A roadside scan of a structure is combined with location and environmental data and run through a computer program that identifies the types and locations of all the energy leaks. A heat flow model calculates the amount of energy loss, and — using the estimated efficiency of the building’s heating and cooling system — that energy loss is then converted to dollars.
The system was developed off of MIT Mechanical Engineering Professor Sanjay Sarma’s concept of “negawatt mining,” which is the idea of recapturing energy loss from buildings.
Buildings lose about 25 percent of input energy due to various inefficiencies. Long Phan, a doctoral candidate at MIT who worked on the research and development, said the U.S. spends $400 billion per year on energy inputted into residential buildings and loses about $100 billion of that due to leaks.
Theoretically, by correcting the leaks, that money should be recouped. In addition, the recaptured energy — the wasted watts — essentially constitutes a new power supply.
“Imagine Google Street View with the ability to drive around a city and capture all the long-wave infrared images in that city and being able to map all that energy loss onto a map, therefore having a well defined carbon or energy leak map of that city,” Phan explained.
The prototype system consisted of one long-wave infrared imager containing a GPS locator and various sensors, all mounted on a car. The system allowed MIT students to geo-tag each image they captured. Phan said the idea was to turn handheld infrared scanning from an ad hoc process to a systematic method that captures all the heat loss information from a structure.
The thermal images and data are examined using a computer vision algorithm that automatically finds leaks, targets them and — based on the size, shape and texture of the leaks — classifies what type of leak it is, such as a door or window. The program uses a prototype library to compare each result to classify the leak.
The model was used successfully in various neighborhoods in Cambridge, Mass., Ft. Drum in New York and a number of other locations, according to Phan. The final system, which Phan said is in the final phase of production, is made up of a camera array of 14 high-resolution imaging units. The units will also be able to be mass produced.
Jonathan Jesneck, a research scientist in the MIT Field Intelligence Lab, who along with Phan created the multiphase process of geospatial data integration and scalable prediction analysis that the drive-by thermal imaging system uses, said the program can identify something as trivial as a window that’s leaking and costing $15 in monthly energy costs. With a little caulk, that could be reduced to $8 per month.
“[We] build up a database of how expensive each leak is and have an estimate on how expensive it would be to fix each one, so you can do a financial analysis to figure out the return on investment of fixing each leak,” Jesneck said. “You’ll know exactly where to put your money for the biggest bang for the buck.”
The system has been so successful that Phan and Jesneck, along with various colleagues, have started a spinoff company from their work at MIT. Called Eye-R Systems, the company is mass producing the scanning technology, which is called Energy Diagnostic for Global Efficiency. Phan is president and CEO of Eye-R Systems, while Jesneck is vice president of research and product development.
The business also has its first customer — the U.S. Department of Defense (DoD). Eye-R Systems and MIT will be working with the U.S. Army Engineer Research and Development Center to demonstrate its technology as a tool for the DoD to make better decisions regarding building design and retrofit projects.
Ultimately, however, Phan said his company’s major goal is to develop a national energy database using the technology.
“Imagine every city in the U.S. mapped onto a national energy database that will allow customers to log in and generate energy reports of their home,” Phan said. “That report will come in the form of several criteria [and] have an energy efficiency rating score associated with that house.”
