Bayesian verification of an energy conservation measure

Abstract Most building owners eventually invest in energy conservation measures for their buildings. Faced with a variety of options, ranging from trivial (replacing light bulbs) to major (better facade insulation), they want to know which measure will yield the highest return on investment—and with what confidence. Moreover, once the measure has been applied, they will want to know how well it performs and whether their money was well spent—also known as Measurement & Verification (M&V). But conventional M&V mandates the establishment of a baseline. The building should be instrumented during a year or more before applying the measure, driving the costs of M&V to sometimes unacceptable levels. Furthermore, a typical M&V study will report a single number for the measure’s efficiency, ignoring any uncertainty surrounding that estimate. To solve these two problems (expensive baseline and absence of uncertainty), we have developed a method, based on Bayesian statistics, that will 1) rely on historic utility bills and climate data to establish the baseline, and 2) estimate, with confidence intervals, how effective the energy conservation measure was. We have tested this method after installing, in March 2016, a model-predictive controller for space heating for a medium-sized (about 2700 m2) office building in Switzerland. The baseline was established from historic oil tank refill records and climate data from a nearby weather station. The total heat loss coefficient of the building was assessed 46 days after the installation and found to have decreased by 31.9%, with 17.8 percentage points standard error.