Language selection


Synthetic remote community energy load using machine learning techniques

Enabling the off-diesel transition for remote communites using digital solutions

As Canada’s remote communities move toward reducing and potentially eliminating their reliance on diesel fuel, there is a need to critically evaluate alternative energy opportunities with metrics regarding energy production, diesel fuel displacement and the costs of various fossil-fuel alternatives. The development of more robust, user-friendly software with AI-enabled optimization techniques would result in better-informed decision making and the transition of remote communities to sustainable energy solutions in a shorter timeframe.

Project objectives

Understanding the load profiles of Canada’s remote communities is an essential starting point for investigating alternative clean energy generation that will lead to transitioning away from diesel generation that currently powers these communities. This type of community load data is generally not publicly available. As a means to move forward with investigation of alternative technologies, CanmetEnergy-Ottawa (CE-O) is collaborating with the Digital Accelerator team to investigate using existing time series (e.g. the few available community load profiles, as well as wind, solar and weather data) to create typical time series for Canada’s remote communities through the application of modern data science techniques.

Expected results

This project is intended to produce synthetic electrical load profiles for remote communities in the form of hourly time series for a representative year. The outputs from the machine learning application would be hourly community electric load time series for the roughly 175 communities for which measured load data is not publicly available. There will be opportunities for future validation as and when measured load data becomes available.

Data sources

  • Metadata for 190 remote communities (e.g. lat/lon, population, annual fuel consumption and electricity generation, number and size of diesel generator sets etc.)
  • Measured hourly electrical load profiles for roughly 15 communities spanning 2-3 years for each community
  • Meteorological/environmental data in the form of hourly time series for 190 community locations (e.g. wind speed, wind direction, temperature, humidity, pressure, solar irradiance, albedo etc.)


Collaborators and Partners

  • NRCan - CanmetENERGY-Ottawa - Buildings and Renewables Group
  • NRCan Digital Accelerator


Ryan Kilpatrick
Research Engineer, CanmetENERGY-Ottawa

Page details

Report a problem on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: