US 11,748,632 B2
Analysis of anomalies in a facility
Refael Horev, Tel Aviv (IL); Alaa Ghanaiem, Sakhnin (IL); Bar Vinograd, Yaffo (IL); Firas Ismael, Nahif (IL); and Mohammad Badarna, Arraba (IL)
Assigned to SENSAI NETWORKS LTD, Haifa (IL)
Filed by SENSAI NETWORKS LTD., Haifa (IL)
Filed on Oct. 30, 2019, as Appl. No. 16/668,412.
Prior Publication US 2021/0133593 A1, May 6, 2021
Int. Cl. G06F 11/34 (2006.01); G06N 5/02 (2023.01); G06N 20/20 (2019.01); G05B 23/02 (2006.01)
CPC G06N 5/02 (2013.01) [G05B 23/0254 (2013.01); G06F 11/3447 (2013.01); G06N 20/20 (2019.01); G05B 2219/32201 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A method of analysing solving an anomaly in an operation of one or more electronic appliances including at least one computer, the method comprising, by a processor and memory circuitry:
upon detection of the anomaly corresponding to a deviation of a given parameter representative of the one or more electronic appliances from an operational state,
obtaining a given computer-implemented machine learning model associated with the given parameter, wherein the given computer-implemented machine learning model links one or more other parameters to the given parameter, wherein the one or more other parameters affect the given parameter,
based at least on the given computer-implemented machine learning model, identifying, among the one or more other parameters, at least one parameter Pj for which a change in its value allows bringing back the given parameter to the operational state,
determining, based at least on the given computer-implemented machine learning model and an operative range for which the given parameter is in an operational state, at least one value Vj for the at least one parameter Pj, which allows bringing back the given parameter to the operational state, wherein, when the at least one parameter Pj is a parameter which is mutable based on one or more other parameters, the method comprises:
(1) obtaining another computer-implemented machine learning model associated with the at least one parameter Pj, wherein said another computer-implemented machine learning model links one or more other parameters Pj′ to the at least one parameter Pj, wherein the one or more other parameters Pj′ affect the at least one parameter Pj,
(2) based at least on said another computer-implemented machine learning model and Vj, identifying, among the one or more other parameters Pj′, at least one parameter Pj′* for which a change in its value to a new value Vj′, allows bringing the at least one parameter Pj to the value Vj according to a matching criterion,
wherein, when the at least one parameter Pj′* is a directly mutable parameter of a given electronic appliance of the one or more electronic appliances, the method comprises sending, by the processor and memory circuitry, a command to the given electronic appliance to automatically change the value of the at least one parameter Pj′* to reach the new value Vj′*, thereby enabling bringing back the given parameter representative of the one or more electronic appliances to its operational state, to solve the anomaly.