Meta-heuristic design of a light-weight homologous backup structure of the primary reflector for the Large Submillimeter Telescope

TL;DR

This paper presents a genetic algorithm-based structural optimization for lightweight, high-precision backup structures of large submillimeter telescope reflectors, achieving significant weight reduction and surface accuracy improvements.

Contribution

It introduces a novel GA-based multi-objective optimization method for designing homologous backup structures with minimized weight and actuator stroke, considering both axisymmetric and non-axisymmetric configurations.

Findings

Optimized structures achieved surface errors of ~5 μm RMS.

Significant reduction in backup structure weight.

Homologous deformation maintained with minimal active control.

Abstract

The development of large-aperture submillimeter telescopes, such as the Large Submillimeter Telescope (LST) and the Atacama Large Aperture Submillimeter Telescope (AtLAST), is essential to overcome the limitations of current observational capabilities in submillimeter astronomy. These telescopes face challenges related to maintaining high surface accuracy of the main reflector while minimizing the weight of the telescope structure. This study introduces a genetic algorithm (GA)-based structural optimization, previously applied in related works, to 50 m-class backup structures (BUSes) with a variable focal position, addressing the challenge of achieving both lightweight construction and high surface accuracy through the consideration of homologous deformation. We model the BUS as a truss structure and perform multi-objective optimization using a GA. The optimization process considers two structures: axisymmetric and non-axisymmetric between the top and bottom. The optimization aims to find structures that simultaneously minimize the maximum stroke length of actuators and the mass of the BUS under practical constraints. The optimized structures show improved surface accuracy, primarily due to the minimization of the maximum actuator stroke length, and reduced weight, both achieved under the imposed constraints. Notably, we find a homologous BUS solution that achieves a surface error of down to $\sim 5\,\mu\mathrm{m}$ RMS with a tiny portion of the truss nodes being actively controlled. The results highlight the potential of GA-based optimization in the design of next-generation submillimeter telescopes, suggesting that further exploration of non-axisymmetric structures could yield even more effective solutions. Our findings support the application of advanced optimization techniques to achieve high-performance and cost-effective telescope designs.