Inverse design of stochastic, voxelated thermo-viscoelastic digital materials

Polymer material jetting enables the fabrication of voxelated, multi-material structures with material control at the microscale. However, current work often neglects viscoelastic effects and designing voxelated digital materials remains challenging due to the complexity of the vast design freedom and intractability of efficiently modeling macroscale voxel structures. We present an efficient representation of stochastically mixed, voxelated digital materials and develop a generalized viscoelastic temperature-dependent material model to design and simulate digital materials mixed from two constituent polymers. The material model is based on an extended percolation theory considering frequency and temperature. An artificial neural network is trained on the material model to directly estimate target material behavior given arbitrary non-linear, user requirements. The approach is validated using two case studies requiring tailored, non-linear material behavior: a personalized wrist orthosis and a machine damper. These show the newly unlocked possibilities for the design and fabrication of tuned, stochastic digital materials.