In the aerospace industry, the demand for high-performance rocket engines with increased thrust and thrust-to-weight ratios has led to increasingly stringent service environments. As a result, critical components such as turbine housings, which are complex in structure, require more rigorous inspection of profile dimensions. Traditional inspection methods for aerospace castings often suffer from low coverage of profile dimension detection, inefficiency, and high manual labor intensity. To address these challenges, we have developed a three-dimensional online measurement system based on 3D laser scanning technology. This system enables non-contact, rapid, and high-precision acquisition of surface parameters and point cloud data, overcoming the drawbacks of conventional methods. In this paper, we present the design, implementation, and testing of this system, emphasizing its application to aerospace castings. We will detail the modular architecture, incorporate formulas and tables to summarize key aspects, and demonstrate its effectiveness through a case study.
The advancement of 3D laser scanning has revolutionized dimensional measurement in various fields, including reverse engineering, digital cities, cultural heritage preservation, deformation monitoring, and precision industrial measurement. For aerospace castings, which often feature intricate geometries and tight tolerances, this technology offers a transformative solution. Prior research has explored automated optical inspection systems with robotic integration and calibration methods for rotational scanning. Building on these foundations, our system integrates robotics, vision systems, and advanced software to achieve full automation in the inspection process for aerospace castings. The core innovation lies in the seamless coordination of measurement, logistics, and control modules, ensuring high coverage and accuracy while reducing human intervention.
Our three-dimensional online measurement system is specifically designed for the dimensional inspection of aerospace castings. It comprises four main modules: the measurement module, logistics module, integrated control module, and safety protection module. Each module plays a critical role in ensuring efficient and accurate inspection. The system leverages blue light scanning technology for high-resolution data capture, robotic manipulators for precise positioning, and vision systems for automated workpiece handling. By integrating these components, we achieve a turnkey solution that automates the entire workflow from workpiece identification to report generation. The design prioritizes scalability and adaptability to accommodate various types of aerospace castings, from small intricate parts to large structural components.
The measurement module is responsible for acquiring 3D point cloud data from aerospace castings. It consists of a measurement robot, scanner, calibration plate, measurement software, analysis software, and a workstation. The measurement robot, with six degrees of freedom, positions the scanner relative to the workpiece. The scanner uses structured blue light to project grating patterns onto the surface, capturing distortions with CCD cameras to generate point clouds. Calibration is performed regularly using a calibration plate to maintain accuracy. The software suite handles data acquisition, stitching, and comparison with CAD models. To summarize the key components, we present Table 1.
| Component | Function | Specifications/Features |
|---|---|---|
| Measurement Robot | Positions scanner with 6-DOF | High repeatability, brake locking for stability |
| Scanner | Captures 3D point cloud data | Blue light, dual CCD, resolution up to 0.01 mm |
| Calibration Plate | Calibrates scanner accuracy | Certified reference artifacts |
| Measurement Software | Controls scanning and data stitching | Automatic feature-based alignment |
| Analysis Software | Compares scan data to CAD models | 3D deviation analysis, report generation |
| Workstation | Processes data and runs software | High-performance GPU for real-time processing |
The logistics module manages the handling and positioning of aerospace castings during inspection. It includes a material handling robot, vision positioning system, rotary table, material stations, and transport fixtures. The material robot, also with six degrees of freedom, transfers workpieces between stations. The vision system uses 2D cameras to identify workpiece QR codes and adjust robot paths for precise picking and placement. The rotary table acts as a seventh axis, rotating the workpiece to enable comprehensive scanning without excessive robot movement. Material stations consist of carts with customized pallets for different aerospace castings, ensuring consistent positioning. Transport fixtures are equipped with quick-change mechanisms to adapt to various part geometries. Table 2 outlines the logistics components.
| Component | Function | Key Characteristics |
|---|---|---|
| Material Robot | Handles workpiece loading/unloading | 6-DOF, payload capacity >50 kg |
| Vision Positioning System | Identifies workpieces and corrects coordinates | 2D camera, large field of view, QR code reading |
| Rotary Table | Rotates workpiece for full coverage scanning | Programmable positions, high torque |
| Material Stations | Holds workpieces for access | Modular pallets,定位轨道 for alignment |
| Transport Fixtures | Grips and moves workpieces | Quick-change grippers, pneumatic locking |
The integrated control module orchestrates the entire system for inspecting aerospace castings. It features a control cabinet with PLCs and a touchscreen interface. The PLC communicates with all subsystems via I/O points or bus protocols, ensuring synchronized operation. The touchscreen provides a user-friendly interface for initiating scans, monitoring progress, and accessing reports. Safety interlocks are implemented to prevent collisions and ensure operational integrity. This module enables fully automated batch inspection with minimal operator intervention, streamlining the workflow for aerospace castings production.
Safety protection is paramount when dealing with heavy aerospace castings and moving robots. The safety module includes mechanical safety switches, door locks, emergency stops, and safety fencing rated at PL e level. The system incorporates electrical interlocks that prevent operation unless all guards are secured. Emergency stops are strategically placed for immediate shutdown. This design protects personnel and equipment, ensuring compliance with industrial safety standards for aerospace manufacturing environments.
The system layout is optimized for efficient flow of aerospace castings. Two material stations allow continuous loading and unloading, while the measurement and material robots operate in coordinated paths to avoid interference. The workspace is enclosed by safety fencing with access doors. Operators place workpieces on material carts outside the enclosure, then initiate the process via the touchscreen. The system automatically identifies, transports, scans, and returns the aerospace castings, producing inspection reports without manual intervention. This layout maximizes throughput and minimizes downtime, crucial for high-volume inspection of aerospace castings.
