The growing energy demand sharpens the search for technological moder- nizations capable of meeting imminent needs, as well as increasing concerns about mitigating the environmental impacts that come with this escalation. The state of the art in Smart Grids refers to evidence of the use of AI techni- ques in IoSGT use cases, aiming to revolutionize the way energy is produced, transmitted, and consumed. In fact, AI is expected to offer unprecedented levels of disruption in the electric sector, through intelligent control methods that can unlock new value streams for consumers, while allowing support for a highly assertive, reliable, and resilient system. However, much research is still needed in this area, such as the positioning of AI-based instances along the edge-cloud continuum, types of techniques and algorithms for each use case, efficient use of predictive analytics capable of predicting future demands, detecting failures and anomalies in the power grid that allow for the adoption of proactive measures and improving network reliability, among many others.

This research proposal aims to address some of the previously mentioned issues through a holistic architecture named IAIoSGT (Artificial Intelligence native in IoSGT). IAIoSGT is designed with the assumption of accelerating the use of AI techniques in an approach based on the continuous edge-cloud continuum. The assessment of the IAIoSGT architecture’s compliance, as well as its behavior and feasibility of use, was conducted on two distinct test benches, addressing both physical devices and Machine Learning algorithms. It is noteworthy that two comprehensive tests were carried out: the first one pertains to the classification and identification of electroelectronic devices connected in the same electrical network, involving Machine Learning algorithms such as KNN, SVM, MLP, NB, and DT. The second test focused on energy consumption

xi

prediction, utilizing the LSTM algorithm.