# US 11,999,392 B2 · Condition controlling of a wear and tear element > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,999,392 B2 - Document ID: US11999392B2 - Status: Granted - Kind: B2 - Application number: 16/616,720 - Filed: 2018-04-19 - Published: 2021-05-20 - Granted: 2024-06-04 - Earliest claimed: 2017-05-24 - Assignees: Siemens Mobility GmbH - Applicants: Siemens Mobility GmbH - Inventors: Maria Davidich; Dennis Klingebiel; Udhayaraj Sivalingam - Web page: https://patentgrep.com/patent/US11999392B2 ## CPC - B61L 15/0081: {On-board diagnosis or maintenance} - B61L 27/57: for vehicles or trains, e.g. trackside supervision of train conditions - G06N 20/00: Machine learning - G07C 5/004: {Indicating the operating range of the engine (economical driving B60R16/0236, F02B77/084; measuring liquid fuel of an engine G01F9/00)} - B61L: GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC (brakes or auxiliary equipment B61H, B61K; point or crossing construction E01B) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G07C: TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE ## Abstract An accurate and reliable maintenance method for controlling a condition of a wear and tear element of a track-bound vehicle includes determining a second condition of the element, being a chronological subsequent condition to a first condition, by starting from the first condition of the element, by using a machine learning algorithm, representing a chronological behavior of the element. A first action performable on the element is determined by using the determined second condition of the element and at least one predefined conditional criterion for the element. A resulting third condition of the element is determined by using a conditional change of the element. The conditional change is a consequence of the first performable action. A second performable action performable on the element is determined one more time by using the resulting third condition of the element and at least one predefined conditional criterion for the element. ## Claims ### Claim 1 A method for controlling a condition of a wear and tear element of a track-bound vehicle, the method comprising: determining a second condition of the wear and tear element, being a chronologically subsequent condition to a first condition, by starting from the first condition of the wear and tear element and using a machine learning algorithm representing a chronological behavior of the wear and tear element; providing a supervised learning algorithm as the machine learning algorithm, wherein the supervised learning algorithm is trained by an ordinary least square method; determining a first performable action performable on the wear and tear element by using the determined second condition of the wear and tear element and at least one predefined conditional criterion for the wear and tear element; determining a resulting third condition of the wear and tear element by using a conditional change of the wear and tear element, the conditional change being a consequence of the first performable action; determining a second performable action performable on the wear and tear element by using the resulting third condition of the wear and tear element and the at least one predefined conditional criterion for the wear and tear element; carrying out at least one action selected from the group consisting of the first performable action and the second performable action as machining of the wear and tear element; and adjusting a wear prediction model with the machine learning algorithm, and using the wear prediction model to reduce a number of inspections of the wear and tear element. ### Claim 2 (depends on 1) The method according to claim 1, which further comprises checking if the determined conditional change of the wear and tear element complies with at least one plausibility criterion, and basing the conditional change on at least one measured or determined conditional parameter referring to the wear and tear element. ### Claim 3 (depends on 2) The method according to claim 2, which further comprises selecting training data for the machine learning algorithm by using the at least one plausibility criterion. ### Claim 4 (depends on 1) The method according to claim 1, which further comprises carrying out a third or a follow-up performable action as a re-profiling or an exchange of the wear and tear element. ### Claim 5 (depends on 1) The method according to claim 1, which further comprises determining at least one of the first or the second or a third or a follow-up performable action as a YES/NO decision by using a decision model. ### Claim 6 (depends on 5) The method according to claim 5, which further comprises mapping the conditional change in the decision model as a YES/NO decision. ### Claim 7 (depends on 1) The method according to claim 1, which further comprises at least one of: describing at least one of the first or the second or the third condition by using at least one conditional parameter, or selecting the conditional parameter from the group consisting of: a diameter, a flange height, a flange width, a flange shoulder width, a slope of a flange shoulder, a groove, a rolled layer, a running tread width, a running tread surface, a work load of the wear and tear element, a travelled mileage of the wear and tear element, a thickness of a brake pad or shoe, an age of the wear and tear element, a profile of the wear and tear element and a roughness of the wear or tear element. ### Claim 8 (depends on 1) The method according to claim 1, which further comprises selecting the wear and tear element from the group consisting of a wheel, a brake pad, a brake shoe, a screw, a spring, and a lock. ### Claim 9 (depends on 1) The method according to claim 1, which further comprises selecting the wear and tear element as a wheel of the track-bound vehicle, wherein the track-bound vehicle is a train. ### Claim 10 (depends on 2) The method according to claim 2, which further comprises selecting the at least one predefined conditional criterion as an operational limit of the at least one measured or determined conditional parameter. ### Claim 11 (depends on 1) The method according to claim 1, which further comprises: determining a further second condition of a further wear and tear element, being a chronologically subsequent condition to a further first condition, by starting from the further first condition of the further wear and tear element and using the machine learning algorithm; and determining a follow-up performable action performable on the further wear and tear element by using the further second condition of the further wear and tear element, the third condition of the wear and tear element and at least one predefined conditional criterion for the further wear and tear element. ### Claim 12 (depends on 11) The method according to claim 11, wherein the wear and tear element and the further wear and tear element are both parts of a mutual structure. ### Claim 13 (depends on 1) The method according to claim 1, further comprising, when the first performable action is a non-action: determining a resulting fourth condition of the wear and tear element, being a chronologically subsequent condition to the second condition, by starting from the second condition of the wear and tear element and using the machine learning algorithm representing the chronological behavior of the wear and tear element; and determining a third performable action performable on the wear and tear element by using the resulting fourth condition of the wear and tear element and the at least one predefined conditional criterion for the wear and tear element. ### Claim 14 A control system for controlling a condition of a wear and tear element of a track-bound vehicle, the control system comprising at least one determining device for: determining a second condition of the wear and tear element, being a chronologically subsequent condition to a first condition, by starting from the first condition of the wear and tear element and using a machine learning algorithm representing a chronological behavior of the wear and tear element; providing a supervised learning algorithm as the machine learning algorithm, wherein the supervised learning algorithm is trained by an ordinary least square method; determining a first performable action performable on the wear and tear element by using the determined second condition of the wear and tear element and at least one predefined conditional criterion for the wear and tear element; determining a resulting third condition of the wear and tear element by using a conditional change of the wear and tear element, the conditional change being a consequence of the first performable action; determining a second performable action performable on the wear and tear element by using the resulting third condition of the wear and tear element and the at least one predefined conditional criterion for the wear and tear element; and adjusting a wear prediction model with the machine learning algorithm, and using the wear prediction model to reduce a number of inspections of the wear and tear element; wherein at least one action selected from the group consisting of the first performable action and the second performable action is a machining of the wear and tear element.