Abstract:

The increase of smart home culture for improved efficiency and comfort in the present energy sector requires paying much attention to big data analytics. Here, the data refers to the energy consumption readings that are continuously captured through smart meters and transmitted to the central computing centres. The entire analysis and decision making in such cases depend on the availability of quality data. However, this data often contains anomalies such as redundancy (duplicated data), which affects their quality. Thus, a systematic approach with three steps (exploration, behavior analysis, and visualization) is proposed in this paper to precisely analyze the redundant data anomalies and their behavior. In exploration, the identification and quantification of redundant data anomalies will be done for all appliances for all available days. This provides the information of the highest and lowest counts of redundancies for all appliances. In behavior analysis, the behavior of redundant data anomalies during various parts of a day will be analysed. The visualization finds the occurrence of redundant data anomalies at the day/hour/minute level. Altogether, these three steps provide a comprehensive analysis of redundant data anomaly behavior that is present in the smart home energy consumption dataset. For the analysis, this paper considers a real-time smart home dataset ‘Tracebase’. From this dataset, the appliance ‘WaterKettle’ is used as an example for the proposed analysis as it exhibits the highest redundancy count when compared to all other appliances. Form the implementation of the proposed approach, it is revealed that there is a high occurrence of redundancy during Daylight hours and is visualized.

Keywords: