A WEATHER-DAYTYPING PROCEDURE FOR DISAGGREGATING HOURLY END-USE LOADS IN AN ELECTRICALLY HEATED AND COOLED BUILDING FROM WHOLE-BUILDING HOURLY DATA

摘要

Hourly monitoring of a building s energy use has been shown to be an effective method for verifying building computer simulations. Ideally, a model should be calibrated with end-use data measured at several strategic monitoring points such as the lighting panels, motor-control-centers, and the heating and cooling equipment in order to match the simulation outputs. Unfortunately, in many cases, the cost of extensive sub-metering of various energy using components is not justifiable. This constrains engineers and architects to rely on hourly whole-building electricity data, weather data from nearby National Weather Service (NWS) weather stations, and selected field measurements to calibrate a simulation to hourly whole-building data.Methods have been developed to classify hourly whole-building data into daytypes or load shapes. Daytyping techniques used today vary from simple average-hourly, weekday/weekend profiles to complex statistical methods including principle component and cluster analysis. One method that has proven to be effective includes daytyping by weather-daytypes. Weather daytyping provides additional insight over simple average-hourly daytypes because it can be used to identify weather dependent (i.e., heating or cooling) and non-weather dependent (e.g., lighting or receptacle loads) 24-hour weekday/weekend profiles.In order to improve upon previously established techniques, this paper presents an improved weather daytyping procedure that is capable of disaggregating a whole-building electricity signal into hourly end-use loads. This methodology utilizes "representative days," to designate the non-weather dependent building base load for occupied and unoccupied hours which is then sorted into two additional weather-daytypes; one for the heating period and one for the cooling period. This method provides six end-use categories including: weekday/weekend, heating, cooling, and non-heatingon-cooling end-use load disaggregations from the hourly whole-building electricity load. The three daytypes provide a convenient match to the heating,cooling, and non-heatingon-cooling loads from a simulation program such as DOE-2.Results from the application of the methodology to a case study building located in Washington, D.C. are presented. In the case study building, end-use loads derived from nine months of hourly whole-building electricity data and site-specific weather data are compared against end-use loads predicted by a calibrated DOE-2. 1D simulation of the building. For the case study building heating, cooling, lighting, and receptacle loads calculated with the weather-daytypes compare favorably with loads calculated by a calibrated DOE-2 simulation for the same building.