Assessment of the impact of missing values in the southwest residential experiment photovoltaic array data records

Data films used by Sandia National Laboratories, Albuquerque (SNLA) to develop computer simulation programs for photovoltaic (PV) systems have data gaps as a result of malfunctions of the data acquisition system or human error. The gaps introduced error into the summary statistics computed from the incomplete files and disrupt the standard methods used to input data to the simulation programs. To improve the quality of the PV data files, SNLA conducted a study to determine (1) the frequency and length of the data gaps and (2) the effect of filling the data gaps with values determined by linear interpolation. The study results indicate that the amount of data missing from the files is not serious and that linear interpolation is an adequate solution. These results provide a method for estimating values to fill in the gaps in incomplete files, allow the analyst to compute summary statistics from these files, allow easy input of data to computer simulation programs, and, in general, increase the quality and usefulness of the data base.