US 12,406,192 B2
Service location anomalies
Carlos Iga Garza, Powhatan, VA (US)
Assigned to Landis+Gyr Technology, Inc., Alpharetta, GA (US)
Filed by Landis+Gyr Innovations, Inc., Alpharetta, GA (US)
Filed on Apr. 29, 2022, as Appl. No. 17/732,788.
Claims priority of provisional application 63/216,375, filed on Jun. 29, 2021.
Prior Publication US 2022/0414484 A1, Dec. 29, 2022
Int. Cl. G06N 20/00 (2019.01); G01R 19/25 (2006.01); G01R 31/08 (2020.01); G06N 5/022 (2023.01); H02J 3/00 (2006.01); H02J 13/00 (2006.01)
CPC G06N 5/022 (2013.01) [G01R 19/2513 (2013.01); G01R 31/086 (2013.01); H02J 3/0012 (2020.01); H02J 13/00002 (2020.01); H02J 2203/20 (2020.01)] 18 Claims
OG exemplary drawing
 
1. A method of using machine learning to detect an electrical anomaly in a power distribution system, the method comprising:
accessing a first plurality of voltage measurements measured at an electric metering device;
calculating, from the first plurality of voltage measurements and for each of a first plurality of time windows, a first corresponding average voltage and a first corresponding minimum voltage;
training a machine learning model by:
accessing a set of training data pairs, wherein each training data pair comprises one or more of (i) a training set of average voltages and a set of minimum voltages, (ii) a training set of average voltages of all the electric metering devices connected to a distribution transformer, or (iii) a training set of average voltages of one electric metering device that is behind the distribution transformer and an expected classification that indicates one or more electrical anomalies;
providing one training data pair of the set of training data pairs to the machine learning model;
receiving, from the machine learning model, a determined classification;
calculating a loss function by comparing the determined classification and the expected classification; and
adjusting internal parameters of the machine learning model to minimize the loss function;
applying the machine learning model to the first average voltages and the first minimum voltages, wherein the machine learning model is trained to identify, from voltage measurements, a first voltage signature that corresponds to the electrical anomaly;
receiving, from the machine learning model, a first classification indicating a first loose connection;
based on the first classification, sending a first alert to a utility operator;
calculating, from a second plurality of voltage measurements and for each of a second plurality of time windows, a second corresponding average voltage and a second corresponding minimum voltage, wherein the second plurality of time windows occur before the first plurality of time windows;
applying the machine learning model to the first average voltages, the first minimum voltages, first voltage signature, the second average voltages, and the second minimum voltages;
receiving, from the machine learning model, a second classification identifying a second voltage signature indicating a second loose connection; and
based on the second classification, sending a second alert to the utility operator.