Assembly Of Large Components Research Articles

Optimization of the production process using simulation modelling

The aim of this paper is to optimize the production process of graphics card assembly in an unnamed company. For the optimization of the production process, we will use the Tecnomatix Plant Simulation software from Siemens. The first and most important step to optimize the manufacturing process is to analyse the current process so that we can pinpoint where bottlenecks or downtime is occurring. For this reason, the first simulation is dedicated to the current production process. In the second simulation, we have successfully implemented all the proposed changes within the upgraded manufacturing process. These changes included replacing the human workforce with two robots in the tenth stage, which involves the assembly of large components. The implementation of the upgraded process allows 44 graphics cards to be assembled in one hour, an increase of 15 graphics cards compared to the current process. In an eight-hour shift, 353 graphics cards can be assembled, which is 118 graphics cards more than the current production process.

Type Design Method and Application for Spacecraft Large Component Assembly Robot Based on Features Constraint Relationship

Type Design Method and Application for Spacecraft Large Component Assembly Robot Based on Features Constraint Relationship

Combined Measurement Based Wing-Fuselage Assembly Coordination via Multiconstraint Optimization

In the field of aircraft manufacturing, the demand for high aerodynamic performance and long fatigue life is growing. This makes high-precision assembly of large components and assembly coordination increasingly important. This is particularly prominent for wing-fuselage joining due to the large size, measurement difficulty, and multi-constraint coordination. In this paper, we propose a novel optimization method for the wing-fuselage assembly coordination problem, as well as an assembly gap measurement approach. First, a gap measurement framework combining laser tracker, 3D scanner and photogrammetry is adopted to solve the occlusion problem that makes measurement difficult in traditional methods. A compatible tooling is utilized to transform the scanned point cloud data in the local coordinate system to the global aircraft assembly coordinate system. Based on this, a comprehensive wing-fuselage pose coordination model is then established, which takes engineering constraints into account to realize assembly gap redistribution for wing-fuselage joining. The transformation parameters of pose adjustment are solved by the proposed multi-constraint optimization algorithm based on the Gauss-Newton method. Finally, experiments on synthetic data and actual data have verified the effectiveness of the coordination method. Results show that our method significantly optimizes the gap distribution within the given tolerance and outperforms traditional methods in both accuracy and efficiency. With advantages of high efficiency and accuracy, our method can be well applied to the wing-fuselage joining and satisfy the requirement of the large component assembly.

Fiber optic sensor array fork-lug flexible monitoring of large components.

To realize the flexible monitoring and online pose correction of the complex assembly of large components, an optical fiber sensor array-based tooling condition monitoring network was built. First, an ANSYS simulation analysis was conducted to obtain the strain sensitive position of the tooling, and the maximum deformation was 0.0368 mm; then, the combination relationship between the monitoring position of the fiber grating sensor and the test data was determined. Next, a flexible sensor packaging structure was designed to achieve optimal strain sensitivity for ensuring flexible assembly. Finally, a three-dimensional reconstruction calculation model of stress field distribution and assembly pose adjustment was developed. The changes in the stress field distribution caused by different assembly problems were analyzed, and the corresponding pose correction parameters were solved. The results show that the inversion accuracy of the position deformation of the optical fiber sensor was better than 0.0058 mm, and the minimum correction accuracy of the position was 0.17 mm. The optical fiber flexible detection system works with the assembly structure to realize the intelligent assembly of large components and addresses the limitations of existing three-dimensional digital measurement processes due to environmental vibration, test openness, and other factors.

Correction to: Configuration optimization of the feature-oriented reference system in large component assembly

Correction to: Configuration optimization of the feature-oriented reference system in large component assembly

Configuration optimization of the feature-oriented reference system in large component assembly

Multi-station measurement is a key technology for the high accuracy and efficiency of large component assembly. The unity of measurement coordinate systems (MCSs) and assembly coordinate system (ACS) is achieved by registration of enhanced reference system (ERS) points. As the transformation errors of assembly features are related to the configuration of ERS points, and to minimize them as possible, the feature-oriented reference system is of great importance, and a configuration optimization method is proposed in this paper. The detailed analyses with regard to structures, inspections, and tolerances of assembly features are conducted, and the complex constraints are established. By applying them and considering the direct connections between MCSs, the modified configuration model is built. The proposed optimization method is based on the improved binary particle swarm optimization (IBPSO), which involves a two-stage strategy and a novel mutation. The mutation is developed by using feasible centers to attract and correct infeasible particles, and simultaneously, to maintain the particle diversity. The performed experiments show that the method can effectively output optimal positions of ERS points, and the reference system is eventually hybrid on the premise of meeting accuracy requirements. The increase of ERS points is beneficial, but no further optimization happens when the amount reaches the upper limit. The limits are 12 and 18 when the measurement instruments are located at two and three different stations respectively.

