ANOVA-Based Variance Analysis in Smart Home Energy Consumption Data Using A Case Study of Darmstadt Smart City Germany

The evolution of smart grids (SG) is rapid and ubiquitous with the advent of information and communication technology. SGs enable utilities and prosumers to monitor energy consumption in real-time, thereby possessing effective supply and demand management. The subsets of SGs namely smart homes/smart buildings are tailored to take the benefits of SGs. These smart homes continuously record energy consumption data through smart meters, sensors, and smart appliances and facilitate consumers to track/manage their energy usage in real-time. Usually, the energy consumption of renewable energy-integrated smart homes depends on consumer behavior and weather conditions. These aspects lead to variance in the recorded energy consumption data from the desired levels. This variance in energy consumption impacts pattern finding, forecasting, financial risk, decision-making, and several other grid functionalities. Hence, comprehension of variance in energy consumption is essential to properly manage the energy. With this aim, this paper proposes the variance analysis on the smart home energy consumption readings using a statistical method named “Analysis of Variance (ANOVA)”. It is implemented on the Tracebase dataset, which is a smart city database and contains data for ten months. The data were collected in the city of Darmstadt, Germany, in 2012. The proposed ANOVA is applied to all these months’ data. As an initial step, the energy consumption readings recorded for every month at each day and at each hour are enumerated and this information is further used to perform day-wise variance analysis using ANOVA. The results show that there is a significant variance in several days in each month. Further, it is revealed that out of ten months, two months have high variability. Thus, this proposed variance analysis helps the stakeholders of SGs to take the necessary precautions for smooth grid functionalities as well as properly estimate future energy requirements.