A framework for sensitivity analysis of data errors on home energy management system

This paper investigates the impact of data errors on home energy management systems (HEMSs) that reduce energy cost and maintain comfort for residential consumers. In particular, we conduct a sensitivity analysis of HEMS subject to various types of input data such as the predicted energy consumption, the forecasted outdoor temperature, the consumers' comfort settings, static and dynamic operation constraints for home appliances, and the demand response (DR) signal. Using the perturbed Karush-Kuhn-Tucker (KKT) condition equations from the HEMS optimization formulation, we develop a linear sensitivity matrix to assess the impact of data on optimal solutions for: (1) electricity cost; (2) consumer's dissatisfaction cost; (3) the energy consumption for home appliances; and (4) the indoor temperature. The results of a simulation study using the developed sensitivity matrix provide HEMS operators with unique insight into factors that account for the relationships of HEMS operations to the change in the various data. Furthermore, these results can be used to provide insights for residential consumers and to evaluate the security risks of HEMS to cyber attacks through data manipulation.