Adaptive Spatial Augmented Reality for Industrial Site Assembly

Today, industrial assembly of large components is mostly organized as site assembly. The human-centered information provisioning for assembly tasks still bears significant optimization potential. The paper presents first results of the integration of a dynamic projection system with visual human position and gesture recognition. Work instructions and other relevant information is displayed relative to the humans’ position within a large workspace. A high-resolution RGB-camera as well as the real-time object recognition algorithm YOLOv3, have been integrated into a proof-of-concept demonstrator. An evaluation at the TU Wien Pilot Factory shows the potential in terms of efficiency, usability, and cognitive workload.

An Error Identification and Compensation Method of a 6-DoF Parallel Kinematic Machine

Kinematic Calibration is an effective and economical way to improve the accuracy of the six degree-of-freedom (DoF) parallel kinematic machine (PKM), named as Stewart platform, for the large component assembly in aviation or aerospace. The conventional online calibration requires a powerful and complicated control system, whereas the current offline calibration methods are not satisfactory in terms of the compromise between efficiency and accuracy. This paper proposes a semi-online calibration method in which the geometric errors are identified offline and compensated online. The geometric errors are inserted into the inverse kinematic model. Instead of formulating the linear mapping model between geometric errors and the pose error of moving platform, the error model is written as the function of geometric errors with respect to the actuation inputs. Hence, a nonlinear error model is obtained. Without worrying about the identifiability, the error identification equations are converted into an optimization problem and solved by the hybrid genetic algorithm (HGA). In the traditional offline compensation, the identified kinematic parameters are adopted to modify the nominal kinematic model, which is inconvenient when the control system is not transparent to the users. A new control block that calculating the equivalent actuation inputs from the identified errors is added to the control flow. The errors are compensated in an efficient manner. Simulations and experiments are implemented to validate the accuracy, efficiency and convenience of the proposed method. The results indicate that our approach improves position and orientation accuracy of the Stewart platform by 85.1% and 91.0%.

High-Efficiency Posture Prealignment Method for Large Component Assembly via iGPS and Laser Ranging

High-efficiency posture prealignment for large components is an important application in aircraft manufacturing. In this article, a new prealignment method for large components via indoor global positioning and laser ranging is proposed. Based on the smart automated guided vehicle (AGV) system, the component is transported from the previous processing station to the assembly station under the guidance of an indoor global positioning system (iGPS), and a predictive correction algorithm is proposed to solve the potential drift problem. The subsequent close-range prealignment is achieved with the proposed laser alignment method, and to improve the accuracy, the estimation of posture error is presented in detail, along with the layout optimization of laser displacement sensors (LDSs). The analysis indicates that the symmetric form with two LDSs on each edge is the most desirable layout. A prototype system is developed, and the verification experiment shows that the posture error is no more than 0.3 mm in translation and 0.02° in orientation. Besides the outstanding performance in precision and efficiency, it possesses great repeatability and applicability.

One-Fits-All vs. Tailor-Made: User-Centered Workstations for Field Assembly with an Application in Aircraft Parts Manufacturing

Field assembly of large components for aircraft, ships, trains or big machinery is often done in a rather unstructured environment, with little concern for the individual needs of the involved personnel. Confronted with a shortage of qualified personnel and a generally aging workforce, workstation designers need to put more emphasis on improving ergonomics and allowing for individual adjustments and support for a particular worker. Human factor engineering traditionally uses a “percentile-approach” to work-design, preventing an exact adjustment to the ergonomic needs of particular users. New technological advances provide opportunities for major improvements; novel assistance and information systems allow the accommodation of the needs of an increasingly diverse assembly workforce, while minimizing overall physical and mental stress.Based on a recently developed synopsis of options for individualization of assembly workstations, an example of an application in field assembly is being presented and illustrated by a full-scale demonstrator for the assembly of large scale composite parts for the aerospace industry. The pilot-installation combines a projector-based augmented reality environment for dynamic worker-information with automated operator position tracking and field of vision positioning. Finally, a comprehensive concept, that includes other dimensions of individualization, is presented.

Model Predictive Control Approach for assembling Large Components in Motion

The assembly of large components is characterized by a large number of varieties and a high process influence by environmental factors. For this reason the assembly in the field of truck manufacturing is in most cases carried out as a moving line with a high percentage of manual processes. Automation of these assembly process, albeit desired from a quality viewpoint, is presently limited as it requires the workpiece to be stationary. The ability to keep the workpiece on the moving assembly line while keeping the used handling systems flexible can enable business benefits. By assembling for example the windshield of a truck in motion, lead times can be shortened and floor space requirements can be reduced while at the same time the quality can be increased. To enable assembly in motion of large components, it must be possible to synchronize the movement between the handling system and the moving workpiece placed on an unguided transport system. Within this article we describe a Model Predictive Control Approach for robotic assembly of large components on moving assembly lines. One approach to synchronize the motion between product and robot is a direct metrological reference between them. This can be done by using an iGPS metrology system in which every handling system can be referenced to every work piece throughout the whole production line. The measurement system must be able to detect sufficiently small movements to ensure a reliable synchronization and thereby a high quality of the assembly process. The interaction of several systems makes for a number of restrictions. Here especially the delay times of the measuring system and the preset position for the robot have to be taken into consideration. To control the assembly process, a model predictive control loop is necessary. Only this control approach can meet the restrictions which are strictly necessary for the assembly. This article shows the first results and describes a demonstrator which has been set up to mount a windshield on a moving truck cabin with the help of a robot.

