Algorithmically Developing Efficient Time-of-Use Electricity Rates

Abstract

The disconnect between price variability of wholesale electricity sales and the uniformity of the retail prices observed by most consumers results in well documented inefficiencies. Time-of-use (“TOU”) electricity tariffs offer a middle-ground between dynamic wholesale prices and flat retail rates. A customer subject to a TOU rate has their consumption binned into discrete seasons (e.g., summer, winter) and periods (e.g., off-peak, mid, on-peak) – with prices varying by season and period. TOU rates can be challenging to design and the results often rely on a regulator’s subjective sense of reasonableness.

In this paper, I outline a new method for developing multi-season, multi-period TOU rates using optimization techniques. This method identifies optimal TOU tariffs based on granular price and load data, using a curve-fitting approach. These TOU tariffs include explicit tradeoffs between different cost categories and between different season or period specifications. It also offers a simple method for comparing the relative efficiency of candidate rate designs, even if those designs have different numbers of seasons, periods, or price differentials. After describing the approach, I apply this methodology to a utility in California to demonstrate how it could be implemented in practice.

Keywords: Rate Design; Time-of-Use Rates; Tariffs; Optimization