Aircraft Assembly Snags: Human Errors or Lack of Production Design?

To err is an intrinsic human trait, which means that human errors, at some point, are inevitable. Business improvement tools and practices neglect to deal with the root causes of human error; hence, they ignore certain design considerations that could possibly prevent or minimise such errors from occurring. Recognising this gap, this paper seeks to conceptualise a model that incorporates cognitive science literature based on a mistake-proofing concept, thereby offering a deeper, more profound level of human error analysis. An exploratory case study involving an aerospace assembly line was conducted to gain insights into the model developed. The findings of the case study revealed four different causes of human errors, as follows: (i) description similarity error, (ii) capture errors, (iii) memory lapse errors, and (iv) interruptions. Based on this analysis, error-proofing measures have been proposed accordingly. This paper lays the foundation for future work on the psychology behind human errors in the aerospace industry and highlights the importance of understanding human errors to avoid quality issues and rework in production settings, where labour input is of paramount importance.

Implementation of Object Recognition Algorithm to enhance Manufacturing and Maintenance Tasks on an Aircraft

The objective of this thesis was to evaluate the viability of implementation of an object recognition algorithm driven by deep learning for aerospace manufacturing, maintenance and assembly tasks. Comparison research has found that current computer vision methods such as, spatial mapping was limited to macro-object recognition because of its nodal wireframe analysis. An optical object recognition algorithm was trained to learn complex geometric and chromatic characteristics, therefore allowing for micro-object recognition, such as cables and other critical components. This thesis investigated the use of a convolutional neural network with object recognition algorithms. The viability of two categories of object recognition algorithms were analyzed: image prediction and object detection. Due to a viral epidemic, this thesis was limited in analytical consistency as resources were not readily available. The prediction-class algorithm was analyzed using a custom dataset comprised of 15 552 images of the MaxFlight V2002 Full Motion Simulator’s inverter system, and a model was created by transfer-learning that dataset onto the InceptionV3 convolutional neural network (CNN). The detection-class algorithm was analyzed using a custom dataset comprised of 100 images of two SUVs of different brand and style, and a model was created by transfer-learning that dataset onto the YOLOv3 deep learning architecture. The tests showed that the object recognition algorithms successfully identified the components with good accuracy, 99.97% mAP for prediction-class and 89.54% mAP. For detection-class. The accuracies and data collected with literature review found that object detection algorithms are accuracy, created for live -feed analysis and were suitable for the significant applications of AVI and aircraft assembly. In the future, a larger dataset needs to be complied to increase reliability and a custom convolutional neural network and deep learning algorithm needs to be developed specifically for aerospace assembly, maintenance and manufacturing applications.

Implementation of Object Recognition Algorithm to enhance Manufacturing and Maintenance Tasks on an Aircraft

The objective of this thesis was to evaluate the viability of implementation of an object recognition algorithm driven by deep learning for aerospace manufacturing, maintenance and assembly tasks. Comparison research has found that current computer vision methods such as, spatial mapping was limited to macro-object recognition because of its nodal wireframe analysis. An optical object recognition algorithm was trained to learn complex geometric and chromatic characteristics, therefore allowing for micro-object recognition, such as cables and other critical components. This thesis investigated the use of a convolutional neural network with object recognition algorithms. The viability of two categories of object recognition algorithms were analyzed: image prediction and object detection. Due to a viral epidemic, this thesis was limited in analytical consistency as resources were not readily available. The prediction-class algorithm was analyzed using a custom dataset comprised of 15 552 images of the MaxFlight V2002 Full Motion Simulator’s inverter system, and a model was created by transfer-learning that dataset onto the InceptionV3 convolutional neural network (CNN). The detection-class algorithm was analyzed using a custom dataset comprised of 100 images of two SUVs of different brand and style, and a model was created by transfer-learning that dataset onto the YOLOv3 deep learning architecture. The tests showed that the object recognition algorithms successfully identified the components with good accuracy, 99.97% mAP for prediction-class and 89.54% mAP. For detection-class. The accuracies and data collected with literature review found that object detection algorithms are accuracy, created for live -feed analysis and were suitable for the significant applications of AVI and aircraft assembly. In the future, a larger dataset needs to be complied to increase reliability and a custom convolutional neural network and deep learning algorithm needs to be developed specifically for aerospace assembly, maintenance and manufacturing applications.

