Research On Key Technologies Of Assembly Process Optimization For The Double-Curved Antenna Sheet-Metals

Assembly precision control for the large-scale double-curved reflector is an essential topic that fulfils the antenna performance designed. This reflector is an assembly of many solid rivet-linked sheet-metals that are tighted based on the fixtures at the lie state and then set to the stand state. At assembly locale, the often occurred oversize RMS at the riveted state or the stand state always induces a number of loops of structural reassembly and RMS retest, or even the structure dismissal. Rapider process response and lower assembly cost require an accurate and rapid riveting-oriented computing method for the dimensional error reduction of the large antenna assembly with thousands rivets. Based on the analysis of the antenna assembly process, this dissertation adopts the kinematic formulation, static and dynamic finite element (FE) analyses, and the related numerical and analytical methods to find the proper modeling for the dimensional faults of the parts in the assembly and an accurate mapping that gathers the influence of the part locating error, self-weight, local riveting distortions and the lie-to-stand conversion into the global dimensional error in three-dimensional (3D) form. Thus this research yields a series of new technologies, such as the assembly chain decomposition method and the algebra for the system fault analysis, the 3D precision analysis method, the local-to-global dimensional error calculation method and the assembly process optimization method. After the discussion for the influence of error sources to the global error and the survey on the performance of assembly process optimization, this research gives born to a software tool that calculates the global dimensional error by setting the detailed process parameters and optimizes the process parameters. In global the main research fruits comprise the following points.1. The decomposition approach for the assembly and the algebra method for the system faults are proposed respectively based on the spatial distribution of the parts in an assembly and the key elements in the static and dynamic fault analysis. The expression of 3D motions of key points for the rigid and compliant motions of parts is unified firstly. According to the assembly sequence and the spactial dependency of the antenna parts, the decomposition approach for the antenna assembly is proposed that fullfils the rapid decomposition of complex assemblies. The logical union that is made up by events is considered as one variable. Each variable is represented by an array:(1) it has three elements; (2) each element, is a set; (3) the first set contains the Boolean state for this variable; (4) the second set contains the state timing; (5) the last set contains the probability at the starting moment for the beginning from the normal state to failure state; and (6) the Boolean state represents the variable value. The variable array interprets the logical union with two variables as the temporal or mixed operation using four operators:sequence, simultaneity, disjunction, and the AND operators. The algebra yields theorems including priority, idempotent, commutativity, associativity, absorptivity and distributivity. This algebra is propitious to the efficient determination and reduction of the minimal cut sets and the analytic solution of top event unreliability. Results of the simulations for the antenna assembly system and the other systems validate this algebra has advantages in the concise expression of fault in systems and the efficient reduction of the Boolean functions.2. The 3D precision analysis method that is based on the assembly chain is proposed by adropting the kinematic formulation and the static FE analysis method. The rigid motion and deformation relating to the locating errors and the gravity deformation satisfy a linear accumulation. The assembly is represented by the FE mesh. Numbering the classifications of nodes of the mesh according to the sequence of parts is used to distinguish different parts in assembly chains. To set the locating errors, three non-collinear nodes are selected from each last assembled part near the mating surface in the chain. By setting the locating errors, the kinematic derivations that are based on an approximation of velocity to micro motion yield the 3D rigid motion calculation method that can modify the coordinates of the related mesh nodes. Static FE method can compute the deformation relating to the locating errors and gravity, which is also used to modify the coordinates of the related mesh nodes. The 3D precision analysis method is integrated by the FE method and 3D rigid motion calculation method. Simulation and experiment for a parallel assembly validate that this 3D precision analysis method is accurate for the calculation of the global dimensional error that caused by the part gravity and locating errors.3. The local to global dimensional error calculation method is proposed based on the dynamic and static FE analysis method and the numerical interpolation. Based on the antenna riveting parameters, the dynamic FE method and the riveting experiments valdidate that the deformation distribution around the rivet hole can represent this riveting, which gives born to a new inherent strain database for the riveting. Then the local to global dimensional error calculation method is made true by the inherent strain data reloading and the iterative static FE analyses. At last the quantitative relation between the global dimensional error and the factors including the part gravity, the mating gaps and the rivet deformations can be integrated by 3D precision analysis method and local to global dimensional error calculation method. The experiment and simulation for a parallel assembly validate the calculation efficiency and precision of the method.4. The antenna assembly process optimization method is proposed based on the local to global dimensional error calculation method and related optimization algorithms. The overall assembly process and local riveting process for the large-scale antenna assembly with 1093 rivets are studied. Parallel and sequential optimization frameworks are integrated based on the classification of rivets and the local to global dimensional error calculation method by the integration of the genetic and ant colony algorithms that targets to reduce the global dimensional error (denoted by RMS). Simulations for the assembly with 1093 rivets validate the efficiency and feasibility of the sequential optimization framework. Results indicate that (1) both the rivet UDs and the Assembly Sequence (AS) are the main RMS influence factors, (2) the proposed method can efficiently optimize the specific process parameters for the large-scale assembly with abundant rivets and (3) the effective optimization prefers to solve rivet UDs and AS step by step. Finally the simulation and experiment for an assembly with 48 rivets validate the optimization effect.5. Both the influence from locating errors, structural gravity, riveting process and the change of constraints to the global dimensional error and the process optimization performance are investigated by the application of the above proposed methods to the parallel assembly, the assembly with 1093 rivets, etc. Two experimental tests are finished by the optimized and the traditional process parameters for the assembly with 1093 rivets. Two groups of dimensional changes between the tested errors before and after riveting indicate that the optimized process can eliminate more gap-induced RMS. Meanwhile the investigation on the results of the simulations and experiments in this thesis opens out:(1) the partial stack-up and reduction are common phenomenon under the influence of gaps, process parameters and gravity; (2) the gravity and constraint are the main factors for the error changes before and after the state conversion; (3) process optimization can be handled in the lie or stand state. Finally the rivet UD optimization and joint sequence optimization methods are respectively proposed. The case study of engineering application shows the method performance and effectiveness. Meanwhile the software interfaces and components are designed and developed.