3, resulting in annual labor cost savings of approximately 300,000 to 500,000 yuan;

[0055] Reduced material waste: Improved processing precision reduces the defect rate from the traditional 3%-5% to below 0.5%, resulting in annual material cost savings.

[0056] 5. Intelligent Upgrade

[0057] Edge computing enables localized data preprocessing to reduce the pressure on core servers and achieve real-time response (processing latency ≤5ms).

[0058] Adaptive parameter adjustment: Automatically matches collision count parameters based on sheet material and size, and dynamically optimizes servo parameters based on machining feedback, reducing reliance on manual experience. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a schematic diagram of a hardware system composition structure provided in an embodiment of the present invention;

[0061] Figure 2This is a schematic diagram of the software system architecture provided in an embodiment of the present invention;

[0062] Figure 3 This is a schematic diagram of the implementation process of the method provided in the embodiments of the present invention;

[0063] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0064] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0065] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0066] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0067] In this embodiment of the invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0068] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0069] I. Technical Objectives

[0070] This invention effectively solves the pain points of low efficiency, insufficient precision, and poor compatibility in traditional sheet metal processing through an innovative architecture of "offline automated contact measurement + intelligent algorithm calibration + real-time data interaction," providing a set of efficient, high-precision, and intelligent position calibration solutions for the high-end manufacturing field.

[0071] The following section will detail the application implementation plan of this system.

[0072] II. System Overall Architecture

[0073] 2.1 Hardware System

[0074] like Figure 1 As shown, the hardware architecture of this system consists of four parts: an offline data processing workstation, a data center server cluster, an intelligent processing center, and supporting auxiliary hardware. The functions of each component (described with reference to the example selection) are as follows:

[0075] 2.1.1 Offline Data Matching Workstation

[0076] The offline data acquisition workstation is the core hardware unit for realizing board coordinate acquisition and position calibration, and mainly includes the following components:

[0077] High-precision contact counter: Uses Renishaw SP25M contact scanning probe with repeatability of ±0.002mm and adjustable trigger force range of 0.1-1.0N, suitable for plates of different materials and thicknesses (e.g., 0.2N trigger force for sheet metal and 0.5N trigger force for titanium alloy plates); the probe head is made of ruby ​​material, which has high hardness, low wear, and a service life of up to 1 million triggers.

[0078] High-precision XYZ motion platform: Utilizing a Bosch Rexroth linear motor to drive the XYZ platform, with a positioning accuracy of ±0.001mm, repeatability of ±0.0005mm, maximum speed of 500mm / s, and a travel range of 1500×1200×500mm, it can accommodate plates with a maximum size of 1400×1100mm. The platform is equipped with a linear encoder position feedback system with a resolution of 0.1μm, providing real-time feedback of the motion position to ensure the accuracy of the acquired coordinates. Existing three-axis motion platforms can be used.

[0079] Intelligent positioning fixture and pneumatic locking system: The fixture adopts a modular design and is equipped with a quick-change positioning reference block, which can adapt to plates of different shapes (rectangular, circular, irregular shapes); the pneumatic locking system uses SMC pneumatic components, with a locking pressure of 0.5MPa. The locking force is evenly distributed on the edge of the plate, ensuring that the plate does not shift or deform during the collision process, with a pressure error ≤±0.05MPa. Existing positioning fixtures and locking equipment can be selected.

[0080] Industrial PC: An Advantech IPC-610L industrial PC is selected, equipped with an Intel Core i7-12700 processor, 32GB DDR4 memory, a 1TB SSD, and running Windows 10 IoT Enterprise operating system. It features multi-threaded data acquisition and processing capabilities and supports 24-hour continuous and stable operation. Alternatively, a PLC controller can be used.

[0081] Data acquisition and preprocessing module: Utilizes NI 9215 data acquisition card with a sampling rate of 1000Hz, capable of simultaneously acquiring the trigger signal of the touch sensor, the position signal of the XYZ platform, and the pneumatic locking pressure signal; the module incorporates median filtering and mean filtering algorithms to remove noise signals caused by environmental vibration and electromagnetic interference in real time, with a signal processing delay of ≤1ms.

[0082] Panel identification module: Honeywell HF800 barcode reader and RFID reader are used. It supports reading QR codes on the surface of the panel and embedded RFID tags. The identification speed is ≤0.1s and the identification rate is 100%. It can quickly associate the CAD design data, processing parameters and other information of the panel.

[0083] 2.1.2 Data Center Server Cluster

[0084] Data centers are the core nodes for data storage, computing, and interaction, and they employ a hybrid architecture of local servers and cloud servers.

[0085] Local core server: A Dell PowerEdge R750 server is selected, configured with 2 Intel Xeon Gold 6330 processors (28 cores and 56 threads), 128GB DDR4 memory, 8×4TB SAS solid-state drives, running Linux CentOS 7 operating system, and deploying MySQL database and Redis caching system to store core data such as raw collision data, calibration parameters, and processing parameters, with data read and write speed ≥1000MB / s.

[0086] Edge computing nodes: Deploy 5 Huawei Atlas 500 Pro edge servers, each equipped with an Ascend 310 AI chip (16 TOPS computing power), for local real-time data preprocessing (such as coordinate data noise reduction and outlier removal), with a processing latency of ≤5ms, reducing the computing pressure on the core server and enabling hierarchical data processing.

[0087] Cloud server: Alibaba Cloud ECS g7 instance is selected, configured with 8 cores, 16GB memory, and 500GB cloud disk, for backing up local server data, remote access and monitoring, to achieve off-site disaster recovery of data, with a disaster recovery time of ≤30 minutes.

