Session: 02-02: Performance of Shape Memory Alloys

Paper Number: 88882

88882 - Influence of the Phase Transformation Behaviour of NiTi Shape Memory Alloy Wires on the Predictability of Strain During Operation 

Predicting the stroke (respectively the strain) of binary NiTi shape memory alloy (SMA) wires is a challenging, but interesting task for designers of SMA components. Due to fatigue strain and the resulting stroke decreases with increasing cycle count. In the past several approaches were investigated, regarding this topic based on models resulting from the energy conversation equation. By heating the wire energy goes into the system and via heat content, convection, radiation, heat conduction, mechanical power (lifting a weight) and phase transformation energy leaves the system. All approaches use the same model to predict the phase transformation energy.

In 2004 Oelschläger proposed a model to calculate the resulting stroke of an SMA wire based on the electric energy. The presented work aims to extend Oelschläger’s model and takes the fatigue of the PTT into account, by monitoring parameters such as electric resistance, strain and energy consumption of the wire.

The leading question for this work is: How well does this model perform for SMA wires generating a full stroke within small temperature ranges? The developed model is tested based on 15 pre-cycled binary NiTi wires (4% strain, PTT between 120 °C and 125 °C, 300 MPa stress, diameter 0.25 mm).

The influence of the input parameters of the simulation model are investigated with a focus on geometric parameters that may deviate due to production influences. The stroke predicted by the model strongly depends on these parameters.

Finally, all results are described and discussed in detail. The fatigue tests are analyzed using statistical methods (distribution models, outlier test, fitting of data, computing of confidence intervals, etc.) in order to develop a model for the target range. The interval is computed at a designated confidence level (e.g. the 95 % confidence level). The confidence interval is also determined for the developed simulation model, so that the simulation model can furthermore be compared with the model of the target range (based on real data). If the simulation model is within the target range, it is suitable to consider the fatigue of the PTT in the prediction/ simulation of the stroke.

The target range model also allows further analyses with respect to anomaly detection, deviations from "set points", indications of impending failures, and degradation behavior predictions.

Presenting Author: Benedict Theren Ruhr-University Bochum