This Mexico City office building didn't collapse until 50 minutes after a 2017 quake—after people had gone back inside.


Is your building safe after an earthquake? These cheap sensors could tell you

JANTETELCO, MEXICO—On 19 September 2017 at 1:14 p.m., the ground lurched under Mariano Matamoros Elementary School here. The earth had ruptured just 40 kilometers away, at the epicenter of a 7.1-magnitude earthquake that killed nearly 400 people across central Mexico. Adobe buildings, common in this small town, cracked and crumbled. At the school, a second-floor landing tilted precariously above a cracked concrete column. All the students and teachers evacuated safely, and the building was repaired, but for months, young students feared returning to second-floor classrooms, says Principal Casimiro Enríquez Vergara. "Any loud sound we hear now, we think it's an earthquake," he says.

Now, this school has become one of the first in Mexico to be equipped with a cheap sensor system that, in a future earthquake, will monitor shaking and automatically assess whether damage has occurred. The ideas behind the system, called Pulse and sold by the Mexico City–based company Grillo, are not new. Earthquake engineers have long put sensors in large, critical structures such as bridges and skyscrapers, so they can look for clues to hidden, deadly damage after a quake. The plummeting cost of sensors and the cloud computing needed to process all the data is allowing researchers in both the public and private sectors to deploy the sensors in many parts of the world. Scientists hope the systems can save lives and help prioritize repairs in the wake of deadly earthquakes.

"It's been a dream of the engineering community for a long time," says Thomas Heaton, a civil engineer and geophysicist at the California Institute of Technology (Caltech) in Pasadena. He compares the field, known as structural health monitoring, to "taking a building's blood pressure" and hopes for a day when all buildings in seismically risky regions are continuously monitored. Thanks to the newly affordable equipment, "it's really feasible" now, Heaton says.

In recent years, small accelerometers—devices that precisely measure movement in 3D space over time—have gotten better and cheaper. Used in smartphones, cars, and video game controllers, the accelerometers are now sensitive enough to detect faraway tremors and subtle building vibrations.

This stoplight, when installed outside a building, can warn if it's not safe to go inside.


Researchers at Grillo say installing Pulse sensors on each floor of a building lets them measure a crucial indicator of structural health, a ratio called interstory drift—how much floors suddenly jerk out of alignment during a quake. "Buildings can withstand some interstory drift," says Monica Kohler, a civil engineer and seismologist at Caltech. But beyond a certain threshold that varies by building, they become permanently damaged and can even collapse, she says. To determine damage, these systems compare a building's interstory drift to limits determined by computer models—ideally crafted from structural drawings for that particular building.

That's not realistic for individual homes or schools in a developing country such as Mexico. Instead, Grillo uses standard blueprints for various building types. The concrete-frame, two-story elementary school here, for example, is "as typical as you're going to get in Mexico," says Andrés Meira, an architect and the founder and CEO of Grillo. "Unless there's some trickiness in how they constructed it, we should have a pretty good idea of how it should move." Grillo has declined to install Pulse in buildings that don't follow standard designs.

Sticking to standardized models helps keep costs down. Grillo and similar systems like P-Alert, made by the company San Lien Technology in Taipei and deployed in buildings in Taiwan and the Philippines, cost $1000 or less, as opposed to tens of thousands of dollars for traditional sensor networks used by seismologists at government agencies and specialized building monitoring companies.

On a recent day, Meira and the Grillo team mount orange plastic boxes containing Pulse sensors on classroom walls on each floor of the school here and connect them with ethernet cables. Then, they bolt a thin cylindrical stoplight to the outside of the building. After a quake, if the green light goes on, the building is very likely safe, they say. Red or orange lights mean damage or possible damage, respectively, and no one should enter the building until engineers can perform an inspection, says Luis Rodríguez Abreu, a seismologist at Grillo.

Pulse doesn't replace those expert inspections, but should serve as a tool that experts can use to prioritize their work in the chaotic hours and days after a strong quake, he says. It can also warn people of hidden damage, which can be deadly. For example, 50 minutes after the shaking stopped in Mexico's 2017 earthquake, a four-story office building in Mexico City collapsed—after some people had returned to work inside.

"If you have sensors installed, you can quickly identify damaged and not-damaged buildings, and focus on the potentially damaged buildings," agrees Yih-Min Wu, a seismologist at National Taiwan University in Taipei who developed P-Alert. Equally as important, the systems can let people know their homes and offices are safe.

The low-cost systems "do make for a very good first pass" at determining which buildings are damaged, Kohler says. As part of Caltech's Community Seismic Network project, however, she has gone further, partnering with the Los Angeles Unified School District in California to install sensors calibrated to customized building models in about 300 public schools. She hopes to see a day when almost every building is monitored, and structural drawings are routinely available, so scientists can tailor models to each one.