[0088] Data backup system: It adopts a combination of tape backup and cloud backup. Local tape backup is performed once a day, and cloud backup is performed once an hour. The data redundancy is ≥3 to ensure that no data is lost.

[0089] 2.1.3 Intelligent Machining Center

[0090] The machining center is the terminal unit that performs sheet metal processing tasks. The Shenyang Machine Tool i5M8.5 intelligent machining center is selected, and its main configuration is as follows:

[0091] CNC Control System: Equipped with SIEMENS 840D sl control system, supporting multi-axis linkage control, positioning accuracy ±0.001mm, repeatability ±0.0005mm, supporting industrial communication protocols such as Modbus TCP and Profinet, and enabling real-time data interaction with data centers.

[0092] Servo drive system: SIEMENS S120 servo drive is selected, equipped with an absolute encoder with a resolution of 13 bits, providing real-time feedback on motor speed and position, with a response time of ≤0.5ms, ensuring the accuracy of the motion trajectory.

[0093] Position feedback and compensation module: Equipped with a grating ruler and laser interferometer, it collects worktable position data in real time with a resolution of 0.1μm, and can realize tool length compensation, radius compensation, and thermal deformation compensation, further improving machining accuracy.

[0094] Pneumatic locking and feeding auxiliary system: The locking system parameters are consistent with those of the offline contact workstation, ensuring the positioning accuracy and stability of the workpiece during processing.

[0095] 2.1.4 Supporting auxiliary hardware

[0096] Industrial Ethernet switch: Huawei S5735S-L24T4S-A switch is selected, with 24 electrical ports + 4 optical ports, 1000Mbps bandwidth, and support for VLAN segmentation and QoS priority settings to ensure the stability and real-time performance of data transmission.

[0097] Industrial-grade UPS power supply: SANTAK C10KS UPS is selected, with a rated power of 10kVA and a backup time of 2 hours, to ensure that the equipment can be shut down normally and data is not lost in the event of a sudden power outage.

[0098] The following is an example of the data communication and control architecture of the hardware system:

[0099] This system adopts a hierarchical communication architecture and a distributed control mode to realize real-time data interaction and collaborative work between hardware systems. The specific design is as follows:

[0100] Communication connection method:

[0101] Backbone network: Adopts industrial Ethernet architecture, constructs star topology through Huawei S5735S switches, and each hardware unit is connected through Category 5e shielded twisted pair cable, supports 1000Mbps full-duplex communication, and end-to-end transmission latency ≤2ms;

[0102] Protocol Standards: The offline data processing workstation and the data center use the Modbus TCP protocol (port 502), with a data frame format of MBAP+PDU, supporting coil read / write and register operations; the intelligent processing center communicates with the data center through Profinet IO (IRT mode), with a periodic communication rate of 125μs and an aperiodic data transmission bandwidth of ≥100Mbps;

[0103] Wireless backup: Configure a Huawei 5G industrial router (model AR550) to support NSA networking mode as a redundant backup for the wired network. The communication latency is ≤20ms to ensure uninterrupted transmission of critical data.

[0104] Control method:

[0105] Centralized-distributed control: The data center serves as the control core, using an OPC UA server (Kepware EX6.5) to monitor device status and distribute parameters; edge computing nodes deploy the PLCopen motion control algorithm to locally process motion commands on the XYZ platform, with a control cycle of ≤1ms;

[0106] Data interaction process:

[0107] 1. Offline data matching stage: The industrial control computer collects coordinate data through the NI-DAQmx driver, and after preprocessing by the edge node (filtering, outlier removal), it is uploaded to the data center via the MQTT protocol (topic: / calibration / data);

[0108] 2. Parameter distribution phase: The data center pushes calibration parameters (X0, Y0, Z0, α, β, γ) to the machining center CNC system via a RESTful API, encapsulated in JSON format, and the data verification uses the CRC32 algorithm;

[0109] 3. Machining Feedback Stage: The machining center uploads spindle load (sampling rate 100Hz) and tool position (resolution 0.1μm) to the data center in real time, and establishes a two-way communication channel through WebSocket to realize dynamic adjustment of the machining process.

[0110] Network security mechanisms: Deploy an industrial firewall (Hillstone Networks SG-6000) to achieve area isolation, and use 802.1X authentication and MAC address binding to restrict access devices; data transmission uses the AES-256 encryption algorithm, and critical commands must be verified by digital signature (RSA-2048) to prevent data tampering and unauthorized access.

[0111] 2.2 Software System Module Design

[0112] like Figure 2As shown, the software architecture of this system adopts a layered design, divided into three layers: device layer software, data layer software, and application layer software. The specific functions of each layer's modules are as follows:

[0113] 2.2.1 Device Layer Software

[0114] Offline data matching workstation software: Developed using C#, it runs on an industrial control computer and mainly includes the following modules:

[0115] 1. Collision Count Path Planning Module: Import the CAD model of the board (supports STEP, IGES, and DWG formats), and automatically generate the optimal collision count path based on the shape, size, and processing requirements of the board. For rectangular boards, 9 sampling points are collected by default: 4 corner points, 4 midpoints of the sides, and 1 center point. For irregular boards, feature points (holes, slots, edges) are automatically identified and sampling paths are generated. The path planning time is ≤10s. Manual adjustment of the number and position of sampling points is supported.

[0116] 2. Data Acquisition and Preprocessing Module: Real-time acquisition of the trigger signal of the touch sensor and the position signal of the XYZ platform. The acquired coordinate data is processed by median filtering (window size 5) and mean filtering (sampling times 3) to remove noise signals. The data processing delay is ≤1ms. It supports viewing the comparison curve between the original data and the processed data.

