Language

HomeSoftware

TIANYI Core software family

Constructing a future welding factory model featuring "digital twin as the mirror, AI algorithm as the brain, and intelligent equipment as the hands", not only addresses the pain points of traditional welding, but also achieves multiple breakthroughs in production efficiency, quality, and humanistic care through technological innovation, providing a replicable practical path for the implementation of intelligent manufacturing.

Contact Us

AI+ Welding Factory Comprehensive Solution

Traditional welding workshops have long grappled with persistent challenges, including poor working environments, high physical labor intensity, significant health risks, and inconsistent efficiency and quality. Aiming for full automation, unmanned operations, and intelligent control, the Smart Lighthouse Welding Factory constructs "Dark Factory"-grade welding production lines. By leveraging technologies such as digital twins, AI, and intelligent sensing, it liberates workers from hazardous operating environments while simultaneously achieving a comprehensive upgrade in efficiency, quality, and management—thereby setting a new benchmark for intelligent manufacturing.

Core Technical Architecture: Seven Intelligent Workstations

Centered on digital-driven operations and intelligent control, it encompasses a complete closed-loop process spanning from design to inspection, with seven major work sections operating in seamless collaboration.

  • 1. Digital Twin Remote Monitoring and Control

    • Construct a 1:1 virtual mirror of the production line, synchronizing equipment, progress, and environmental data in real time.
    • Remote monitoring, dispatching, and command issuance from the office—no need to enter the workshop.

  • 2. Digital simulation process design

    • Finite element / thermodynamic simulation, virtual welding trial, early prediction of penetration depth and deformation
    • Replace empirical trial-and-error, quickly output the optimal process, reduce welding trial costs

  • 3. Intelligent Sensing and 3D Modeling

    • Line laser + 3D vision high-precision scanning, generate workpiece point cloud data
    • Automatically extract geometric features, gaps, and bevels, precisely plan welding path

  • 4. AI intelligent process generation and distribution

    • Machine learning trains process decision-making model, automatically match parameters for material / thickness / bevel
    • One-click issue to the robot and power supply, achieve intelligent adaptation of the process

  • 5. Dynamic correction and process control

    • Real-time online detection of molten pool, arc light, and arc
    • Welding while adjusting, closed-loop control, digital twin synchronization visualization, full traceability throughout

  • 6. Intelligent stress elimination

    • Automated heat treatment / vibration aging, automatic processing after welding
    • Based on simulation data, preset parameters, no intervention throughout the process

  • 7. Full-dimensional detection and quality traceability

    • Ultrasonic / X-ray / visual multi-dimensional defect detection
    • Data is linked, forms a complete quality archive, traceable throughout the life cycle

Weld Seam Tracking

Intelligent Laser Vision, Precise Seam Positioning, and Automated Welding Solutions: Leveraging machine vision and laser triangulation, this system achieves automated seam recognition, path planning, and real-time tracking—thereby resolving critical industry pain points such as low precision, poor efficiency, and insufficient stability inherent in traditional welding processes.

Technical Overview | The Core of Intelligent Manufacturing Welding

Technically defined, weld seam tracking is an intelligent, machine-vision-based welding assistance technology. By utilizing sensors to scan the weld seam location, it automatically plans the welding path, ensuring that the welding torch remains precisely aligned with the seam trajectory; it serves as a critical supporting technology for industrial welding automation.

  • Core Technological Advantages

    • High-Precision Positioning: Guidance accuracy of ±0.5 mm
    • High-Speed ​​Processing: Single-frame processing in 50–100 ms
    • Rapid Response: System response time of <500 ms
    • Strong Adaptability: Compatible with multiple materials and welding methods

  • Pain Points of Traditional Welding

    • Manual teaching is inefficient and requires lengthy programming time.
    • Workpiece assembly errors lead to weld misalignment.
    • Thermal deformation compromises the stability of weld quality.
    • Complex weld seams are difficult to automate.

Core Algorithms | Precise Recognition and Calibration

Image Processing Algorithm Workflow: Image Preprocessing → Laser Stripe Center Extraction → Weld Feature Point Calculation → Real-time Weld Tracking.
    Algorithmic Advantages:
    • Mature and stable; high computational speed
    • No training data required; plug-and-play functionality
    • Low resource consumption; suitable for embedded systems
    • High interpretability; facilitates easy debugging
     
    Hand-Eye Calibration System:
    • Coordinate Chain: Image Coordinates → Camera Coordinates → Robot Base Coordinates
    • Two Mounting Configurations: Eye-in-Hand / Eye-to-Hand
    • Calibration Accuracy: Reprojection Error < 0.5 mm

Teach-free Welding Robot

Leveraging 3D vision and intelligent algorithms, this system enables teach-free, fully automated, and high-precision welding for ultra-long workpieces and ribbed components—thereby significantly reducing manual programming costs.

