GREEND: An Energy Consumption Dataset of Households in Italy and Austria (1405.3100v2)

Abstract: Home energy management systems can be used to monitor and optimize consumption and local production from renewable energy. To assess solutions before their deployment, researchers and designers of those systems demand for energy consumption datasets. In this paper, we present the GREEND dataset, containing detailed power usage information obtained through a measurement campaign in households in Austria and Italy. We provide a description of consumption scenarios and discuss design choices for the sensing infrastructure. Finally, we benchmark the dataset with state-of-the-art techniques in load disaggregation, occupancy detection and appliance usage mining.

• The paper introduces the GREEND dataset, the first 1Hz energy consumption dataset from households in Italy and Austria, collected over time from nine homes.
• The dataset supports advanced applications like Non-Intrusive Appliance Load Monitoring (NILM), occupancy detection, and appliance usage modeling using techniques like Particle Filtering and Bayesian Networks.
• The GREEND dataset has extensive implications for smart grid technology, demand-side management, and home energy management systems (HEMS), enabling development and validation of energy conservation strategies.

An Analysis of the GREEND Energy Consumption Dataset

The paper "GREEND: An Energy Consumption Dataset of Households in Italy and Austria" introduces a comprehensive dataset that aims to aid researchers and engineers working on energy management systems. This dataset was collected over a period of time in nine households located in Austria and Italy, marking it as the first known 1Hz consumption dataset for these regions. The GREEND dataset serves a crucial need in the domain of energy modeling and management by providing fine-grained power usage data that support detailed analysis and application of various machine learning techniques.

Methodology and Data Collection

The dataset was gathered as part of the MONERGY project, a cross-border initiative between Austria and Italy directed at reducing energy consumption in these regions. The dataset encompasses time-series data capturing active power at each second, spanning diverse household scenarios with varied appliances and inhabitant profiles. Sampling at this rate provides an advantage over existing datasets, offering a high-resolution insight into household electrical consumption.

The implementation involved deploying an ARM-based platform with a Plugwise Basic Kit set in each household. The setup ensured a robust and continuous recording of consumption data for up to a year, thus enabling the capture of seasonal variations in energy use.

Applications and Case Studies

The paper highlights three primary applications of the GREEND dataset: load disaggregation, occupancy detection, and appliance usage modeling.

1. Non-Intrusive Appliance Load Monitoring (NILM): Using a Particle Filtering (PF) approach applied to Hidden Markov Models, the paper demonstrated how the dataset could facilitate the disaggregation of energy consumption into component appliances. The analysis achieved high accuracy, showcasing the dataset's potential application in optimizing smart home energy systems through effective device-level monitoring.
2. Occupancy Detection: The paper applies the NIOM algorithm to infer occupancy through energy consumption patterns. By using weekday and weekend data, the derived probabilities adeptly reflected household behavioral patterns, providing an innovative means to detect occupancy without intrusive sensors.
3. Appliance Usage Modeling: Here, Bayesian Networks were employed to model and predict appliance usage patterns. By learning from the dataset, the research forecasts appliance operation and thus assists in building intelligent home management systems that adapt to user habits.

Implications and Future Directions

The implications of this dataset are extensive, spanning both theoretical and practical applications in the fields of smart grid technology, demand-side management, and HEMS optimization. By offering open access to a high-frequency dataset, the GREEND project empowers researchers to develop and validate sustai...

Looking forward, the GREEND dataset provides a foundation for extending research into microgrid simulation and renewable energy integration, offering a holistic view of home energy ecosystems. The project's next phase will focus on aggregating data to encapsulate total household demand, offering a more inclusive vantage for simulation and strategy assessment. Additionally, integrating renewable energy production data will present comprehensive insights necessary for the exploration and emulation of net-zero energy homes.

The GREEND dataset thus marks a significant contribution to the paper and evolution of smart home energy systems, enabling the application of advanced analytical techniques in real-world energy conservation strategies.