[0117] 3. Panel Information Management Module: Associates panel information such as ID, material, size, thickness, processing parameters, and production batch, and automatically matches corresponding parameters such as collision trigger force and platform movement speed, without requiring manual settings; supports batch import of panel information, with an import speed of ≥100 records / minute.

[0118] 4. Communication and Data Upload Module: Utilizes Modbus TCP protocol to communicate with the data center, packaging and uploading raw collision data, calibrated position parameters, board information, etc. The data packets are encrypted using AES-256 to ensure data transmission security; upload success rate ≥99.99%, upload speed ≥10MB / s.

[0119] CNC machining center software: developed based on the SIEMENS 840D sl open platform, mainly including the following modules:

[0120] 1. Data retrieval and verification module: Initiates board data requests to the data center via the Profinet protocol, verifies the matching of board ID with processing task, and retrieves data such as collision calibration parameters and processing parameters; supports local caching function, if the data center network is interrupted, cached data can be called to continue processing, and the cache validity period is 24 hours.

[0121] 2. Machining parameter pre-adjustment module: Based on the calibrated plate position parameters, automatically calculate the initial position of the worktable, the tool compensation parameters (length compensation, radius compensation), and the servo drive gain parameters; pre-adjustment time ≤ 2 minutes, pre-adjustment error ≤ ±0.002mm.

[0122] 3. Motion control and machining execution module: Controls servo motors according to pre-set motion trajectories and process parameters to achieve multi-axis linkage machining; collects data such as cutting force, tool wear, and position feedback in real time, with a data sampling rate of 100Hz.

[0123] 4. Real-time feedback correction module: The actual position data of the worktable is collected in real time by the grating ruler and laser interferometer (sampling frequency 1000Hz) and compared with the theoretical machining trajectory. When the deviation exceeds ±0.003mm, the PID compensation algorithm is automatically triggered to dynamically adjust the output torque and movement speed of the servo motor. The compensation response time is ≤10ms, ensuring the stability of position accuracy during the machining process.

[0124] 4. Real-time feedback correction module: Compares the real-time acquired position data with the calibration parameters. If the error exceeds ±0.003mm, it automatically adjusts the servo drive parameters and corrects the motion trajectory. The correction response time is ≤10ms to ensure machining accuracy.

[0125] 2.2.2 Data Layer Software

[0126] Developed using the Java language and deployed on local core servers and cloud servers, it mainly includes the following modules:

[0127] Data storage and management module: It uses a MySQL database to store collision data, calibration parameters, processing data, board information, etc., and uses Redis to cache frequently used data (such as process parameters of high-frequency processed boards), with a cache hit rate of ≥95%; it supports functions such as adding, deleting, modifying and querying data, batch exporting, and historical querying, with a query response time of ≤1s.

[0128] Algorithm calculation module: Built-in weighted least squares board position calibration algorithm, data filtering algorithm, and process parameter optimization algorithm, which can automatically process the raw data uploaded by the offline data matching workstation and generate calibrated position parameters; the algorithm calculation time is ≤5s and the calculation accuracy is ≤±0.001mm.

[0129] Data interaction interface module: Provides multiple communication interfaces such as Modbus TCP, Profinet, and HTTP, and supports integration with third-party systems such as offline data processing workstations, machining centers, production planning systems, and quality inspection systems; the interface adopts the RESTful design specification, supports high-concurrency requests, and has a concurrent processing capacity of ≥1000 times / second.

[0130] Status monitoring and early warning module: Real-time monitoring of the operating status of offline data matching workstations, processing centers, and data centers (such as equipment temperature, CPU utilization, memory utilization, and network bandwidth). When equipment malfunctions (such as data matching device failure, network interruption, or CPU utilization exceeding 90%), early warnings are issued to administrators via SMS, email, client pop-ups, etc., with an early warning response time of ≤1 second. Supports the generation of daily, weekly, and monthly equipment operation reports.

[0131] 2.2.3 Application Layer Software

[0132] Developed based on web technology, it supports access from both PC and mobile devices, and its main functions include:

[0133] System configuration management: Set parameters such as trigger force of the touch sensor, movement speed of the XYZ platform, data upload frequency, and warning threshold; supports batch configuration of multiple devices.

[0134] Production progress monitoring: View the contact status, processing status, and completion rate of each board to track production progress in real time;

[0135] Data statistics and analysis: Statistics on collision time, processing time, accuracy error, defect rate, etc., are compiled to generate visual reports such as bar charts and line charts to provide data support for process optimization;

[0136] User permission management: Set permissions for different roles such as administrators, operators, and technicians to ensure data security and operational standards.

[0137] The software's control logic and interaction flow with the hardware are as follows:

[0138] The software system achieves precise control of hardware devices through a layered collaborative mechanism. The interaction process between the software and hardware systems at each level is as follows:

[0139] Layered control architecture:

[0140] 1. Application Layer-Data Layer Interaction: The application layer sends device control commands (such as starting a data processing task and adjusting processing parameters) to the data layer via a RESTful API. The data layer persists the commands through a MySQL database and caches real-time status data through Redis. The data layer pushes the hardware status (such as the position of the XYZ platform and processing progress) to the application layer in JSON format, with an update frequency of ≤1 second.

[0141] 2. Data Layer-Device Layer Interaction: The data layer sends path planning parameters (sampling point coordinates, movement speed) to the offline data processing workstation via the Modbus TCP protocol, and pushes calibration parameters (X0, Y0, Z0, α, β, γ) to the machining center CNC system via Profinet IO; the device layer feeds back the real-time running status (such as current sampling point, trigger force value, and machining progress) to the data layer via the MQTT protocol (QoS=1).