Intelligent Robotic Welding System

Teachless Programming | Intelligent Vision-Guided Seam Finding | Digital Twin Simulation | One-Touch Operation

  • Offline Programming and Simulation

    Toolpath Planning, Process Simulation, and Digital Twin Pre-visualization

  • Robotic Welding Control System

    Bus Control, Real-time Algorithms, Multi-axis Coordination

  • 3D Vision Guidance Module

    Initial Positioning + Laser Seam Tracker + Line Scan Camera

Core Strengths

System Intelligence

  • • Digital Twin: The nation's first welding digital twin, enabling full-process simulation and rehearsal.
    • One-Touch Operation: Single-click extraction of weld seams, process parameter matching, and path planning.
    • Integrated Vision & Intelligence: Automatic visual positioning and blueprint-free reverse modeling.
    • Automated Calibration: Rapid, one-touch calibration to restore robot precision.

Simplified Learning and Usage

  • • Beginner-Friendly: Algorithms replace manual programming—no specialized background required.
    • Built-in Expert Process Library: Automatically matches parameters to ensure stable weld quality.

Production Digitization

  • • Cloud Connectivity: Multi-device Monitoring, Remote Mapping, Mobile Management.
    • Smart Display: Real-time Device Status Acquisition, Data Visualization.

Off-line Programming & Digital Twin

Development, application, and optimization of VC, Simufact Welding, Ignition SCADA, and AI technologies; offline programming, virtual commissioning, digital twins, welding process simulation, and intelligent applications; optimization of the integration of multi-brand equipment and vision systems; construction of an intelligent WPS process library; and the establishment of end-to-end intelligent manufacturing service capabilities.

    • Offline Programming: Efficient Adaptation, Intelligent Optimization

      Optimize the VC post-processing program to ensure compatibility across multiple equipment brands, thereby breaking down equipment barriers. Develop a cross-platform VC programming module that integrates WPS parameters from Simufact Welding and leverages AI technology to recommend optimized programming paths. Establish an intelligent WPS process library to facilitate the efficient linkage between programming and process parameters, thereby enhancing both programming efficiency and precision.

    • Virtual Commissioning: Reduce Costs, Boost Efficiency, Standardize Delivery

      Simulate the collaborative operation of multi-brand equipment on a virtual commissioning platform to validate robot programs and sensor interaction logic. Integrate AI to optimize the debugging workflow, significantly reducing on-site commissioning time and lowering project delivery costs. Establish a standardized virtual commissioning process to ensure project delivery quality and enhance customer satisfaction.

    • Welding process simulation: Precise prediction, scientific optimization

      Leveraging Simufact Welding, accurately simulate a wide range of welding scenarios—including arc welding, spot welding, and laser welding. By integrating AI machine learning models, optimize welding sequences, heat input, and clamping strategies, while predicting welding deformation and residual stress in advance. Generate standardized WPS parameters and synchronize them with an intelligent WPS process library, enabling the reuse and optimization of process parameters.

    • Digital Twin: Real-time Monitoring, Intelligent Decision-making

      By integrating Ignition SCADA with AI technology, a digital twin model for the entire production line is constructed, integrating welding process simulation data. Real-time monitoring of the production line, AI-driven fault prediction and optimization decision-making are achieved, enhancing production stability. Expanding applications such as cross-device collaborative optimization, customer remote support, and predictive maintenance, enabling enterprises to achieve lean production.

    • AI Technology Integration: Empowering the Entire Process, Enhancing Intelligence Level

      Developing AI models such as machine learning and deep learning to optimize Simufact Welding parameters, VC programming paths, and Ignition data analysis efficiency. Creating an AI-driven WPS process library to achieve intelligent parameter recommendation, adaptive adjustment, and iterative learning based on historical data. Strengthening the predictive capability of digital twins, with fault prediction accuracy reaching 90%, reducing equipment failure rates.

    • Full-cycle Service: From Implementation to Empowerment, Full-Scale Support

      Project implementation: Responsible for the full-process application of core technologies in production line design, programming, and simulation verification, meeting customer customized requirements. Training support: Providing core technology training for internal employees and offering system integration, welding optimization, and AI value-added services to customers. Market cooperation: Deeply collaborating with industry-leading manufacturers and technical communities, tracking market trends, and formulating intelligent manufacturing strategies.

