Bayesian Inference of Visco-Elastic Visco-Plastic Material Model Parameters for SLS-printed polyamide lattices

The response of polymeric materials can be represented using complex finite-strain visco-elastic visco-plastic material models. Such a model requires the identification of tens of parameters in order to remain accurate for a wide range of strain rates and stress states, the response in compression being different than in tension. A complex experimental campaign involving dynamic mechanical analysis (DMA), and compressive and tensile cyclic loading at different strain rates is thus required.
Besides, when considering lattice structures obtained by additive manufacturing, the struts response is not similar to the macro-bulk material response. Because a complex experimental campaign cannot be conducted at the level of the struts, the parameters identification also needs to be conducted at the level of the lattice response.
However, when performing the parameters identification using these different loading cases, a unique set of parameters cannot usually reproduce all the experimental tests because of the model limitations and errors, in particular when considering nonlinear responses. Besides, the data are inevitably entailed by experimental errors. These difficulties can be circumvented by considering a Bayesian Inference (BI) process. In this presentation we consider experimental tests conducted at different scales on polyamide lattices in order to infer the model parameters of a complex finite-strain visco-elastic visco-plastic material model.