Abstract:
|
Energy efficiency program administrators in the Northwest offer holistic strategic energy management (SEM) coaching to industrial and commercial sectors to reduce energy consumption in the region. As evaluators, we estimate energy savings resulting from SEM programs using regression models of daily or monthly energy consumption, an indicator for the SEM engagement, and other drivers of energy use including industrial facility production, commercial building occupancy, and operations and scheduling as well as time dependencies. We combine results across participants to estimate program total savings and use standard errors of the estimated SEM engagement effects to calculate confidence intervals. Until recently, SEM programs have seen low participation rates (10-20 customers per year) and evaluating all participants has been feasible. With growth in participation rates, this is no longer cost effective. Therefore, we sample facilities for evaluation, introducing sampling uncertainty. In this paper, we describe a typical SEM evaluation and describe the methods we use to combine regression and sampling error to quantify uncertainty in program total savings estimates.
|