Molten Pool Online Monitoring

By capturing images of the weld pool in real-time using high-speed cameras—and integrating deep learning algorithms—we achieve precise extraction of weld pool features, intelligent prediction of weld quality, and proactive intervention in the welding process, thereby comprehensively elevating the levels of welding automation and intelligence.

Real-time Monitoring | Deep Learning | Multimodal Fusion | Defect Warning | Remote Transmission

  • Core Technological Advantages:

    1. Deep Learning-based Intelligent Recognition: Enhanced CNN architecture for precise identification of weld penetration status.
    2. Multimodal Data Fusion: Integration of image, audio, and temperature data for highly accurate predictions.
    3. High-Speed ​​Data Acquisition (30 fps): Captures transient changes in the weld pool with clear and stable imagery.
    4. Intelligent Parameter Optimization: Automatically generates the optimal image processing configuration.
    5. Three-Tier Warning Mechanism: Audio-visual alerts → Parameter correction → Emergency shutdown.
    6. Remote Data Transmission: Real-time data upload via WebSocket, enabling cloud-based monitoring.

  • Core Function Modules:

    1. System Initialization: Comprehensive configuration of cameras, sensors, communication modules, and process parameters.
    2. Acquisition Control: Dynamic data acquisition, area statistics, and optimal frame selection.
    3. Process Recipe Management: Parameter saving, rapid recall, and batch management.
    4. Deep Learning Inspection: Automated segmentation, annotation, and model inference.
    5. Intelligent Defect Pre-warning: Real-time identification of five major categories of welding defects.
    6. Remote Data Transmission: Real-time uploading of images and process parameters.

Real-time Monitoring Parameters

Weld Pool Geometry: Length, Width, Area, Perimeter, Aspect Ratio, Coordinates; Welding Status: Wire Stick-out, Wire Diameter, Arc Spot Position, Travel Angle; Environmental Perception: Weld Pool Temperature, Welding Sound, Operational Status
Welding Defect Recognition
Defect Type
    
Stoma
 Weld Pool Bubbles / Gas Interruption / Wire Sticking
Level 1 Audible and Visual Alarm
Missed weld
 Abnormal welding wire, intermittent welding
Level 2 Audible and Visual Alarm
Undercut
Excessive current, unstable speed
Level 2 Audible and Visual Alarm
Weld Misalignment
Weld bead offset, workpiece displacement
Level 2 Audible and Visual Alarm
Lack of fusion
Blurred edges, asymmetric temperature field
Level 3 Audible and Visual Alarm

Application Value

  • • Visualization, Digitalization, and Intelligentization of the Entire Welding Process
    • Predict quality in advance, intervene proactively, and reduce the scrap rate.
    • Replaces manual inspection; offers high precision, high efficiency, and zero errors.
    • Accumulate welding big data to support process optimization.
    • ACompatible with automated production lines, meeting the demands of high-end manufacturing.

Weld Inspection

The weld surface quality inspection system, based on line-structured light 3D vision and integrated with robotic automated scanning, enables non-contact, high-precision, and end-to-end intelligent detection of weld defects. It provides stable and reliable quality control solutions for welding applications across sectors such as automotive components, steel structures, and new energy enclosures.

 
Testing Process: Utilizing line-structured light triangulation, the system acquires a point cloud of the workpiece's contour via sensors; through image stitching and algorithmic analysis, it automatically evaluates the quality of the weld seam.

Testing Process

  • Constant-Speed ​​Scanning

    The robot carries a sensor and scans along the weld seam at a constant speed.

  • Image Stitching

    Acquire multi-frame contour point clouds and stitch them together to form an image.

  • Automatically Segment Weld Regions

    Fit the base metal and weld profiles, and automatically segment the weld region.

  • Identify and Classify Defects

    Match defect models to identify and classify defects.

Core Advantages of the Solution：
  • • High-Precision Inspection: Repeatable positioning accuracy of ±0.1 mm, enabling precise identification of minute defects.
    • Comprehensive Defect Coverage: Capable of detecting standard defects such as weld length, width, and height, as well as burn-through, weld deviation, porosity, undercut, and intermittent welds.
    • Flexible Adaptability: Supports various welding processes, including MIG/MAG and laser welding.
    • Stable and Interference-Resistant: Optimized for industrial environments; features an integrated light-shielding solution to effectively mitigate the impact of strong ambient light.
    • Data Traceability: Inspection results are linked directly to specific workpieces, integrating with MES systems to enable end-to-end process control.