[0142] 3. Device Layer - Hardware Execution: The offline data acquisition workstation software uses the NI-DAQmx driver to control the data acquisition card and outputs analog signals (0-10V) to control the XYZ platform servo motors; the machining center CNC software uses the PLCopen motion control function library to generate pulse commands to drive the servo driver to achieve tool trajectory control, with a position loop control cycle ≤1ms.

[0143] The core control interaction flow of the software system is as follows:

[0144] 1. Offline number matching stage:

[0145] The application layer issues a collision detection task (board ID, sampling strategy) → the data layer calls the algorithm to calculate path parameters → the device layer software parses the path and generates G-code → the XYZ platform moves via the motion control card → coordinate data is collected when the collision detection device is triggered → the data is preprocessed, encrypted, and uploaded to the data center.

[0146] Hardware status feedback: The platform's movement position is transmitted back in real time via a grating ruler (resolution 0.1μm), the trigger force signal is acquired via an analog input module (sampling rate 1kHz), and abnormal states (such as collisions or overtravel) are triggered by digital I / O to initiate an emergency stop.

[0147] 2. Processing execution stage:

[0148] The CNC software of the machining center retrieves calibration parameters from the data center → configures servo driver parameters (position loop gain, acceleration feedforward) through Profinet IO → collects grating ruler position feedback in real time when executing G code → and dynamically compensates through PID algorithm after comparing with the theoretical trajectory (compensation cycle 125μs).

[0149] Data closed-loop process: Machining process data (cutting force, spindle temperature) is uploaded to the data layer through the OPC UA server, and edge computing nodes perform real-time analysis. When tool wear is detected (vibration amplitude > 0.05mm), a tool replacement warning is automatically triggered.

[0150] 3. Exception handling mechanism:

[0151] Hardware fault response: When the loss of the trigger signal is detected (no trigger for 200ms), the device layer software immediately sends an emergency stop command (DO output low level), and reports to the data layer via Modbus TCP. The application layer pops up a window to display the fault code and troubleshooting suggestions.

[0152] Network interruption response: A local caching + breakpoint resume mechanism is adopted. After the network is restored, the device layer automatically uploads the cached data within 30 minutes, and the data layer ensures data integrity by comparing timestamps.

[0153] Control timing guarantees:

[0154] Nanosecond-level clock synchronization is achieved between different levels through a time synchronization protocol (PTP IEEE 1588), ensuring that the timestamp deviation between command issuance and data acquisition is ≤10μs; critical control commands (such as emergency stop and parameter issuance) are processed using a priority queue to ensure that the response delay is ≤5ms, meeting the requirements of real-time control.

[0155] III. Introduction to the Core Algorithm: Weighted Least Squares Method for Optimizing Board Position Calibration

[0156] 3.1 Algorithm Principles and Formula Definitions

[0157] The position and attitude of the plate can be described by 6 degrees of freedom parameters: X0 (X-axis coordinate of the origin of the plate's own coordinate system in the global coordinate system), Y0 (Y-axis coordinate), Z0 (Z-axis coordinate), α (rotation angle around the X-axis, roll angle), β (rotation angle around the Y-axis, pitch angle), and γ (rotation angle around the Z-axis, yaw angle).

[0158] Assuming n contact points are collected, the actual collected coordinates of each point are ( , , (i=1,2,...,n), the theoretical design coordinates of this point in the plate's own coordinate system are ( , , Then the three-dimensional spatial coordinate transformation relationship is:

[0159] ,

[0160] ,

[0161] ,

[0162] The symbols are defined as follows:

[0163] (1). Coordinate parameters

[0164] ( , , ): The actual three-dimensional coordinates of the i-th collision point in the global collision coordinate system, in mm. It is acquired in real time by the XYZ motion platform grating ruler of the offline collision workstation with a sampling resolution of 0.1μm. The data is obtained after median filtering (window size 5) and mean filtering (sampling times 3).

[0165] ( , , ): The theoretical three-dimensional coordinates of the i-th contact point in the design coordinate system of the plate itself, in mm. It is extracted from the CAD model of the plate and obtained by the plate recognition module through reading the QR code / RFID tag. The coordinate accuracy is consistent with the CAD design accuracy (≤±0.001mm).

[0166] (2). Translation parameters

[0167] X0, Y0, Z0: Translation parameters of the origin of the plate's own coordinate system in the global coordinate system, in mm, representing the positional offset of the plate in space, obtained by weighted least squares algorithm, with a calculation accuracy ≤ ±0.0005 mm.

[0168] (3). Attitude parameters (Euler angles)

[0169] α: The rotation angle (roll angle) of the plate about the X-axis, in radians (rad), with a value range of [-π / 6, π / 6]. It represents the degree of tilt of the plate in the X-axis direction and is calculated by the algorithm based on the fitting of multiple sets of collision point coordinates.

[0170] β: The rotation angle (pitch angle) of the plate about the Y-axis, in radians (rad), with a value range of [-π / 6, π / 6]. It represents the degree of tilt of the plate in the Y-axis direction and is calculated in the same way as α.

[0171] γ: The rotation angle (yaw angle) of the plate about the Z-axis, in radians (rad), with a value range of [-π / 3, π / 3]. It represents the rotational offset of the plate in the plane and is solved by the coordinate transformation matrix.

[0172] cosθ and sinθ are trigonometric function operators, where θ represents α, β, or γ. They are used to convert angular parameters into coordinate transformation coefficients. Double-precision floating-point numbers (64-bit) are used to ensure the accuracy of the calculation.

