| dc.description.abstract | In recent years, the introduction of Smart Grids has provided us with access to a new layer of energy consumption patterns.
This new layer of consumption could be analyzed to boost efficiency, prevent power loss, and generate significant economic and environmental benefits.
For the analysis of this data layer, a relatively new paradigm has emerged: fog computing or fog networking.
Fog computing seeks to offload a cloud computing architecture by decentralizing data processing from the typical cloud computing platform to edge nodes that have less computing power and are closer to the data source.
In this study, we intend to implement an anomaly detection algorithm on Smart grid data using a model that employs a generative adversarial network and long term short therm memory to classify anomalies in data from smart grid customers.
This algorithm will be evaluated on multiple platforms, including Edge devices and Cloud virtual machines, and run-time metrics will be collected for comparison purposes. | |
