Method for predicting a remaining failure or lifetime of an electrical component of an electrical circuit

Patent metadata

Field	Value
Publication number	US-12339328-B2
Application number	US-202218549401-A
Country	US
Kind code	B2
Filing date	Mar 11, 2022
Priority date	Mar 11, 2021
Publication date	Jun 24, 2025
Grant date	Jun 24, 2025

Abstract

The present invention provides a method for predicting a remaining lifetime of an electrical component of an electrical circuit, the electrical circuit being part of a building management device. The method comprises: estimating (S 1 ) two or more estimated temperatures of the electrical component by using a trained machine learning model of the electrical component that is trained based on training data. Further, the method comprises: generating (S 2 ) a temporal course of temperature of the electrical component based on the two or more estimated temperatures; and computing (S 3 ), based on the temporal course of temperature of the electrical component, an indicator for the remaining lifetime of the electrical component. The present invention also provides a method for predicting a remaining lifetime of an electrical circuit being part of a building management device.

First claim

The invention claimed is:

1. A method for predicting failure or a remaining lifetime of a building management device or an electrical component of an electrical circuit, the electrical circuit being part of the building management device, wherein the method comprises:
observing or estimating at least one characteristic parameter of the building management device or the electrical component;
generating (S 2 ) a temporal course of the characteristic parameter(s) based on the observed or estimated characteristic parameter(s); and
computing (S 3 ), based on the temporal course of the characteristic parameter(s), an indicator for the failure or the remaining lifetime of the building management device or the electrical component;
wherein computing the indicator for the failure or the remaining lifetime comprises determining a degree of similarity between the temporal course of the characteristic parameter(s) with a pattern of the respective characteristic parameter(s) retrieved from a memory, wherein the pattern stored is associated with a failure to be expected and the characteristic parameter(s) is/are observed parameter(s).

2. The method according to claim 1 , wherein
the estimated characteristic parameter is a temperature of the building management device or the electrical component and the method comprises:
estimating (S 1 ) two or more estimated temperatures of the building management device or the electrical component by using a trained machine learning model of the building management device or the electrical component that is trained based on training data comprising a plurality of data sets, wherein each data set of the plurality of data sets comprises:
a real temperature of the building management device, the module, or electrical component at a respective time of an operation of the building management device, or the electrical circuit, in association with
a plurality of parameters of the building management device, the module or the electrical circuit at the respective time, wherein the plurality of parameters comprises one or more physical parameters of the building management device or the electrical circuit and/or one or more operation parameters of the building management device or the electrical circuit;

estimating (S 1 ) each estimated temperature of the two or more estimated temperatures at a respective time by inputting (S 11 ) the plurality of parameters for the respective time to the trained machine learning model that computes (S 12 ) the estimated temperature of the electrical component at the respective time based on the plurality of parameters for the respective time.

3. The method according to claim 2 , wherein the machine learning model corresponds to:
a regression model, or
a neural network model, comprising a deep neural network model.

4. The method according to claim 1 , wherein the machine learning model comprises at least one algorithm differently weighting the plurality of parameters to compute, as an output, the estimated temperature at the respective time, wherein the weighting is determined by the training of the machine learning model based on the training data.

5. The method according to claim 1 , wherein computing the indicator for the failure or the remaining lifetime comprises integrating the temporal course of temperature.

6. The method according to claim 5 , wherein computing the indicator for the failure or the remaining lifetime further comprises weighting the integrated temporal course of temperature with a weighting function or weighting factor, wherein the weighting function or the weighting factor is dependent on at least one parameter of the plurality of parameters.

7. The method according to claim 1 , wherein information of a time relation of the pattern and the failure is stored associated with the pattern.

8. The method according to claim 7 , wherein the method comprises:
computing the further indicator for the failure or the remaining lifetime of the electrical component by comparing the one or more voltage ripple values with one or more initial voltage ripple values, wherein an increase of the one or more voltage ripples values with regard to the one or more initial voltage ripple values indicates a decrease in the remaining lifetime of the electrical component.

9. The method according to claim 8 , wherein the method comprises:
monitoring the temporal course of voltage ripple of the voltage of the electrical component, and
in case a change of the temporal course of voltage ripple within a time period is greater than a threshold for the change:
outputting an alarm indicating a reduced remaining lifetime of the electrical component; and/or
automatically adjusting at least one setting of the building management device; and/or

the method comprises:
monitoring the temporal course of capacitance of the electrical component, and
in case a change of the temporal course of capacitance of the electrical component within a time period is greater than a threshold for the change:
outputting an alarm indicating a reduced remaining lifetime of the electrical component; and/or
automatically adjusting at least one setting of the building management device.

10. The method according to claim 1 , wherein the temporal course of the characteristic parameter(s) is determined to correspond to a specific pattern, if the degree of similarity exceeds a predefined threshold.

11. The method according to claim 1 , wherein the pattern(s) are determined from observation of the characteristic parameter(s) until a failure occurs in a test environment or from an analysis of the failure and stored observed data on the characteristic parameter(s).

12. The method according to claim 1 , wherein the parameter(s) is/are selected from a group of parameters comprising:
a temperature of a control unit for controlling the electrical circuit;
a temperature of a substrate, comprising a printed circuit board, at which the electrical circuit is arranged;
a temperature inside a housing of the building management device;
a supply voltage received by the electrical circuit from an electrical energy source;
a supply current received by the electrical circuit from the electrical energy source;
frequency of the supply voltage and/or current received by the electrical circuit from the electrical energy source,
an output voltage output by the electrical circuit;
an output current output by the electrical circuit;
a power loss of the electrical circuit;
an energy conversion efficiency of the electrical circuit;
a value for setting an operation state of the building management device, and
in case the building management device is a lighting device, a dimming level at which the electrical circuit is operated for electrically supplying lighting means.

13. The method according to claim 1 , wherein the electrical component is a capacitor, the method further comprising:
obtaining one or more voltage ripple values of a voltage of the electrical component by obtaining for one or more time periods a respective voltage ripple value of the voltage of the electrical component; and
computing, based on the one or more voltage ripple values, a further indicator for the failure or the remaining lifetime of the electrical component.

14. The method according to claim 13 , wherein the method comprises:
obtaining two or more voltage ripple values of the voltage of the electrical component by obtaining for two or more time periods the respective voltage ripple value of the voltage of the electrical component,
generating, based on the two or more voltage ripple values, a temporal course of voltage ripple of the vo

Assignees

• Tridonic Gmbh & Co Kg

Inventors

• Klein Tobias
• Zengerle Thomas
• Schmölzer Andreas
• Mayrhofer Markus
• Carracedo Cordovilla Luis Javier
• Hüttinger Ulrich

Classifications

• G01R31/2849

  Environmental or reliability testing, e.g. burn-in or validation tests (of individual semiconductors G01R31/2642; of printed circuits boards G01R31/2817; of IC's G01R31/2855) · CPC title

• G01R31/2846

  using hard- or software simulation or using knowledge-based systems, e.g. expert systems, artificial intelligence or interactive algorithms · CPC title

• G01R31/282

  Testing of electronic circuits specially adapted for particular applications not provided for elsewhere (G01R31/2801 and G01R31/2851 take precedence) · CPC title

• G05B23/0254

  based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks · CPC title

• G01R31/64 Primary

  Testing of capacitors · CPC title