The report describes the activity conducted on voltage monitoring in medium-voltage networks, involving the first large-scale systematic comparison (~30,000 voltage signals) between the outcomes of applying new criteria based on Deep and Machine Learning for characterizing voltage sags in terms of validity and origin (DELFI RMS and FExWaveS) and current criteria (2nd harmonic and Global Method), which are active in Italian monitoring systems developed to monitor the Power Quality of distribution networks. The comparison, on a national and macro-area scale, was conducted using confusion matrix-type tables, evaluating the percentage consistency of the new criteria versus the current ones. The evaluations focused on severe events originating in the medium-voltage network, which are relevant for regulation: the adoption of the new criteria would lead to an average increase of 43% in the synthetic index N3b that counts them. The new criteria have been implemented in the QuEEN system to facilitate comparisons with future data sets and make the results available on the updated related website. The acquired experience may help optimize the current criteria used in national monitoring, confirming the existing one-to-one relationship between research-use monitoring systems (QuEEN) and regulatory systems (MonNaLiSA). Regarding advanced voltage monitoring methods, higher-order moments (HOM) were applied to analytically generated continuous disturbance sets (harmonics, including overlapping events, flicker) and to simulated sequences of continuous disturbances and events occurring in portions of a distribution network. The characteristic parameters of HOM distributions (statistical approach) were evaluated, constructed from the generated voltage signal sets (~500,000 signals). Skewness and kurtosis are capable of detecting the presence of even and odd harmonics in voltage signals, respectively, while a global Power Quality (PQ) disturbance index would be sensitive only to dominant harmonics. Comparing signals affected by random harmonic disturbances within normal limits with those where harmonics exceed these limits by a factor of three, an increase of 160% / 220% in the standard deviation of skewness/kurtosis distributions is observed. The presence of impactful events, such as voltage sags, prevents HOM from detecting continuous disturbances on short time scales.