LHL-Capstone

The Shift: Analyzing and Forecasting Alberta's Energy Grid

Link to tableau: (https://public.tableau.com/views/TheShiftAnalyzingandForecastingAlbertasEnergyGrid/TheShift?:language=en-US&publish=yes&:display_count=n&:origin=viz_share_link)

Project Intro / Goals:

This project analyzes hourly, time series data provided by the Alberta Electric System Operator (AESO) from January 1, 2010, to July 31, 2023, of each energy-producing asset under AESO. My goal for this project is to answer the following questions:

1. What is the trend of fuel types from 2010 to 2023? Are we on track for Net-Zero in 2035?
2. Alberta is historically known for Oil and Gas; are we accelerating the use of clean energy projects?
3. What patterns can we infer from wind and solar generation?

Results:

1. Natural gas produces the majority of electricity generation in the Province, nearly 70% each year; Wind is second.
2. Electricity generated by clean energy assets is growing exponentially.
3. There is twice the amount of fossil fuel generating capacity compared to clean energy, but fossil fuels supply the power grid with 8x more electricity.
4. Why is this? Asset Utilization = the amount of power an asset generates relative to its full capability. Fossil fuels have, on average, an 80% utilization rate as these assets can be set to a desired output to send out to the grid. Clean energy assets, however, have an average asset utilization rate of 24%.
5. Wind and Solar generation complement one another in Alberta. When the sun is down at night, wind generation is typically at its average daily peak, and when wind speeds are at their lowest average speed during summer months, solar power is at its peak.
6. Alberta needs more energy storage, 100x more than it currently has. Grid-scale batteries can augment wind and solar farms to increase asset utilization rates by storing energy when demand is low and then releasing it steadily when required. This mimics the ability of fossil fuel plants to run at a desired output.

Challenges and Next Steps:

This data is highly seasonal, as is to be expected with power generation time series. This presents an issue for building models that accurately represent the data, but also for forecasting future values. Due to wind and solar power generation relying on the weather, improvements can be made by adding wind and sun data during feature engineering.

Additional elements to add which are important pieces of the conversation regarding the clean energy transition are economics and emission reductions. These two topics can be projects of their own in this series of analyzing the Albert Energy Grid.