New Indices for Evaluating Vibration Characteristics of Flexible-Joint Robots

Structural vibration is a significant consideration for robotic applications such as machining where the robot is subject to large dynamic loading. Aiming at providing an efficient means to evaluating the vibration characteristics of industrial robots for these applications, this work proposes two new indices to quantify the elastic displacement of the tool mounted on the robot caused by the vibrations induced by external process loading for flexible-joint robots. For this purpose, a structural dynamic model is first developed to derive the frequency responses of the tool displacement. Then, the displacement-force and displacement-torque frequency response ratios are defined, which represent the mapping from the amplitudes of an external harmonic force and torque to the amplitude of tool displacement respectively. The upper bounds of the two ratios are used as evaluation indices for the vibration characteristics of the robot, which represent the worst situation of the tool displacement due to harmonic excitation with amplitude of unit force and unit torque respectively. With these indices, an efficient method is provided to predict whether the tool misalignment caused by periodic loading is acceptable for process quality requirement. Numerical simulation demonstrates the effectiveness of the proposed method for a robotic riveting system being developed for aerospace assembly.

Hybrid Nonlinear Variation Modeling of Compliant Metal Plate Assemblies Considering Welding Shrinkage and Angular Distortion

Ship assembly involves thousands of large dimensional compliant metal plates. These compliant metal plates are fully welded together by seam welding in the assembly process. Different from the automobile and aerospace assembly process, the final variation of ship assembly is significantly influenced by the geometric nonlinearity and welding deformation generated during the seam welding process. This paper develops a nonlinear variation model (NVM) to consider the geometric nonlinearity, welding shrinkage, and angular distortion based on elastic mechanics. Furthermore, the nonlinear variation model is calibrated by the composite Gaussian process (CGP) to compensate for other factors that are not considered in the nonlinear variation model. The proposed model is validated by a case study on the deviation prediction of an assembly of two compliant metal plates and compared with the existing methods. The results show that the proposed model has a significant improvement in prediction accuracy of assembly deviation.

Smart Multi-Sensor Monitoring in Drilling of CFRP/CFRP Composite Material Stacks for Aerospace Assembly Applications

Composite material parts are typically laid out in near-net-shape, i.e., very close to the finished product configuration. However, further machining processes are often required to meet dimensional and tolerance requirements. Drilling, edge trimming and slotting are the main cutting processes employed for carbon fiber-reinforced plastic (CFRP) composite materials. In particular, drilling stands out as the most widespread machining process of CFRP composite parts, chiefly in the aerospace industrial sector, due to the extensive use of mechanical joints, such as rivets, rather than welded or bonded joints. However, CFRP drilling is markedly challenging: due to CFRP abrasiveness, inhomogeneity and anisotropic properties, tool wear rates are inherently high leading to superior cutting forces and detrimental effects on workpiece surface quality and material integrity. Damage such as delamination, cracks or matrix thermal degradation is often observed as the result of uncontrolled tool wear or improper machining conditions. Sensor monitoring of drilling operations is, therefore, highly desirable for process conditions’ optimization and tool life maximization. The development of this kind of automated control technologies for process and tool state evaluation can notably contribute to the reduction of scraps and tool costs as well as to the improvement of process productivity in the drilling of CFRP composite material parts. In this paper, multi-sensor process monitoring based on thrust force and torque signal detection and analysis was applied during drilling of CFRP/CFRP laminate stacks for the assembly of aircraft fuselage panels with the scope to evaluate the tool wear state. Different signal-processing methods were utilised to extract diverse types of features from the detected sensor signals. A machine-learning approach based on an artificial neural network (ANN) was implemented to make smart decisions on the timely execution of tool change, which is highly functional for CFRP drilling process automation.

An Improved Robust Method for Pose Estimation of Cylindrical Parts with Interference Features

Horizontal docking assembly is a fundamental process in the aerospace assembly, where intelligent measurement and adjustable support systems are urgently needed to achieve higher automation and precision. Thus, a laser scanning approach is employed to obtain the point cloud from a laser scanning sensor. And a method of section profile fitting is put forward to solve the pose parameters from the data cloud acquired by the laser scanning sensor. Firstly, the data is segmented into planar profiles by a series of parallel planes, and ellipse fitting is employed to estimate each center of the section profiles. Secondly, the pose of the part can be obtained through a spatial straight line fitting with these profile centers. However, there may be some interference features on the surface of the parts in the practical assembly process, which will cause negative effects to the measurement. Aiming at the interferences, a robust method improved from M-estimation and RANSAC is proposed to enhance the measurement robustness. The proportion of the inner points in a whole profile point set is set as a judgment criterion to validate each planar profile. Finally, a prototype is fabricated, a series of experiments have been conducted to verify the proposed method.