This work presents a GPU-accelerated numerical framework based on the Lattice Boltzmann method (LBM) coupled with the immersed boundary method (IBM) for simulating microscale rotating structures. The framework is developed to support the design and optimization of MEMS-based actuators where accurate modelling of fluid–structure interactions is critical. High-resolution simulations are enabled through GPU parallelization, capturing unsteady flow behaviour with enhanced spatial and temporal fidelity. A Smagorinsky-based large eddy simulation (LES) model is incorporated to resolve turbulence effects at high Reynolds numbers (
. The model is validated using canonical benchmarks including a simple lid-driven-cavity flow and flow past a circular cylinder showing good agreement with reference velocity profiles and drag coefficients. MEMS-oriented case studies are then presented comprising a micro-cantilever embedded in a lid-driven cavity and a microscale rotating rotor to evaluate confined flow-structure interactions, hydrodynamic loading and torque generation. A case study involving a microscale rotor evaluates key actuation metrics including torque generation, pseudo-power coefficient and flow structures such as vortex shedding and tip-induced turbulence. The results demonstrate the framework's capability to resolve complex flow physics relevant to micro-actuation while maintaining computational efficiency. This approach is particularly applicable to MEMS, biomedical microdevices and microfluidic actuators where multiphysics coupling and accuracy at small scales is essential.