[0173] Operating Mechanism: The above parameters together constitute the core variables for plate position calibration. The translation parameters (X0, Y0, Z0) address the "position offset" problem during plate placement, while the Euler angles address the "attitude tilt" problem. Together, they achieve a precise mapping of the plate from the design coordinate system to the machining coordinate system, providing a data foundation for subsequent toolpath planning in the machining center. In practical applications, the system collects data from at least five feature points (…). , , The data was substituted into the formula to construct an overdetermined system of equations, and finally the optimal values ​​of the six parameters were obtained.

[0174] The formula describes the three-dimensional coordinate transformation process from the plate's own coordinate system to the global collision number coordinate system. It achieves the mapping from theoretical design coordinates to actual measured coordinates by integrating translation and rotation transformations. The specific operating mechanism is as follows:

[0175] Coordinate transformation logic: the left side of the formula ( , , The numbers represent the actual 3D coordinates collected during the touchdown process. The right side constructs a mapping relationship in the form of "translation parameters + rotation matrix × theoretical coordinates".

[0176] Translation components: (X0, Y0, Z0) represent the offset of the origin of the plate in the X / Y / Z axis of the global coordinate system, respectively, which are used to eliminate the translation error of the plate placement position;

[0177] Rotation components: A rotation matrix is ​​constructed using three Euler angles (α roll, β pitch, γ yaw), where trigonometric function terms such as cosθ and sinθ constitute the elements of the rotation matrix, achieving rotational transformations in three-dimensional space; , , ) represents the theoretical design coordinates in the CAD model of the plate, and the attitude adjustment is achieved through matrix multiplication.

[0178] This formula solves two core problems in sheet metal processing:

[0179] Attitude calculation: The tilt state of the plate in space is accurately described by three Euler angles, which solves the problem that angle error cannot be quantified by manual calibration;

[0180] Coordinate unification: Transform the coordinates of scattered contact points to the same global coordinate system, providing a mathematical basis for subsequent weighted least squares parameter optimization;

[0181] In practical applications, the system collects data from at least 5 feature points. , , Substitute these parameters into the formula to construct an overdetermined system of equations, and finally solve for the optimal six parameters X0, Y0, Z0, α, β, and γ to achieve precise calibration of the plate position.

[0182] 3.2 Algorithm Operation Mechanism

[0183] The weighted least squares method constructs an objective function to solve for the six degrees of freedom parameters (X0, Y0, Z0, α, β, γ) that minimize the deviation between the actual acquired coordinates and the theoretical coordinates. The specific steps are as follows:

[0184] (1) Error model construction

[0185] Define the coordinate residual vector of the i-th touch point. for:

[0186] ,

[0187] in( , , ) is the coordinate transformation function in the three-dimensional spatial coordinate transformation formula.

[0188] (2) Definition of weighted objective function

[0189] Considering the differences in measurement reliability at different touchpoints (e.g., edge points have higher accuracy than center points), a weight matrix is ​​introduced. = The weights are dynamically allocated based on the sampling point type. By dynamically allocating the weights of different sampling points using a diagonal matrix (e.g., corner points w=1.5, edge points w=1.2, center points w=1.0), higher weights are assigned to high-reliability points (such as corner points), reducing the interference of noise points on parameter solving and improving calibration accuracy.

[0190] The objective function is:

[0191] T is the transpose. The deviation between theoretical and measured coordinates is quantified in the form of a weighted sum of squares, and parameter optimization is achieved by minimizing this function. The weight matrix W ensures that the error of key feature points (such as corner points) accounts for a higher proportion in the objective function, guiding the algorithm to prioritize fitting high-precision sampling points.

[0192] (3) Parameter iterative solution

[0193] The minimum value of the objective function is found iteratively using the Gauss-Newton method:

[0194] Initial value settings: Preset initial parameters (X0,Y0,Z0=0), (α,β,γ=0) based on the CAD model of the sheet metal;

[0195] Jacobian matrix calculation: Taking the partial derivative with respect to the three-dimensional coordinate transformation formula yields the sensitivity matrix of the residuals with respect to the six parameters. (6×3D);

[0196] Incremental equation construction:

[0197] ,in The parameter correction amount is used, and the operating mechanism is as follows: the nonlinear objective function is linearized based on the Gauss-Newton method, and the left side... To ensure the uniqueness of the solution, a positive definite matrix is ​​used. The right-hand side JᵀWr represents the gradient term, and the parameter correction is obtained by the least squares method. This guides the parameters to be updated in the direction of decreasing residuals;

[0198] Iterative updates: ( The step size factor (initial value 1.0, which is halved if the residual increases) is introduced. The operating mechanism is as follows: The step size factor λ (initial value 1.0) is introduced to dynamically adjust the parameter update amplitude. If the residual increases, λ is halved to avoid divergence. When the residual decreases, a large step size is maintained to accelerate convergence and achieve efficient search of the parameter space.

[0199] Convergence criterion: When Alternatively, the iteration count can be stopped at 50, and the optimal parameters can be output. The parameter update amount is measured by the L2 norm, and the parameters are considered to have converged when the change is less than 1μm. Setting an upper limit of 50 iterations can prevent getting trapped in local optima and balance computational efficiency with solution accuracy (in actual applications, the average number of convergences is about 15-20).

[0200] (4) Algorithm performance metrics

[0201] Calculation accuracy: Calibration parameter error ≤ ±0.0005mm (angle error ≤ ±0.001°);

[0202] Stability: In 100 consecutive calibration experiments, the standard deviation of the parameter is ≤0.0003mm;

[0203] Efficiency: Calibration time for a single board (9 sampling points) ≤ 300ms, supports parallel computing (5 edge servers processing concurrently, throughput ≥ 20 boards / minute).

