The aim of the work is to compare the advantages and disadvantages of two feasible methods of evaluation of voltage dips "origin" intended as the network where the "fault" has probably occurred. The first method, mainly based on the simultaneous monitoring of an event by nearby primary substations connected to a common high voltage network, has been extensively applied since 2009 to the QuEEN (Quality of the Electric Energy) monitoring system data, by RSE within its research activity. The second one requests the monitoring of both primary substation medium voltage bus-bars and has been proposed as a possible method for the new ongoing Italian national monitoring system, developed by the Italian DSOs, by a technical committee on National Monitoring. This method, which has been shared as a possible and feasible method by all the participants to the committee (DSOs, TSO, researchers and the Italian regulator), can be applied at a few QuEEN substations compliant with its requirements. The first method requires the knowledge of the high voltage network topology and, when applied at research level, requires time-expending visual analysis of voltage dips waveforms to compensate the lack of "confidential" information. The second method has lower effectiveness but, being a local method, can be applied also to networks including only one substation.

The Italian MV networks voltage quality monitoring system has been collecting power quality data since 6 years by 600 instruments, 400 of which installed on HV/MV substations MV bus-bars and statistical representative of the network, the other 200 installed at some nodes of the MV lines [1]. The paper is focused on the analysis of voltage dips monitored at some of the 200 measuring points considered more significant. The aims were to confirm, on a wider scale, the representativeness of the statistics gathered on the HV/MV stations MV bus-bars with respect to that expected along the MV lines and to investigate the events way of propagation along different voltage levels. The analysis has been performed by both statistical methods on real data and DIgSILENT 1 simulations. The possibility to get the MV network voltage dip performance along the lines by measurements at HV/MV stations level has been widely confirmed. Measurements at MV lines have also confirmed the percentages of probable HV origin voltage dips, evaluated at national and macro area levels. The voltage dips duration distributions have shown a good accordance with the typical tripping times of the HV and MV protection system.

The subject addressed is the assessment of the origin (High Voltage HV / Medium Voltage MV / Low Voltage LV) of voltage dips monitored on networks and its effect on both voltage dip statistics and indices evaluation at both national and macro area level from the point of view of regulation. The method applied to assess the HV origin of the events refers to both the signal coming from the distance protection of HV lines and to the correlation of the occurrence time of those events which have been monitored on MV bus bars belonging to different HV/MV Primary Substations but underlying a common part of the HV grid. The MV bus bars monitored in this research are those of the Italian MV network monitoring system QuEEN. This method has been partially extended also to the events propagation from the MV to the LV networks thanks to the late installation of a small number of Measurements Units (MU) at specific Secondary Substations in order to realize a sort of "disturbed network tree". The estimation of the percentage of MV events with a probable HV origin at macro area level, for the 2010÷2011 period, has shown a range of variation of about 6%÷15% in comparison with the 37% evaluated at national level. As to the voltage dips performances, indices N 2a and N 3b , their evaluation at national level for 2011 is affected respectively by the 18,6% and 14,7% by voltage dips with a HV origin. As to the events propagation from the MV to the LV networks a good correspondence between the events monitored at the two voltage levels has been observed.