Cooperative and self organizing paradigm for wide area synchronized monitoring of Smart Grids: Performance analysis in real operating scenarios

Synchronized Wide-Area Monitoring Systems (WAMS) are very important in the development and the use of power system-wide measurements to avoid large disturbances and reduce the probability of catastrophic events. Typically in WAMS the global absolute time reference for sensors synchronization is obtained by satellite-based timing signals processing. Since these signals are extremely vulnerable to radiofrequency interference, effective countermeasures aimed at increasing the resilience of synchronized WAMS to external and internal interferences need to be designed. To this aim several synchronization approaches are explored in literature. In particular, to avoid the large volume of raw data involved in typical client-server architectures, many researchers are analyzing the possibility offered by decentralized architecture based on consensus policies. In this scenario many consensus-based synchronization methods are proposed in literature but the attention has rarely been paid to the influence factors that characterize real environments and real smart grid applications as: (i) the presence of hardware characterized by finite processing time, and finite resolution of nodes involved in the consensus procedure; (ii) the effect of typical clock instability as drift, jitter and wander; (iii) the effect of latencies and limited bandwidths available on a radio system or when common commercial wireless communication are adopted. All the above-mentioned influence factors can be thought of as causes of uncertainties that could affect the overall performance. This paper proposes a study these influence factors both in simulated and real scenarios. The final aim is to find an uncertainty model to evaluate and forecast the reliability and the quality of the time synchronization procedure.