COMPRESSIVE SENSING-BASED METHODOLOGIES FOR SMART GRID MONITORING

Modern distribution networks, commonly known as Smart Grids, will be characterized by strictly requirements in terms of reliability and efficiency of the power supply. This will require a high empowerment in the management of the distribution, and transmission, networks by the system operators. Problems such as the identification of the prevailing harmonic sources and the fault location are characterized by criticality which must be appropriately taken into account, in order to fully exploit the capabilities of the Smart Grids. The analysis of both phenomena requires an appropriate monitoring of the networks, which are currently characterized by the availability of a limited number of measurements. This increase the complexity of the analysis of distribution networks, and the necessity of developing ad-hoc algorithms and solutions aimed at supporting the system operators while managing the networks. In this thesis, Compressive Sensing-based algorithms for detecting the main harmonic polluting sources, and for identifying the location of faults occurring in distribution systems have been presented. With reference to the identification of the main harmonic sources, two algorithms have been proposed: one for detailed analysis, with reference to a specific harmonic order, and one for more general analysis, which allows to investigate multiple harmonic orders simultaneously. The performed tests have proved how both methodologies are robust with respect to the measurement uncertainties, underlying the different capabilities of the two methods. Contrarily, the performance of the fault location algorithms are more influenced by the higher uncertainties in measuring the dynamic signals involved during the fault. The analysis performed considering the proper uncertainty scenarios have underlined how the use of modern devices for branch current measurements allow to increase the performance of the fault location algorithms; providing additional information which are useful for locating the fault.