Session: 02-03: Shape Memory Alloys

Paper Number: 91582

91582 - Shape Memory Alloy Rendering of Experimental Analysis and Calibration Tool 

The fundamental understanding and industrial application of shape memory alloys (SMAs) continues to grow. However, a laborious workflow of experimental analysis and constitutive model calibration is required for rigorous SMA characterization, especially for newcomers to the field or those engaged in multi-disciplinary efforts. The various external state variables that govern shape memory material behavior (i.e., temperature, stress, strain) often require conglomeration of multiple instruments to properly record sufficient data and can result in inefficient use of time when synchronizing various datasets from different instruments. Further, when such data has been properly analyzed, there are myriad methods of arriving at an accurate constitutive model to describe the specific SMA material behavior. Thus, a streamlined open-source python code, named the SMA Rendering of Experimental Analysis and Calibration Tool (REACT), is herein described to help scientists and engineers working on SMAs analyze and calibrate their complex data sets.

Our tool provides an intuitive workflow that imports and processes raw unfiltered shape memory alloy mechanical (tensile/compression), thermal (DSC), or thermomechanical (tensile/compression with environmental chamber) data to produce customizable figures and systematically derived material data. This toolset can extract data from multiple inputs such as tensile test data and external thermocouples and automatically synchronize them onto the same time series. With raw force and displacement data, the SMA REACT can calculate strains and stresses based on various sample geometries. Coupling temperature, stress, and strain data, this tool can apply customizable filters and remove systematic errors within the dataset, periodically prompting the user for filter approval. The program then produces various figures to help visualize the complex shape memory alloy material behavior.

Furthermore, a model calibration routine is developed to find the best fit of constitutive model parameters (martensitic elastic modulus, austenite start temperature, etc.) given experimental data. Following the thermodynamically consistent model derived by Lagoudas, et al., the developed calibration routine leverages global optimization strategies to minimize error between model prediction and experimental data. The tool described herein enables the user to customize the optimization routine as well as the model parameters to be optimized (e.g., bounds and free variables). Outputs from the calibration routine include a set of model parameters to be used in future analyses (i.e., material properties for FEA) and a thermodynamically consistent phase diagram based on calibrated model parameters. The current workflow attempts to minimize tribal knowledge contained within the SMA constitutive modeling community by demystifying processes used for calibration. We hope this tool can provide an efficient workflow and salient guidance to others in the shape memory alloy community for years to come.

Presenting Author: Patrick Walgren Texas A&M University