Physics- and statistics-based quality control of weather files

Dynamic building energy simulation programs require accurate weather data to produce reliable energy consumption results. Traditionally, hourly weather files have been vetted using “physical” quality control procedures (QCPs), such as wind speeds rarely changing more than 15 m/s in an hour. However, weather data that proves to be suspect upon careful inspection may pass these physical QCPs unnoticed. The quality control strategy for yearly hour weather files proposed in this paper supplements current physical QCPs in the literature with statistical QCPs based on historical local weather data, pulled from ASHRAE’s Weather Data Viewer (2017). These statistical QCPs use historical data going back up to 30 years at the weather station to flag hourly magnitudes, hourly steps, daily profiles, or monthly averages of meteorological variables that fall outside what is historically expected at that weather station. This enhanced methodology successfully catches suspect weather data that would otherwise pass traditional QCPs.