Session: 01-12-02: Spacecraft Structures 2

Paper Number: 155538

155538 - Shape Errors of Modular Lattice Structures During and After Assembly 

With growing demand for larger structures in space, in-space assembly has become a promising approach. These structures, such as solar arrays, space-based telescope reflectors, or habitats, require precision in the assembled shape to maintain functionality.

NASA’s ARMADAS effort has made significant progress in designing the hardware and software for in-space assembled structures. The ARMADAS approach is to assemble identical frame-based building blocks called “voxels” into large lattice-like structures using simple robots. While structural stiffness, material strength and assembly algorithms have been extensively studied, assembly shape errors caused by manufacturing imperfections remain an ongoing research challenge. We address the prediction of shape errors of the structure during and after assembly due to manufacturing shape errors in the individual voxels.

We present a novel finite element analysis (FEA) simulation procedure designed to predict shape errors during voxel assembly. Unlike previous methods that focus solely on the final structure configuration, the proposed approach captures the deformation and shape evolution at each step of the assembly process. The methodology incorporates a stepwise simulation procedure, in which intermediate assembly stages are modeled sequentially. Experiments were conducted to verify the simulation, but due to measurement noise, the results were inconclusive. Instead, the validity of the simulation was confirmed by comparing it to a high-fidelity model.

The goal of this approach is to predict shape errors of the assembled structure at intermediate states during assembly. Abaqus FEA is used throughout. Geometrically imperfect voxels with random nodal position errors are automatically generated. Voxel struts are modeled as beams with linear elastic material properties. To simulate intermediate states of assembly, a complete model including voxels, connectors, and boundary conditions is constructed in the initial step of the simulation. The process is divided into three steps

1. The removal step uses model change interaction to hide all the connectors and parts except for the voxels to be assembled.

2. In the displacement step, the voxels are moved to a predetermined location where the connecting nodes coincide with each other.

3. In the equilibrium step, the connectors between the two connecting voxels are reinstated using the model change interaction, the displacement boundary conditions are deactivated, and the structure equilibrates as forces from shape mismatches are redistributed across the voxels' structural members.

The displacement and equilibrium steps are repeated for sequential voxel assembly. This approach captures elastic stresses in voxels during assembly, including self-stress, and the resulting elastic strains and deformations.

To validate the simulation approach, experiments were conducted. The as-manufactured shape errors of individual voxels were measured. These voxels were assembled into larger structures, and the resulting shape errors of the structures were measured. Measurements were taken using a non-contact laser line scanner, which showed pre-assembly errors on the order of 0.1 mm. However, measurement noise was also of a similar magnitude, which made the experimental validation inconclusive.

As an alternative approach to verify the simulation, a high-fidelity model was developed. The high-fidelity model uses solid elements, beam connectors, and contacts between voxels to achieve more realistic results. The comparison between the high-fidelity model and the simplified model was conducted by simulating a two-voxel assembly. Nodal coordinates were compared and showed a deviation of less than 10 microns with a 1-millimeter manufacturing error.

The proposed simulation approach successfully models common structural configurations, such as one-dimensional cantilever beams, two-dimensional plates, and ring formations.

In conclusion, this study presents an efficient method for simulating the assembly process of modular space structures, capturing both intermediate states and final shape errors. This approach provides critical insights for ensuring functional integrity in space structures and can aid in the optimization of assembly processes to minimize shape errors and material failures.

Presenting Author: Szu-Jui Huang Stanford University

Presenting Author Biography: Szu-Jui Huang is a Masters student in the Department of Aeronautics and Astronautics at Stanford University. He conducts research in the Morphing Space Structures Lab. He received his Bachelors in Civil Engineering from the National Taiwan University.