A Methodology for Detection of Power Quality Disturbances in the Context of Demand Side Management

With the wide use of non-linear loads and the integration of multiple power systems, there is an increased risk of damaging power quality. Automatic detection of disturbance is the first step in dealing with power quality problems. Most of the related works founded in the literature, to detect power quality disturbances, use high computational cost techniques, making difficult to board in hardware. Thus, this work proposes a methodology with a low computational cost for disturbances detection in electrical power quality aiming embedded in hardware. In this way, the pre-processing stage employed a sliding window, with a one-point step, and for each window, two features are calculated: the root mean square and the harmonic distortion, to be used in the disturbance detection. From the results obtained through synthetic data, it was possible to observe that the proposed methodology can efficiently and rapidly detect the presence of disturbances with an accuracy rate greater than 90% for signals with more than 25 points of disturbance. The observed results can also be considered relevant for power quality analysis in Smart Grids because it can be shipped in a low-cost smart meter.