[0204] The above-mentioned collaborative mechanism of "weighted error quantization → sensitivity analysis → linearization solution → dynamic iteration → convergence control" enables high-precision calibration of the six degrees of freedom parameters of the plate, with a calculation accuracy of ≤ ±0.0005mm, meeting the needs of high-end manufacturing scenarios such as aerospace.

[0205] IV. Data Communication and Interactive Control of Hardware and Software Systems

[0206] 4.1 Communication Architecture Design

[0207] The system adopts a "layered distributed" communication architecture, with real-time data interaction between each layer via industrial Ethernet. The specific communication links are as follows:

[0208] Communication Node Communication Protocol Data types Transmission rate Delay Security Mechanism Offline data collection workstation → Edge computing node Modbus TCP Original coordinate data, board ID, status code 100Mbps ≤5ms AES-256 encryption, CRC check Edge computing node → Data center HTTP / JSON Calibration parameters, machining process parameters 1Gbps ≤20ms Digital certificates, data signatures Data Center → Processing Center Profinet Preset parameters, motion commands 1000Mbps ≤1ms Real-time message priority (Class 3) Processing Center → Data Center MQTT Processing status, real-time position feedback 500Mbps ≤10ms Subscription / Publish Access Control

[0209] 4.2 Interactive Control Flow

[0210] like Figure 3 As shown, taking the entire process of "board contact count → data upload → processing execution" as an example, the detailed interaction process is as follows:

[0211] Step 1: Panel loading and identification

[0212] The operator places the workpiece on the positioning fixture of the offline contact counting workstation, and the pneumatic locking system automatically starts (pressure 0.5MPa), triggering the position sensor to confirm that the workpiece is in place;

[0213] The barcode reader / RFID reader reads the board ID (such as "PN-20231015-001") and uploads it to the industrial control computer;

[0214] The industrial control computer requests the CAD model and process parameters of the board from the data center via an HTTP interface. The data center returns a STEP format model file (size ≤ 50MB) and collision point planning parameters (such as 9-point sampling for rectangular boards).

[0215] Step 2: Automatic number matching and data preprocessing

[0216] The path planning module parses the CAD model, generates motion trajectories (such as collecting data clockwise from the top left corner), and sends instructions to the XYZ motion platform.

[0217] The platform moves along the planned path, and a signal is triggered when the SP25M probe contacts the surface of the board. The grating ruler synchronously collects coordinate data. , , The sampling frequency is 1000Hz.

[0218] The data acquisition card filters the raw data:

[0219] Median filtering: Removes impulse noise (such as outliers caused by slight probe bouncing), window size 5;

[0220] Mean filtering: Smooths out mechanical vibration interference by taking the average value after three consecutive samplings;

[0221] The edge computing node (Atlas 500 Pro) receives the filtered data, calls the outlier detection algorithm (3σ criterion) to remove points with a deviation > 0.01 mm, and retains valid sampling points (n≥5).

[0222] Step 3: Calculate and upload calibration parameters

[0223] Edge computing nodes will effectively sample the coordinates of the points ( , , ) and theoretical coordinates ( , , The input is the weighted least squares module, which solves for 6 degrees of freedom parameters.

[0224] The calculation results are packaged in JSON format (including board ID, calibration parameters, confidence level, and timestamp) and uploaded to the data center via HTTPS protocol;

[0225] After verifying the data integrity (MD5 checksum), the data is stored in the MySQL database, and the board status is updated to "pending processing".

[0226] Step 4: Data retrieval and parameter pre-tuning of the machining center

[0227] The machining center operator scans the part ID, and the CNC system sends a request to the data center via the Profinet protocol. An example code is shown below:

[0228] {

[0229] "board_id": "PN-20231015-001",

[0230] "request_type": "calibration_params",

[0231] "timestamp": "2023-10-15T08:30:15Z"

[0232] }

[0233] The data center returns calibration parameters and machining process parameters (such as spindle speed 8000 rpm and feed rate 500 mm / min).

[0234] The processing parameter pre-adjustment module calculates based on the calibration parameters:

[0235] Initial position of the workbench (safe distance / mm): Xstar = X0 + 100, Y_start = Y0 + 100;

[0236] Tool compensation value: The tool length compensation amount is calculated based on angles α and β. ;

[0237] Servo gain: Dynamically adjust the position loop proportional gain (Kp=8.5) and integral time (Ti=0.1s).

[0238] Step 5: Processing Execution and Real-time Correction

[0239] The machining center starts processing according to the preset parameters, and the laser interferometer collects the position of the worktable in real time (sampling frequency 1000Hz).

[0240] The real-time feedback correction module compares the actual position with the theoretical trajectory. When the deviation e (e = |actual position theoretical trajectory|) > 0.003mm, PID compensation is triggered.

[0241] ,

[0242] in( =12), ( =0.5), ( =0.2), the output control quantity u(t) is used to adjust the torque of the servo motor;

[0243] After processing is completed, the CNC system uploads data such as processing time and accuracy error to the data center and updates the board status to "completed".

[0244] V. Detailed Explanation of Methods and Steps

[0245] 5.1 Offline automatic number matching steps (corresponding to steps 1, 2, and 3 of the system workflow)

[0246] Implementation process:

[0247] Panel positioning and clamping: The operator places the panel in the modular fixture and triggers the pneumatic locking button. The system automatically detects whether the panel outline matches the fixture (achieved through edge sensors). If they do not match, an audible and visual alarm is issued.

[0248] Panel information association: The barcode reader scans the QR code on the panel surface to obtain information such as panel ID, material (e.g., 6061 aluminum alloy), and thickness (5mm), and automatically matches the collision parameters (trigger force 0.3N, platform speed 200mm / s).

