Procedures for Filling Short Gaps in Energy Use and Weather Data

Filling short gaps (a few hours) in hourly energy use and weather data can be useful for (i) retrofit savings analysis and calculation, and for (ii) diagnostic purposes. The paper evaluates four methods for rehabilitating short periods of missing data. Single variable regression, polynomial models, Lagrange interpolation, and linear interpolation models are developed, demonstrated, and used to fill 1-6 hour gaps in weather data, heating data and cooling data for commercial buildings. The methodology for comparing the performance of the four different methods for filling data gaps uses 11 one-year data sets to develop different models and fill over 50,000 "pseudo-gaps" which are created by assuming data is missing and then comparing the "filled" values with the measured values. Comparisons are made using six statistical parameters including mean bias error, coefficient of determination, and coefficient of variation of the root-mean-square-error. For filling 1-6 missing hours of cooling data, heating data or weather data, a linear interpolation model or a polynomial model with hour-of-day (HOD) as the independent variable both provide a mean bias error of less than 0.087 % (0.005 F). The Lagrange model exhibits mean bias errors of 0.175 % (-0.010 F) which is better than the SVR model with temperature as the independent variable, which exhibits mean bias errors up to 0.909 % (0.062 F). Based on these findings, the polynomial model with hour-of-day as the independent variable and the linear interpolation model are recommended for filling data gaps of six hours or less in cooling, heating and weather data