Error Modeling and Flexure-Based Calibration of Large-Aperture Optical Adjustment Mechanisms Using an Improved Particle Swarm Optimization Algorithm

In laser inertial confinement facilities, large-aperture parallel adjustment mechanisms face significant alignment errors and calibration challenges due to motion coupling and multi-source error interactions. This paper proposes a kinetostatic modeling and error calibration method based on an improved particle swarm optimization (PSO) algorithm. Compliance modeling and energy-based analysis are employed to equivalently treat structural clearances and geometric errors, establishing an error model with rigid-flexible decoupling. A global sensitivity index (GSI) is introduced to evaluate parameter influence and guide optimization. Furthermore, a global-dynamic multi-subpopulation collaborative PSO is developed to coordinate local refinement and global exploration, mitigating premature convergence and enhancing calibration accuracy. Based on the calibrated error model, adjustment compensation is implemented. Simulations and prototype experiments demonstrate that the method reduces average attitude errors by 72%–97% and maximum errors by 60%–96%, providing an effective approach for error compensation and parameter optimization in large-aperture compliant parallel mechanisms.