Investigation of the Applicability of an iGPS Metrology System to Automate the Assembly in Motion of a Truck Cabin

The assembly of large components is characterized by manual processes and customer-specified products. These are just two of the challenges for the efficient automation of assembly processes. In the field of large component assembly, the goal is to synchronize between the handling system and the moving work piece placed on an unguided transport system. One approach to solve this problem is a direct reference between the product and the assembled work piece. This referencing is done by using a global reference system in which every handling system can be referenced to every work piece throughout the whole production line. This article will show the first results using a global reference to track large components on an unguided conveyor. Moreover, it investigates the question if the metrology system, which was used here, is able to detect even small position deviations on higher frequencies.

Research on Mechanical Manufacturing with Posture Optimization of Installation of the Bolts in Large Component Assembly

To improve the assembly efficiency, an optimization model is proposed to achieve the optimal posture of the component of the bolts installation. The component is measured and aligned by an automatic system, which consists of some NC locators and coordinate measuring device. Based on the measurement data of the bolt hole and the end surface, the maximum diameter of the bolt that can pass through the joint hole is obtained by using morphology theory. The particle swarm optimization method is adopted to solve this problem, and the stability and efficiency of the calculated results are improved by controlling the initial distribution of particles. The numerical simulation experiment shows that the proposed method would be helpful to guide the alignment and docking of large components.

A Novel Algorithm of Posture Best Fit based on Key Characteristics for Large Components Assembly

Measurement and variation control of geometrical Key Characteristics (KCs), such as flatness and gap of joint faces, coaxiality of cabin sections, is the crucial issue in large components assembly from the aerospace industry. Aiming to control geometrical KCs and to attain the best fit of posture, an optimization algorithm based on KCs for large components assembly is proposed. This approach regards the posture best fit, which is a key activity in Measurement Aided Assembly (MAA), as a two-phase optimal problem. In the first phase, the global measurement coordinate system of digital model and shop floor is unified with minimum error based on singular value decomposition, and the current posture of components being assembly is optimally solved in terms of minimum variation of all reference points. In the second phase, the best posture of the movable component is optimally determined by minimizing multiple KCs’ variation with the constraints that every KC respectively conforms to its product specification. The optimal models and the process procedures for these two-phase optimal problems based on Particle Swarm Optimization (PSO) are proposed. In each model, every posture to be calculated is modeled as a 6 dimensional particle (three movement and three rotation parameters). Finally, an example that two cabin sections of satellite mainframe structure are being assembled is selected to verify the effectiveness of the proposed approach, models and algorithms. The experiment result shows the approach is promising and will provide a foundation for further study and application.

Template Based Interference Detection Technology of Large Components Assembly

In order to make quantitative analysis of the interference situation during insertion and fit of large components,a template based interference detection technology is proposed. The template is customized according to the CAD model of insertion and fit characteristics. After match with scanning data,the template is taken as the reconstruction model to take part in the interference analysis. The insertion characteristic is discretized into a series of parallel sections perpendicular to the assembling direction,and the penetration depth,clearance and interference demarcation line of each section relative to the fit characteristic in the process of motion is calculated. The integration of interference situations of all sections constitutes the interference situation of insertion and fit characteristics. This method simplifies the 3D intersection problem to a 2D one,and not only can calculate quantitatively the penetration depth and clearance but also can give the interference distribution of assembly parts. Based on this,the handling of follow-up assembly is guided quantitatively. Finally,through the simulated interference detection and analysis of a pair of insertion and fit characteristics,the interference results including maximum interference quantity and distribution are. given,which verifies the feasibility of the proposed interference detection technology.

Posture Following Fixture Method in Large Components Flexible Assembly

How to keep and fasten reliably the adjusted posture is always a key problem in large components assembly such as airplane. Based on the phase-change principle of low-melting-point alloy,a new posture following fixture method of large components and a posture following fixture are presented,and it is proved feasible and reliable by a series of experiments and calculations. The method is described as follows:heat and melt the low-melting-point alloy in the cylindrical metal,insert the mandrel rigidly jointed with the large component into the melted alloy,and cool the metal to keep and fasten reliably the mandrel without posture change,and lock the inner core rigidly connected with the cylindrical metal,reheat and remelt the low-melting- point alloy again to release the mandrel and to resume the adjustable posture of the large component.