A Methodology to Identify Monthly Energy Use Models from Utility Bill Data for Seasonally Scheduled Buildings: Application to K-12 Schools

摘要

The measured energy savings from retrofitsin buildings is often determined as the differencebetween the energy consumption predicted by abaseline model and the measured energyconsumption during the post retrofit period.Most baseline models are developed either byregressing the daily energy consumption versusthe daily average temperature ( daily model ) orby regressing the monthly energy consumptionversus the monthly average temperature (monthly model ).Savings measurement for buildings such asprimary and secondary schools (k-12 school) isvery difficult due to the special operatingschedules of such buildings. Currently, savingsare either determined by simple pre-post utilitybill comparison or by a method where by thebaseline model consists of two separate models:a 3-P model for non-summer months, and amean model for the summer months.(Landman, 1996).This paper proposes an improvedmethodology for identifying baseline models ofenergy use from utility billing data for buildingssuch as schools which have important daily andseasonal variations in occupancy. By explicitlyconsidering the occupancy rate in the model, weare able to generalize it and retain the distinctionbetween energy use levels during occupied andunoccupied days of the year. Thus the modifiedbaseline model accounts, not only for the effectof weather, but also for the influence of schoolschedules. The proposed methodology has beenevaluated against the previous 3-P-mean proposed by Landman for 10 schools in Texasfor which several yearsof monitored data are available. Incorporation ofscheduling information reduced the average CVof the model from 23.6% using Landman'smethod to 10.9% using our proposed method.