Green Power Labs of Dartmouth is celebrating its 10th year of delivering climate analytics to the power industry by launching an ambitious project that could reduce energy consumption in buildings by well over 10 per cent.

Founded in 2003, the company has developed an international clientele for its predictive analytics product SolarSatData for Utilities, which helps power suppliers determine future patterns of energy supply based on expected change of solar radiation.

The company most recently was awarded a $2.4-million loan from the Atlantic Canada Opportunities Agency’s Atlantic Innovation Fund to develop technology that can optimize energy consumption in commercial buildings.

“We find ourselves positioned at the leading edge of this technology and it really brings some serious promise,” said president Alexandre Pavlovski in an interview Tuesday. “We have lots of interest internationally from well-recognized property management companies and energy management companies.”

Pavlovski and his partners established Green Power Labs to help to determine how much solar power a jurisdiction or company can expect to produce, given the expected weather patterns. The company’s technology analyzes satellite images for cloud patterns and factors in such data as temperature, humidity, pollution to determine how much solar power is expected to be produced in the future, whether it’s the next minute, next day or the next year.

Its main clients so far are utilities, and it has offices in San Diego, Calif., and Adelaide, Australia, and is considering opening an office in Hong Kong.

The company can now apply its analysis to a square kilometre, but its funding from the innovation fund will help it hone that down to analyzing a square metre of a building’s surface. SolarSatData for Buildings will consider similar phenomena to the product for utilities. But it will also analyze how the sun’s beams hit each surface of a building at specific times of day, and whether the structure is in a city or surrounded by nature.

By predicting how much energy will be required, building managers can adjust heating settings and save power.

Marlene Moore, the vice-president of marketing and public relations, said Swiss research has shown that predictive analytics can improve a building’s energy efficiency by up to 41 per cent. Given that buildings account for about one-third of global energy consumption, the savings could be profound.

However, Pavlovski said the company will initially target improvements of 10 per cent.

Green Power Labs has partnered with leading universities in the region to help to develop the project. An Acadia University team headed by Daniel Silver will work on the machine learning and artificial intelligence aspects, while Lukas Swan of Dalhousie University will head a group working on building energy modelling.

“If I know it’s cold in the morning but it will be hot in the afternoon, there are ways to make (a building) more comfortable,” said Silver. “The X factor for Green Power is it will take into consideration those ambient conditions and the amount of energy hitting the building.”

The funding from the innovation fund is for 21/2 years and should allow Green Power to expand its 15-person workforce in Dartmouth. Pavlovski said the company expects to complete the technology development by the autumn of 2014 and a product on the market in late 2015.

The commercialization of the product will require more capital, and the company is now considering how best to raise it, said Pavlovski.