[0249] Path planning and execution: The collision number path planning module imports the CAD model, identifies the feature points of the board (such as the coordinates of the four corner points being (0,0,0), (1000,0,0), (1000,800,0), and (0,800,0) respectively), and generates motion path G-code, as shown in the example below:

[0250] G00 X0 Y0 Z50 (Quickly move to 50mm above the starting point);

[0251] G01 Z-5 F200 (Slowly descend to detection altitude);

[0252] G01 X0 Y0 F100 (Contact detection of the first corner point);

[0253] ... (Complete the detection of 9 points in sequence).

[0254] Data acquisition and filtering: When each sampling point is triggered, the grating ruler records the coordinate data. After median filtering (removing burrs) and mean filtering (smoothing fluctuations), the effective coordinates are obtained (such as the actual coordinates of the first corner point (0.002, -0.001, 0.003) mm).

[0255] Data upload to edge node: The industrial control computer packages the coordinate data of 9 points, board ID, and acquisition time, and sends them to the edge computing node (Huawei Atlas 500) via Modbus TCP protocol.

[0256] Technical principle:

[0257] Contact detection achieves high-precision acquisition of coordinate data through synchronous acquisition of probe trigger signals and grating ruler position feedback;

[0258] Filtering algorithms remove noise based on statistical characteristics. Median filtering can effectively suppress impulse interference, while mean filtering can reduce random errors.

[0259] Technical effects:

[0260] The time for counting a single board is ≤3 minutes (compared to 10-15 minutes for traditional manual counting), improving efficiency by 70%.

[0261] Coordinate acquisition accuracy is ±0.002mm (compared to ±0.01-±0.02mm for traditional manual methods), representing an 80% improvement in accuracy.

[0262] 5.2 Processing parameter pre-adjustment steps (corresponding to step 4 of the system workflow)

[0263] Implementation process:

[0264] Data Request and Verification: The CNC system of the machining center sends a parameter request to the data center through the board ID. The data center verifies whether the board has completed the numbering (status is "pending processing"). If not, it returns the error code "E001".

[0265] Calibration parameter reception: The data center returns the 6 degrees of freedom parameters calculated by weighted least squares (e.g., X0=10.002mm, Y0=5.001mm, α=0.001rad, β=-0.0005rad, γ=0.002rad).

[0266] Machining coordinate system transformation: Convert the workpiece's own coordinate system to the machining center coordinate system, and calculate the offset of the worktable origin. :

[0267] ,in , , () represents the coordinates of the design origin of the sheet metal in its own coordinate system;

[0268] Tool compensation calculation: Calculate the tool length compensation value based on angles α and β to ensure that the actual cutting point of the tool is consistent with the theoretical trajectory;

[0269] Servo parameter optimization: Adjust the position loop gain (Kp=9.0) and speed loop integral time (Ti=0.08s) of the servo drive according to the board material (e.g., titanium alloy requires high rigidity).

[0270] Technical principle:

[0271] The parameter mapping between different coordinate systems is achieved through coordinate transformation matrices, ensuring consistency between theoretical design and actual processing.

[0272] Tool compensation corrects tool wear and installation errors based on spatial geometry, while servo parameter optimization improves dynamic response performance through PID control theory.

[0273] Technical effects:

[0274] Processing parameter pre-adjustment time is ≤2 minutes (traditional manual pre-adjustment takes 5-8 minutes), reducing preparation time by 60%;

[0275] The pre-adjustment error is ≤ ±0.002mm, which ensures the accuracy of subsequent processing.

[0276] 5.3 Processing Execution and Correction Steps (corresponding to step 5 of the system workflow)

[0277] Implementation process:

[0278] Worktable positioning: The machining center drives the XYZ axes to move to the initial machining position according to the preset parameters, and the grating ruler feeds back the actual position (e.g., X=100.001mm, Y=50.002mm). The error compared with the theoretical value is ≤0.001mm.

[0279] Machining process execution: Spindle starts (8000 rpm), tool cuts according to G-code trajectory, and laser interferometer collects position data at 1000 Hz frequency;

[0280] Real-time deviation detection: The real-time feedback correction module compares the actual position with the theoretical trajectory and calculates the deviation value e(t). When e(t) > ±0.003mm, PID compensation is activated.

[0281] Dynamic parameter adjustment: Calculate the control quantity u(t) according to equation (4), adjust the output torque of the servo motor (such as increasing the X-axis motor current by 0.5A), and correct the motion trajectory;

[0282] Processing Completion and Data Upload: After processing is completed, the system automatically detects the processing dimensions (through an in-machine probe) and uploads the processing time (e.g., 25 minutes) and accuracy error (e.g., ±0.0025mm) to the data center.

[0283] Technical principle:

[0284] Laser interferometers achieve nanometer-level position feedback through optical path difference measurement, providing high-precision data for real-time correction;

[0285] The PID compensation algorithm quickly suppresses position deviation and ensures dynamic following performance through the synergistic effect of proportional, integral, and derivative components.

[0286] Technical effects:

[0287] The machining process has a positional accuracy of ≤±0.003mm (compared to ±0.005-±0.01mm in traditional machining), meeting the high-precision requirements of the aerospace field;

[0288] The utilization rate of machining centers has increased to over 85% (compared to 60%-70% in the traditional model), significantly improving equipment utilization.

[0289] In summary, this system, through its innovative architecture of "offline automated contact measurement + intelligent algorithm calibration + real-time data interaction," effectively solves the pain points of low efficiency, insufficient precision, and poor compatibility in traditional sheet metal processing, providing a highly efficient, high-precision, and intelligent position calibration solution for the high-end manufacturing field.