To quantify the performance of our system for aerospace castings, we define key metrics. The coverage rate of profile inspection is critical. It is calculated as the ratio of detected surface area to the total surface area of the aerospace casting. Mathematically, this is expressed as:
$$ \text{Coverage Rate} = \frac{A_{\text{detected}}}{A_{\text{total}}} \times 100\% $$
where \( A_{\text{detected}} \) is the area captured by the scan, and \( A_{\text{total}} \) is the theoretical surface area including any repaired holes. For aerospace castings with complex geometries, achieving high coverage is challenging due to occlusions and surface reflectivity. Our system employs multi-view scanning with automatic stitching to maximize \( A_{\text{detected}} \).
Measurement accuracy is another vital parameter. The scanner’s precision is influenced by factors such as calibration, environmental stability, and data processing algorithms. We model the overall measurement error \( E \) as a combination of systematic and random errors:
$$ E = \sqrt{E_{\text{systematic}}^2 + E_{\text{random}}^2} $$
where \( E_{\text{systematic}} \) arises from calibration residuals and \( E_{\text{random}} \) from noise in point cloud acquisition. Through rigorous calibration and filtering techniques, we minimize \( E \) to meet the tight tolerances required for aerospace castings.
We tested our system on a low-pressure housing blank for an oxygen pump, a representative aerospace casting. The workpiece was scanned using automated routines, and the point cloud was compared to its CAD model. The coverage rate was computed using the formula above. Additionally, specific dimensions, such as diameters and heights, were extracted and compared to design specifications. The results are summarized in Table 3.
| Parameter | Design Value (mm) | Tolerance (mm) | Measured Range (mm) | Compliance |
|---|---|---|---|---|
| Diameter ϕ209.5 | ϕ209.5 | ϕ209 -0/-2 | ϕ209.284 – ϕ209.440 | Yes |
| U-slot Height 24.5 | 24.5 | 24.5 -0/-2 | 24.312 – 24.492 | Yes |
| Surface Coverage | Total area: 1,362,750.49 mm² | N/A | Detected area: 1,337,207.71 mm² | Coverage: 98.1% |
The high coverage rate of 98.1% demonstrates the system’s ability to capture nearly the entire surface of the aerospace casting. This is achieved through optimized scanning paths and the use of the rotary table to expose all angles. The dimensional measurements fall within specified tolerances, confirming the accuracy of the system. The inspection process, from loading to report generation, was completed autonomously, showcasing the efficiency gains for aerospace castings production.
Further analysis involves statistical process control for aerospace castings inspection. We can define a process capability index \( C_p \) for critical dimensions:
$$ C_p = \frac{\text{Tolerance Width}}{6\sigma} $$
where \( \sigma \) is the standard deviation of measured dimensions across multiple aerospace castings. A \( C_p \geq 1.33 \) indicates a capable process. Our system facilitates the collection of large datasets to compute such indices, enabling continuous improvement in the manufacturing of aerospace castings.
The integration of vision systems for workpiece identification adds another layer of robustness. The accuracy of robot positioning via vision correction can be modeled by error propagation. If the camera has a pixel error \( \delta_p \) and a calibration error \( \delta_c \), the total positioning error \( \delta_{\text{pos}} \) in world coordinates is:
$$ \delta_{\text{pos}} = \sqrt{ \left( \frac{\partial f}{\partial p} \delta_p \right)^2 + \delta_c^2 } $$
where \( f \) is the transformation function from pixel to world coordinates. Our vision system minimizes \( \delta_{\text{pos}} \) through high-resolution cameras and precise calibration, ensuring reliable handling of aerospace castings.
In terms of throughput, the system’s cycle time \( T_{\text{cycle}} \) for inspecting one aerospace casting can be broken down as:
$$ T_{\text{cycle}} = T_{\text{load}} + T_{\text{scan}} + T_{\text{process}} + T_{\text{unload}} $$
where \( T_{\text{load}} \) and \( T_{\text{unload}} \) are handled by the material robot, \( T_{\text{scan}} \) depends on scan speed and coverage, and \( T_{\text{process}} \) involves data analysis. With parallel operations and optimized paths, we achieve a cycle time of under 5 minutes for typical aerospace castings, significantly faster than manual methods.
The system’s adaptability to different aerospace castings is enabled by the modular design. Each workpiece type has dedicated pallets and grippers, which can be swapped automatically via the gripper library. The scanning programs are pre-configured based on CAD data, allowing quick changeovers. This flexibility is essential for small-batch production common in aerospace casting foundries.
Comparing our system to traditional methods for aerospace castings inspection, the advantages are clear. Traditional techniques like coordinate measuring machines (CMM) or manual gauging offer limited point coverage and are time-consuming. Our system provides full-field data, enabling comprehensive defect detection and trend analysis. The automation reduces labor costs and minimizes human error, critical for quality assurance in aerospace castings.
Future enhancements could involve integrating artificial intelligence for real-time defect classification in aerospace castings. Deep learning algorithms could analyze point cloud data to identify porosity, cracks, or deviations beyond dimensional tolerances. Additionally, the system could be linked to digital twin platforms for predictive maintenance and process optimization in aerospace casting production.
In conclusion, we have successfully developed and implemented a three-dimensional online measurement system based on 3D laser scanning for the dimensional inspection of aerospace castings. The system’s modular architecture, encompassing measurement, logistics, control, and safety, ensures high coverage, accuracy, and efficiency. Through a case study on an oxygen pump housing, we demonstrated a coverage rate of 98.1% and compliance with dimensional tolerances. The use of robotics, vision, and advanced software automates the entire process, addressing the limitations of traditional methods. This system represents a significant advancement in quality control for aerospace castings, contributing to the reliability and performance of rocket engines. As the aerospace industry evolves, such automated inspection solutions will become increasingly vital for manufacturing complex components with precision and consistency.