BESO topology optimization iterations.
This project develops a BESO-based topology optimization framework for shape memory alloy superelastic structures and actuators. Unlike conventional compliance minimization approaches rooted in linear elasticity, this framework employs a path-dependent transformation energy criterion as the sensitivity measure, inspired by the actuation work and fatigue life frameworks of Wheeler et al. and Chemisky et al. The element-level transformation energy is computed by integrating the conjugate Mises equivalent stress with respect to the conjugate Mises transformation strain over the complete thermo-mechanical loading path, capturing the cumulative energy output during martensitic phase transformation. The global objective maximizes volume-specific actuation energy density subject to a prescribed volume fraction constraint. A stress-controlled adaptive force strategy maintains bounded stress levels across BESO iterations, ensuring realistic isobaric operating conditions for actuator designs. The framework is demonstrated on superelastic structures at constant temperature, thermally driven actuators, and three-dimensional geometries, consistently yielding topologies with superior energy densities compared to conventional approaches. Ongoing work focuses on training neural networks to predict the transformation energy parameter, aiming to significantly accelerate the optimization process.
Graduate Student: Sefa Oksuz