How do we get reliable long-term forecasting?

FERC is pushed by consumer groups to not allow over-building.

A generation facility that sits idle or is curtailed is an expensive asset that must be paid for in the world of generation planning and interconnection. Consumer advocates do not want ratepayers to be on the hook for these costs. This applies to both generation and supporting transmission.

The future is uncertain, and load forecasts range from 50% increase to 400% increase by 2050 (generation plant life for nuclear plants is 80-100 years, and natural gas combined cycle is 40-80 years, with repowering solar and wind can produce for 60+ years). 2050 is well within the life span of plants built today. One forecast indicates that we need to build in the US 80 GW of new generation per year, double the rate we have been.

Plants (even solar) can take up to 15 years from conceptual discussions to commissioning (mostly permits, queues, and supply chain issues).

Skeptics don’t believe that the growth in demand is going to happen, or at least most of it. They do not want projects to even start the conceptual stage and blast the forecasts that exist. These disagreements on forecasts, even using scenario planning in this transitional period have now reached FERC. How do we predict the future?

1)     We can put hard limits on growth. 1GW a year for data centers, 1 GW per year for housing (allocated by state), etc. The US becomes a planned economy and limits the growth of power use. To make this work, the limits would need to be a hard limit for the next 20-25 years with updates every 5 years.

2)     We could make large loads, pay 100% of their cost for energy, permits, interconnection, construction, decommissioning, environmental, etc. Holding harmless existing loads. This does not solve for electric vehicles, or decarbonization, but it takes large loads out of the picture.

3)     We could return to cost of service as the model for generation. Until the 1980s, all generation was based on cost of service and there were very few surprises. Markets were supposed to reduce the costs for customers, but that has only really been true for large customers over the last 30 years. With the increasing wild daily price swings, from solar max to evening no solar.
4)     We could move to a microgrid model and have every neighborhood take full responsibility for their own costs, reliability, maintenance and billing to their members.

5)     We could under build based on existing forecasts, by 10-70% (this is suggested by some consumer advocates - the numbers are theirs) and wait to plan any new facilities until we overrun the existing energy production. [One argued that rolling blackouts would be cheaper than over building.]

6)     Something else.

There is no way to predict energy use in 2035, or 2050 with a level of accuracy that prevents over building (or under) with the timelines for construction.

Your thoughts?