This work presents the implementation of a Novelty Detector (ND) on a Field Programmable Gate Array (FPGA) platform, using the similarity-based comparison technique known as Dynamic Time Warping (DTW). This approach stands out for its ability to analyze data segments of varying lengths, enabling the system to adapt to Power Quality (PQ) disturbances present in voltage and current signals of Electric Power Systems (EPS). This adaptive feature is crucial for identifying previously unseen events that may arise with the increasing deployment of Smart Grids (SG) in EPS. Although DTW typically involves higher computational complexity compared to fixed-segment novelty detection techniques, hardware and software optimizations—combined with the use of soft-core processors—result in an architecture capable of real-time operation. Functional simulation tests confirm that the developed hardware efficiently computes DTW and identifies segments with novelties in signals sampled at a rate of 3840 Hz. This implementation has been validated on a functional prototype, demonstrating the system's feasibility for practical application in real-world scenarios.