In a single half-hour period, we create a supply stack of available generation capacity, and rank it by short-run marginal cost.
We overlay demand.
The price is set where the two meet.
Our input data is not down to plant level, so we create a spline (or a linear fit) between the different technology sub-types to mimic plant-level price differences. It also means our resulting price curve is reasonably smooth when we do this across all half hours.
Typically when wind and solar are low, and demand is high, we get very high prices. Remember, this is without storage or interconnectors. Those are added in next.
While interesting, clearly this does not model reality. Storage and interconnection ease capacity constraints on the system.
Updated 10 days ago