Model Architecture¶
This page covers how CANOE is structured internally: how sectors relate to each other, how their output comes together into a single database, and what happens to that database before it's ready for TEMOA to run.
Sectors as demand models¶
CANOE is organized into seven sectors: Electricity, Agriculture, Transportation, Industry, Residential, Commercial, and Fuels. At their core, each sector is a model of end-use demand: how much energy (and what kind) is needed to meet Canadians' needs over time.
Sectors vary in complexity:
- Some represent demand as a simple aggregated values per period.
- Others include levers that shift both the amount and the type of demand — for example, how many electric vehicles versus internal-combustion vehicles are on the road, which changes both total energy demand and which fuel/commodity it falls on. These are often technologies that the model can scale up or down to meet demand.
Linker modules¶
Fuel and Electricity play a different role from the other sectors: rather than representing an end-use demand on their own, they act as linkers, tying the sector demand models to energy prices, imports, and distribution.
Electricity's dual role
Electricity currently covers both generation and distribution in one module. Separating generation out as its own concern is being discussed but not yet implemented.
From sectors to a master database¶
Each sector and linker module writes into a shared SQLite database following the TEMOA/CANOE schema (see CANOE vs. TEMOA). CANOE compiles all of this into a single master database.
Within that master database, each sector's demand is represented at two different resolutions, side by side:
- High resolution — the sector modules' own bottom-up output: the detailed, module-by-module demand and technologies data described above.
- Low resolution — an alternative built from CEF (Canada's Energy Future) projections, a coarser, top-down aggregate view of the same demand.
Both versions live in the master database at once. Choosing between them is part of what the filtering step below does.
Why two resolutions
The low resolution sectors significantly reduce the computational complexity of experiments. They can be used to assess changes in one sector while keeping a reasonable pressure from the rest of the modules.
Filtering and representative periods¶
Before a run, the master database goes through two narrowing steps:
flowchart TD
A[(Master database
high-res modules + low-res CEF)] --> B[Filtering interface]
B --> C[(Filtered database)]
C --> D[Representative periods]
D --> E[(TEMOA-ready database
reduced temporal resolution)]
E --> F[TEMOA run]- Filtering interface — narrows the master database down to a specific case of interest: which provinces, which scenario, and which resolution track (high-res module output or low-res CEF) to use for each sector, rather than the full compiled dataset.
- Representative periods — often 365 days per year of temporal resolution is finer than what TEMOA can practically optimize, so this step reduces it to a manageable set of representative time periods before the run.
Both steps are necessary today because TEMOA can't optimize efficiently against the full master database directly. They're called out separately here because they show up as distinct stages throughout the docs (and as distinct tools/repos) — the Get Started guide covers how to actually run them.
Where to go next¶
-
Glossary
Plain-language definitions of capacity-expansion and TEMOA terms.