Climate effects in data envelopment analysis for residential energy performance benchmarking: An empirical case validation

Reinforced fault diagnosis to discover inefficiency sources on energy performance of residential systems is critical for taking targeted retrofitting measures to enhance the effectiveness of energy upgrading investments. Data-driven energy benchmarking through nonparametric data envelopment analysis (DEA) has been classically adopted to locate malfunctions of energy systems with its unique capability of isolating the effects of diverse factors to differentiate overall efficiency, scale efficiency, and management efficiency. Nevertheless, extant works tend to subjectively process climate effects which could result in errors during DEA. This paper provides an empirical case to examine the reliability of two candidate methods on handling climate factors in residential energy efficiency evaluation, including the normalization-based and the weighting-based approaches. By investigating the performance of 70 single-family homes within a 3-year dynamic context, the improved robustness of normalization-based DEA approach is observed with statistical testing. It shows that the overall relative low energy efficiency is mainly due to poor management. This research intends to benefit retrofitting managers to make solid decisions.