Many natural and social systems may exhibit abrupt, rare, very large events (e.g., here and the references therein). Examples are a material failure, bursting financial bubbles, epileptic seizures, volcano eruptions, power blackouts, etc. These phenomena typically occur unexpectedly and have catastrophic consequences for the system. However, before these events, precursor activity can be misinterpreted as random noise. Such precursor activity might be leveraged for predicting (and ideally preventing) the upcoming catastrophic events.

In the case of materials loaded well below their limit capabilities over extended periods, failure by damage accumulation and localization can occur in a phenomenon known as creep failure. If this happens, a material can fail catastrophically and unexpectedly, which is especially dangerous in civil structures holding loads for long periods, such as, e.g. bridges or damns. Creep failure has also been related to natural catastrophic events such as landslides, cliff collapses, and some earthquakes and volcanic eruptions. As shown in this work, it’s possible to find so-called precursors of failure, that is, statistical changes in a material’s plastic deformation activity that tells us that failure is coming. However, such statistical changes might occur abruptly, and they might not be clearly distinguishable from noise. The relevant question is, then, to which extent can the precursors be used to predict a possibly imminent catastrophic material failure? If possible, can we do it early enough to prevent it?

Plastic deformation exhibits a complex and stochastic behavior with correlations in space and time, and much information is encoded in such correlations. By applying classical statistical methods, it will be tough to exploit all that information. However, dealing with complex multidimensional and correlated data in an automated way is what machine learning excels at.

A material deforming under creep conditions, and eventually approaching failure. Left: strain-time curve; Right: spatial map of plastic strain.

In this paper by S. Biswas, D. F. Castellanos & M. Zaiser, Random Forest regression is used to learn and leverage spatio-temporal correlations in plastic deformation activity automatically. With this model, we estimated as a function of time the remaining time to fracture of many samples. The model is trained and evaluated with synthetic datasets produced by the stochastic model introduced in this other work. The first question we answered is whether predicting a material’s upcoming fracture is possible. This question is of great importance, with implications even for earthquake predictability.

Moreover, we studied how different material properties and conditions can affect failure predictability. For example, what’s the role of the material’s structural heterogeneity, the temperature, the loading conditions, or the material’s sample size? We found that the machine learning model can predict the approach to failure with higher accuracy than other, more classical methods proposed in the scientific literature. In addition, we found that the remaining time to failure is predictable with a different degree of accuracy that depends on the distance to such time, the material properties, and the loading conditions. Among those, the local strength heterogeneity is a crucial parameter in producing safer materials since it leads to a less abrupt approach to failure. Such behavior means that an upcoming failure can be identified easier and with higher anticipation, which is crucial for safety reasons. Lastly, we found that the bigger a material sample is, the more difficult it is to predict failure. Unfortunately, this might be bad news for earthquake predictability.

In summary, this proof-of-concept study suggests the possibility of training machine learning models on synthetic data to learn the statistical signatures of material failure. Furthermore, using synthetic data for model training can help overcome the scarcity of experimental data since, by definition, reaching failure implies the destruction of the object under study. A trained model could then be deployed in real-world settings to find early-warning signals of failure hidden in the deformation activity of a structure or the earth’s crust.