US 12,468,614 B2
Method for determining heat-sink contamination by means of artificial intelligence
Bernd Roppelt, Unterhaid (DE); and Jens Schmenger, Forchheim (DE)
Assigned to Siemens Aktiengesellschaft, Munich (DE)
Appl. No. 18/023,021
Filed by Siemens Aktiengesellschaft, Munich (DE)
PCT Filed Jun. 22, 2021, PCT No. PCT/EP2021/066980
§ 371(c)(1), (2) Date Feb. 24, 2023,
PCT Pub. No. WO2022/042897, PCT Pub. Date Mar. 3, 2022.
Claims priority of application No. 20192872 (EP), filed on Aug. 26, 2020.
Prior Publication US 2023/0244585 A1, Aug. 3, 2023
Int. Cl. G06F 11/30 (2006.01)
CPC G06F 11/3062 (2013.01) 20 Claims
OG exemplary drawing
 
1. A method for determining contamination of a heat sink provided for cooling an electronic component, the method comprising:
determining with at least one temperature sensor for the heat sink having the electronic component at least one temperature profile of a temperature along a heat transfer chain from the electronic component to the heat sink for a load profile of the electronic component;
in a learning phase, determining of a prototype uncontaminated heat sink, which is defined as a heat sink as supplied to a user for permanent operation of the electronic component in an installation, load profile sections in the component of a heat sink under test whose contamination is to be determined;
determining, from the load profile for the electronic component of the heat sink under test and the at least one temperature profile of the learning sections, a decision function;
based on the determined decision function, ascertaining that the heat sink under test is equivalent to the prototype uncontaminated heat sink when for a section of the load profile for the electronic component of the heat sink under test that corresponds to a corresponding section of the load profile of the prototype uncontaminated heat sink at least one corresponding section of a temperature profile of the heat sink under test is similar to the corresponding section of the respective temperature profile of the prototype uncontaminated heat sink based on a similarity measure having a threshold value that is inversely proportional to a mean square deviation between the at least one temperature profile and the respectively assigned temperature section of the prototype uncontaminated heat sink, and a threshold value that is inversely proportional to an amount of a maximum time deviation between the at least one temperature profile and the respectively assigned temperature section of the prototype uncontaminated heat sink, and ascertaining that the heat sink is contaminated when the section of the load profile for the electronic component of the heat sink under test is different from the corresponding section of the load profile of the prototype uncontaminated heat sink and the at least one corresponding section of a temperature profile of the heat sink under test is different from the sections of the respective temperature profile of the prototype uncontaminated heat sink;
determining in a subsequent classification phase, classification sections corresponding to sections of the load profile that are similar to at least one section of the load profile assigned to a learning section based on the similarity measure;
comparing the temperature profile assigned to a classification section of the heat sink under test to the decision function determined in the training phase; and
detecting a contamination state of the heat sink by a distance measure of the decision function.