Optimizing Demand Management to Enable Renewables: Why the Use of a Marginal Emissions Signal is a Poor Choice

Public Service Company of Colorado has been rapidly transforming from a coal-dominant utility to a wind and solar-dominant utility as they develop and implement plans for 80% renewables by 2030. Above 50% renewables with 15% PV and 39% wind, the utility projects high levels of PV curtailment that effectively begin to limit future PV additions. Although storage is being installed, significant demand shaping and flexibility is not anticipated by the utility before 2030. In analyzing the possibilities to incorporate flexible demand, this paper uses monthly-average hourly projections of load, renewables, and emissions through 2042 from utility regulatory proceedings and hourly data from the Energy Information Agency to conclude that average emissions provide a better signal for demand flexibility to enable renewables than marginal emissions for this specific case of Colorado. Marginal emissions are due to the power plant on the margin providing incremental power when there is an increment in load. Prior to frequent renewable curtailments, this is usually a fossil fuel plant in this balancing area. This provides a weak signal for the cleanest hours with no signal at all indicating renewables availability until renewables are on the margin due to curtailment. In contrast, average emissions per MWh anticipate future renewables in this balancing area at levels currently on the grid. This can be used to further modify demand in anticipation of future renewables. The delayed renewables signal for marginal emissions patterns relative to average emissions is about 5-6 years in this case. The effect is very similar to recent discussion on the advantages of using long-run marginal emissions rates in preference to short-run marginal emission rates to evaluate the effectiveness of various demand profiles on reducing future grid emissions and costs.