Summerside Electric using artificial intelligence to predict city's energy needs

Summerside Electric is using artificial intelligence to better predict the municipality's energy needs. 

The utility first began testing the software, BluWave-ai, about two years ago. The software uses algorithms to determine energy needs. 

It went live with the system on Nov. 1.

The hope is that the technology will help increase the utility's efficiency in determining how much energy customers need, said Greg Gaudet, Summerside's director of municipal services.  

"For us it's replacing a software that we've been using for the last, close to 12 years, in forecasting our needs for energy and how we purchase it in the open market," Gaudet said.  

Forecasting energy in advance

Gaudet said the utility has to be able to forecast how much energy customers in the city will need one hour in advance.

To determine that, officials have to calculate how much energy can be generated from the town's renewable energy sources like wind power, and how much needs to be purchased in addition to that. 

When you're more efficient you're saving on greenhouse gas emissions.— Greg Gaudet, director of municipal services

"A lot of times those decisions, although we make them as best we can, they have some accuracy errors in them," he said. 

"And sometimes we would buy too much energy and then have to spill it back into the grid for basically a loss of revenue."

Over the past two years, the utility has provided data that has helped make the software more accurate. From there, real-time testing was conducted to ensure that the results were accurate. 

'Being more efficient'

While it's too early to gauge how much the software will impact the utility's ability to predict energy needs, Gaudet said he's hoping the utility will be able to improve its forecast accuracy by about 20 per cent.

"If you're more accurate, then you do have the effect of being more efficient and when you're more efficient you're saving on greenhouse gas emissions," he said. 

He said it's too early to tell if electricity rates will change as a result of the technology. 

Over the next year, officials will be monitoring the software closely to determine how effective it is, Gaudet said.

More P.E.I. news