A trivariate T-spline based framework for modeling heterogeneous solids

•A robust framework to perform material modeling suitable for continuously varying material with local control.•The optimization process consists of two alternately executed procedures, progressive fitting and octree-based adaptive refinement.•The goal of progressive fitting procedure is to dynamically update the control points by minimizing the energy functional.•The comparison between trivariate T-spline with adaptive refinement and trivariate NURBS with uniform refinement demonstrates the computational efficiency.