Smart engineering: Technological innovations improve efficiencies and teach young engineers valuable skills
Wearables, thermostats and virtual assistants—the kind of smart devices we’re familiar with today—are just ripples on the surface compared to the tidal wave to come.
By 2020, the research company Gartner predicts there will be 20 billion internet-connected “things,” including jet engines and cars. Smart technology—which incorporates sensors into machinery and parts to collect data then analyze, share and use that data in real time—is fast becoming an essential tool in many industries, including manufacturing, transportation, energy, construction, health care, agriculture and more.
This means up-and-coming engineers need to understand how to incorporate sensors and data analytics into their designs. At Mines, faculty and students are working on a wide array of projects that not only give them hands-on experience with smart technology but help industries advance their knowledge and solve real-world problems.
Conserving resources with sensors and networks
Smart technology can help end users manage resources more efficiently and save money. A smart irrigation system—which Civil and Environmental Engineering Professor Junko Munakata Marr and Computer Science Professor Qi Han are working on under a National Science Foundation grant—shows how technology can make a difference.
Most landscape irrigation systems—the kind used for parks, golf courses and lawns—are operated manually or use a timer, which isn’t the most efficient. Studies show sensors that monitor existing soil moisture can reduce the amount of water used in sprinkler systems by 61 percent. But existing sensor systems are often cumbersome to install, only manage soil saturation by preventing irrigation from occurring or collect information that requires landscape managers to read and compile the data then decide on an action to take. By the time they reach a decision, conditions may have changed.
Munakata Marr and Han are working to automate the decision-making process and direct irrigation controls in real time. They’ll also use “reclaimed water”—treated wastewater—in a way that won’t damage plants.
The project requires multiple sensors connected into a wireless adhoc network, which Han is developing. The system is designed to irrigate only when necessary and will determine the correct proportion of reclaimed water to use. It’s a delicate balance, because reclaimed water is saltier than tap water, and too much salt kills plants. “There are a lot of moving parts that need to communicate smoothly with one another. It’s more complicated than it sounds,” Munakata Marr said.
One complicated step is translating landscape data into actionable information. Civil and environmental engineering graduate student Max Weiss and computer science undergraduate student Jordan Newport are working with Munakata Marr and Han on that part of the project.
To make a proper irrigation decision that balances the need for water conservation while also managing soil salinity, sensors must be able to report moisture and salinity simultaneously. Wireless communications will transmit data to a central location. There, a decision is made about when and how much water is applied to the landscape, sending a command to an irrigation valve to turn water on or off. The goal is to make sure the soil is not too wet, salty or dry, while conserving water and protecting plant health. When salinity gets too high, more water is needed to flush salt out of the soil. However, that requires using additional water, so the “rules of the game” must be determined to solve the optimization problem.
“It’s a big learning curve, but it’s really exciting,” Weiss said. “Knowing how to code and translate data into something useful can only benefit engineers.”
Shifting into overdrive
Below the surface, tunnel boring machines used for roads, subways and water systems are big, heavy and slow, usually advancing just a few inches a minute. Mike Mooney, the Grewcock University Endowed Chair and director of Mines’ Center for Underground Construction and Tunneling, is working with colleagues to speed up the machines and make them smarter.
“We’re using AI techniques to learn from data how to operate the machine better. The idea is to make it move faster, be more productive, avoid stoppages and prevent damage to nearby buildings,” Mooney said.
Data from sensors measuring pressures, forces, movements and vibrations as the machines work provide valuable information and suggest optimal settings in real time. In Seattle, Mooney and his crew installed a monitoring system that warned operators of impending boulders, allowing them to make adjustments and avoid damaging expensive equipment.
Petroleum Engineering Associate Professors Bill Eustes and Jorge Sampaio, Chemical Engineering Assistant Professor Joe Samaniuk, Geophysics Research Professor Rich Krahenbuhl and Mooney are collaborating on a DARPA initiative that aims to dramatically improve the speed for building small-diameter tunnels. Normally, it takes about 24 hours to construct a small-diameter tunnel 500 meters long, but working with key industry collaborators, the Mines Rapid Tunneling Technology Team thinks they can get that down to 83 minutes or less.
Some of the improvement will come from using smart sensor data to make adjustments on the fly. “Our research has shown we can improve drilling and tunneling speed by 30 percent to 100 percent by using AI,” Mooney said.
Mooney believes intelligent drilling will inevitably become an industry standard. “To be cost effective, industry will push the envelope as far as technology will let them. If one contractor does it, the rest will have to adapt to be competitive,” he said.
Taking smart tech to the Moon
Smart drilling techniques could also be used in outer space, and Mines researchers are already working on extraterrestrial applications.
Scientists now know the Moon contains water, which could be used both for drinking and to create fuel for further space exploration. But first, you have to find the water and figure out how to get to it, said Jamal Rostami, Haddon/ Alacer Gold Endowed Chair in Mining Engineering and director of the Earth Mechanics Institute.
NASA’s current plans involve taking samples of the lunar surface and sending them to Earth to determine whether a site contains enough water to make drilling worthwhile. Rostami and his team at Mines have a better idea: analyzing lunar soil, or regolith, to determine drilling parameters that will allow the system to reveal a site’s water content in real time.
“While you’re drilling through a sample, AI and machine learning algorithms will analyze the operational parameters of the drill, and from this data, they will conclude what type of material you’re cutting,” Rostami said. “All we need is data from the drill. If we can do on-site measurement and verification, it would speed up the work by months, if not years.”
In addition to looking for water, NASA plans to collect data and images from multiple space vehicles as they land on the Moon or circle around it. But transmitting the data from each vehicle would be a slow process riddled with omissions and duplications.
Computer Science’s Qi Han is working with Mechanical Engineering Research Assistant Professor Christopher Dreyer and robotic researchers at NASA’s Jet Propulsion Laboratory to develop an inter-spacecraft wireless communication network. The network would let rovers, landers and other vehicles exchange and organize their information, then feed it to a single carrier with high storage capacity and computing power. The carrier would transmit data to Earth.
“It’s about communication and control and making sure spacecraft are covering all the locations they’re supposed to cover,” Han said. “The robots could communicate with each other and adjust their decisions in real time.”
Using sensors and data wisely
From urban underground to outer space, Mines faculty and students are finding innovative ways to incorporate smart technology into engineering decisions. Working their way through these technical problems is proving advantageous for young engineers, providing students with skills in data management and coding that are integral to tomorrow’s technologies and valuable to employers.
Mines’ emphasis on problem-solving ability, working under pressure and creativity enables students to pick up smart technology skills quickly. Even more important, the foundation Mines provides helps shape their decisions about when and how to use such technology.
“Smart technology can be quite dumb if it isn’t used right,” Mooney said. “We need students with strong fundamentals who know how to use technology and judge whether it’s doing what it’s supposed to do.”
Technical knowledge, sound judgment, quick problem-solving ability and creativity all came together last fall for Sumner Evans MS ’19, who was part of a Mines team that beat 20 competitors from across the globe to win a Facebook hackathon.
The team developed an Android app that creates an interactive indoor map from a photo of a floorplan using augmented reality technology. They created the app in just 24 hours on a software platform only one of them had any experience with.
With just an hour left in the competition, the team tested their app only to find it didn’t work. They managed to fix the software bug just as the event organizer was counting down the last seconds.
Their Mines education gave the team the foundation it needed to win, Evans said. “What Mines provides isn’t so much experience with a particular smart technology,” he explained. “Technology comes and goes. What Mines does is prepare students to move into any new technology and contribute.”