[0290] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4 As shown, the electronic device may include the above-mentioned Figure 3 The XXX device shown. Optionally, the electronic device 410 may include a first processor 2001.

[0291] Optionally, the electronic device 410 may also include a memory 2002 and a transceiver 2003.

[0292] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0293] The following is combined Figure 4 A detailed description of each component of electronic device 410 is provided below:

[0294] The first processor 2001 is the control center of the electronic device 410. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0295] Optionally, the first processor 2001 can perform various functions of the electronic device 410 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0296] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 are shown in the diagram.

[0297] In a specific implementation, as one example, the electronic device 410 may also include multiple processors, for example... Figure 4 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0298] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0299] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently and be connected via the interface circuit of the electronic device 410. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0300] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0301] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 4 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0302] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A system for offline automatic contact counting and position calibration of sheet metal parts, characterized in that, include: ‌ An offline data acquisition workstation is used to collect three-dimensional coordinate data of multiple feature points on a board and to preprocess the coordinate data. The data center server cluster is connected to the offline data matching workstation to receive and store the processed coordinate data. It has a built-in weighted least squares position calibration optimization algorithm to calculate the position calibration parameters of the plate's six degrees of freedom based on the coordinate data and the theoretical design coordinates of the plate. as well as The intelligent machining center is connected to the data center server cluster and is used to retrieve the position calibration parameters and adjust the machining coordinate system and tool compensation parameters based on the parameters to perform sheet metal machining.

2. The system according to claim 1, characterized in that, The offline contact counting workstation includes a high-precision contact counter, an XYZ motion platform, and an intelligent positioning fixture; The high-precision contact-type counter is used to contact feature points on the surface of the plate and generate trigger signals under the drive of the XYZ motion platform. The XYZ motion platform is used to drive the contact counter to move along a preset path in three-dimensional space, and the grating ruler on it is used to synchronously collect the actual three-dimensional coordinates of the contact counter tip in the global coordinate system when the contact is triggered. The intelligent positioning fixture is used to fix the board to be tested.

3. The system according to claim 2, characterized in that, The offline data matching workstation also includes: The data acquisition and preprocessing module is configured to: receive the raw coordinate data acquired by the grating ruler, and perform median filtering and mean filtering on the raw coordinate data to remove noise signals; The board identification module is configured to read board identification information through a barcode reader or RFID reader to associate it with the board's CAD design model and theoretical design coordinates.

4. The system according to claim 1, characterized in that, The weighted least squares position calibration optimization algorithm is configured as follows: Based on the actual three-dimensional coordinates, theoretical design coordinates, and dynamically allocated weight matrix, a weighted objective function is constructed. and By iteratively solving for the minimum value of the weighted objective function, the translation parameters (X0, Y0, Z0) and attitude Euler angles (α, β, γ) of the origin of the plate's own coordinate system in the global collision number coordinate system are calculated, thereby realizing the accurate mapping of the plate from the design coordinate system to the machining coordinate system.

5. The system according to claim 1, characterized in that, The data center server cluster transmits coordinate data with the offline data synchronization workstation via the Modbus TCP protocol, and transmits position calibration parameters with the intelligent machining center via the Profinet protocol.

6. The system according to claim 1, characterized in that, The intelligent machining center includes: The data retrieval and verification module is configured to request and obtain the location calibration parameters from the data center server cluster based on the board identification information. The machining parameter pre-adjustment module is configured to calculate the initial working position of the machining center and the tool compensation value based on the position calibration parameters (X0, Y0, Z0, α, β, γ). The motion control and machining execution module is configured to control the machining center to perform machining based on pre-tuned parameters.

7. The system according to claim 6, characterized in that, The intelligent machining center also includes a real-time feedback correction module, which is configured as follows: During the processing, the actual position data of the worktable is collected in real time; The actual location data is compared with the theoretical processing trajectory generated based on the location calibration parameters; When the deviation exceeds the preset threshold, the PID compensation algorithm is automatically triggered to dynamically adjust the servo motor output in order to correct the motion trajectory.

8. The system according to claim 1, characterized in that, It also includes edge computing nodes deployed on the offline data collection workstations and / or data centers, the edge computing nodes being configured to perform localized real-time preprocessing on the collected coordinate data, including filtering and outlier removal. ‌ 9. The system according to claim 1, characterized in that, The offline data processing workstation, data center server cluster, and intelligent machining center are connected via a layered distributed communication architecture built on industrial Ethernet, with each layer exchanging data through encrypted communication.

10. A method for offline automatic contact counting and position calibration of sheet metal parts, applied to the system according to any one of claims 1-9, characterized in that, Includes the following steps: ‌ Panel loading and information identification: The panel is fixed on the offline contact measurement workstation, and the identification module is used to obtain the panel identification information to associate it with its CAD model and theoretical coordinates; Automatic collision counting and data acquisition: The offline collision counting workstation plans and executes the collision counting path, drives the collision counting device to contact multiple feature points on the surface of the board, synchronously collects the actual three-dimensional coordinate data of each point, and performs filtering preprocessing; Calibration parameter calculation: Input the actual three-dimensional coordinate data and its corresponding theoretical design coordinates into the weighted least squares position calibration optimization algorithm to calculate the translation parameters (X0, Y0, Z0) and attitude Euler angles (α, β, γ) of the plate. Parameter transmission and machining pre-adjustment: The calculated position calibration parameters are transmitted to the intelligent machining center, which performs machining coordinate system transformation and tool compensation pre-adjustment based on the parameters; Machining execution and real-time correction: The machining center executes machining according to the preset parameters and provides real-time feedback on the actual position during the machining process, compares it with the theoretical trajectory, and dynamically corrects the deviation.

Images