US 12,253,548 B2
System for monitoring and analyzing electric parameters
Fabio Corti, Scandicci (IT); Giacomo Talluri, Fiesole (IT); Francesco Grasso, Florence (IT); Marco Somma, Florence (IT); and Libero Paolucci, Grosseto (IT)
Assigned to UNIVERSITÀ DEGLI STUDI DI FIRENZE, Florence (IT)
Appl. No. 17/790,051
Filed by UNIVERSITÀ DEGLI STUDI DI FIRENZE, Florence (IT)
PCT Filed Dec. 31, 2020, PCT No. PCT/IB2020/062594
§ 371(c)(1), (2) Date Jun. 29, 2022,
PCT Pub. No. WO2021/137193, PCT Pub. Date Jul. 8, 2021.
Claims priority of application No. 102019000025855 (IT), filed on Dec. 31, 2019.
Prior Publication US 2023/0054387 A1, Feb. 23, 2023
Int. Cl. G01R 19/25 (2006.01); H02J 13/00 (2006.01)
CPC G01R 19/2513 (2013.01) [H02J 13/00002 (2020.01); H02J 13/00022 (2020.01)] 19 Claims
OG exemplary drawing
 
1. A system for monitoring and analyzing electrical operating parameters of a load in an electric network, said system comprising a smart socket arranged to be placed in series between said load and said electric network, said smart socket comprising:
a voltage detection module comprising a voltage sensor configured to measure a voltage value in said electric network, as an electric potential difference between the ends of said load;
a current detection module in said electric network comprising a current sensor arranged to measure a current value adsorbed by said load, when said load is connected to said electric network; and
a microcontroller connected to said voltage detection module and to said current detection module, wherein:
said microcontroller is configured to carry out:
a periodic acquisition, with a predetermined frequency f, of said voltage value in said electric network, obtaining a voltage trend over time; and
a periodic acquisition, with a predetermined frequency f, of said current value adsorbed by said load, obtaining a current trend over time;
said microcontroller includes a neural network arranged to carry out a training that includes the steps of:
definition of a number n of events Ei, each event Ei representing a disturbance or an operating anomaly of said electric network;
association, to each event Ei, of a number nit of patterns mi of predetermined current and/or voltage trends, where pij indicates the j-th pattern associated with the i-th event Ei, with i=1, 2, . . . , n and j=1, 2, . . . , mi, obtaining a number n of classified events Ei;
for each classified event Ei, extrapolation of characteristic parameters cik distinguishing the patterns pij associated with said classified event Ei, where cik indicates the k-th characteristic parameter of the i-th classified event Ei; and
said neural network is also arranged to carry out an analysis of said voltage and/or current trend acquired by said microcontroller, said analysis including the steps of:
comparison of said acquired voltage and/or current trend with predetermined voltage and/or current trends corresponding to the correct operation of said electric network;
definition in said acquired voltage and/or current trend of possible anomalous patterns with respect to said predetermined voltage and/or current trends; and
in case of identification of at least one anomalous pattern, search for said characteristic parameters cik in said or each identified anomalous pattern with consequent:
confirmation of the presence of said classified event Ei in said acquired voltage and/or current trend, in case that said or each identified anomalous pattern comprises at least one predetermined number of characteristic parameters cik of a classified event Ei; and
confirmation of the presence of an unclassified event in said acquired voltage and/or current trend, in case that said or each identified pattern does not comprise at least said predetermined number of characteristic parameters cik of a classified event Ei.