Site icon Wonderful Engineering

Scientists Are Using Machine Learning To Try And Predict Earthquakes

New Type of Machine Learning Helps In Earthquake Risk Prediction

Buildings are only as solid as the ground underneath them. However, that solid base can turn into liquid during an earthquake and collapse the whole building all at once.

Typically known as liquefaction, it was a key factor of the earthquake in 2011 at Christchurch, New Zealand. The earthquake measuring the magnitude of 6.3 destroyed various homes and took nearly 185 lives.

The media covered the unfortunate event on a massive scale as the city already installed various sensors for observing the earthquakes. However, post-event inspection provided further data on the condition of the soil.

“It’s an enormous amount of data for our field, if we have thousands of data points, maybe we can find a trend,” said researcher Maria Giovanna Durante.

In collaboration with Ellen Rathje, an engineer at The University of Texas at Austin and principal investigator of the U.S. National Science Foundation-funded DesignSafe effort, Durante supports the natural hazards community research.

Rathje has been researching liquefaction and hopes to incorporate machine learning into her research; therefore, the Christchurch earthquake seemed an interesting event, to begin with.

In the Christchurch earthquake, a machine learning model developed by researchers predicted the movements that occurred when the earthquake instigated the soil to lose its robustness. The results were published in Earthquake Spectra.

“It’s one of the first machine learning studies in our area of geotechnical engineering,” Durante said.

According to Joy Pauschke, a program director at NSF’s Directorate for Engineering, this paradigm shift in data-sharing and teamwork is the core of DesignSafe and would help us in our goals regarding field progression.

“Researchers are beginning to use AI methods with natural hazards research data, with exciting results,” Pauschke said. “Adding machine learning tools to DesignSafe’s data and other resources will lead to new insights and help speed advances that can improve disaster resilience.”

The Texas Advanced Computing Center has partnered in the DesignSafe project, providing computing resources and software to the natural hazards engineering community. TACC’s Frontera supercomputer was put to work to test the model due to its speed and efficiency.

Exit mobile version