A Machine Learning Model for Predicting Progressive Crack Extension based on DCPD Fatigue Data

TL;DR

This paper presents a BiLSTM neural network trained on DCPD fatigue data to predict crack growth progression across temperatures, capturing crack jumps and aiding understanding of damage mechanisms.

Contribution

The study introduces a BiLSTM model trained on high-frequency DCPD data to accurately predict crack extension during fatigue tests, including complex crack jumps.

Findings

Model accurately predicts crack growth at various temperatures.

Reproduces crack jumps and overall crack progression.

Demonstrates potential for studying damage mechanisms.

Abstract

Time history data collected from a Direct Current Potential Drop (DCPD) fatigue experiment at a range of temperatures was used to train a Bidirectional Long-Short Term Memory Neural Network (BiLSTM) model. The model was trained on high sampling rate experimental data from crack initiation up through the Paris regime. The BiLSTM model was able to predict the progressive crack extension at intermediate temperatures and stress intensities. The model was able to reproduce crack jumps and overall crack progression. The BiLSTM model demonstrated the potential to be used as a tool for future investigation into fundamental mechanisms such as high-temperature oxidation and new damage models.