Validate primary fuel methodology that should be used for assigning a plant's fuel category

[grgmiller]

The OGE pipeline currently can use several methods to identify a plant's primary fuel: based on the fuel consumption, based on net generation, or based on nameplate capacity (see docs). The plant primary fuel identification is important because that is what is used to categorize each plant into the fuel categories that we use to match plant data up with the hourly EIA-930 data, and is also what we use when we calculate fleet-specific emission rates. Thus, ideally, we should be trying to categorize each plant into the fuel category that is used by EIA-930.

We currently default to identifying the primary fuel based on fuel consumption, but there is a chance that this approach, while technically accurate, may not match up with the fuel category used by EIA-930. For example, if there is a plant with a 100 MW coal generator that runs at 50% capacity and a 90 MW natural gas generator that runs at 90% capacity, our pipeline would likely identify this plant as a natural gas generator, because it is likely consuming more natural gas fuel (by mmbtu) than it is coal. However, just looking at the nameplate capacity of the plant, it has more coal generating capacity than gas capacity, so it might be categorized as a coal plant by the ISO and thus in EIA-930 (even if in a specific year, it did burn more gas than coal).

We have never validated whether one primary fuel identification method generally performs better than another. There are a few ways we could consider trying to validate which method we should use: We should try grouping plants into fuel categories based on different primary fuel identification methods, then:

1. Compare the error between total reported net generation for that fuel category in EIA-930 and total net generation in that category from our data.
2. Compare the error between total emissions calculated in OGE to total emissions calculated using the fleet emissions factors multiplied by EIA-930 net generation.

We should also generally look at how much each of these different primary fuel identification methods actually impacts the categorization of plants - how many plants does this affect? Are these plants large or small?

Also, we should start exporting the plant_primary_fuel table as part of our intermediate outputs.

[grgmiller]

See #296 for some work that:

• Exports the primary_fuel_table with all intermediate columns to outputs to help with validation.
• Adds a new validation check to flag when the plant primary fuel assigned by the pipeline does not match the capacity-based primary fuel assignment.

We still need to work on comparing whether one primary fuel assignment method better aligns with total generation reported in EIA-930.