Granted Patent B2

US 12,638,842 · App. 17/597,666 · Granted May 26, 2026

Inventors: Thomas Albrecht (Beilngries, DE); Lukas Schindler (Duggendorf, DE); Benedikt Boettcher (Bruckmuehl, DE)

Assignee: KRONES AG

G05B23/0283G06N20/00

Abstract

Claims (52)

1 . A method for determining an optimized production plan for production of one or more products using one or more production lines in the food industry or the beverage industry, the method comprising:

optimizing a changeover time of the production plan, providing the changeover time of the production plan to a user, and then implementing the changeover time of the production plan, the changeover time comprising a product changeover, bottle changeover, type changeover, format changeover and/or pack changeover in the production of the products, wherein optimizing the changeover time comprises:

breaking down the production plan into a plurality of contiguous units, each of the units corresponding to at least one product, processing step, or at least one device of a production line for producing a product of the production plan; and

arranging the units so that the changeover time has a minimum value; and

implementing optimized production costs of the production plan to be spent on processing the units, wherein the production costs are determined by retrofitting times and production times for the units, wherein the retrofitting times and production times are defined as product- and device-dependent, and wherein implementing the optimized production costs comprises:

assigning the units to a respective production line, and

determining an order in which the units are to be processed by the respective production line and processing the units in the determined order;

wherein the changeover time and/or the production cost additionally comprise a duration of a maintenance action for a detected malfunction of the device of the respective production line.

2 . The method of claim 1 , wherein the method is further for automatically detecting the malfunction of the device in the food industry or the beverage industry, the method further comprising:

generating simulation data by means of a digital twin of the device, the simulation data being generated for a plurality of malfunctions;

training a machine learning algorithm on the simulation data and the plurality of malfunctions so as to generate a trained machine learning algorithm; and

detecting a malfunction of the device by the trained machine learning algorithm using operating data of the device;

wherein the detected malfunction of the device is an event, a failure, and/or a maintenance action to be taken, wherein:

the event comprises increased friction, increased vibration, imbalance, a wear parameter, or decreased lubricity,

the failure comprises an engine failure, a bearing failure, an axle failure, a crack, an incorrectly positioned bottle, a short circuit, a stoppage of the device, or any other form of mechanical damage, and

the maintenance action to be taken comprises cleaning, lubricating, oiling, fixing, adjusting, replacing, or repairing; and/or

wherein detecting the malfunction further comprises detecting a component of the device on which the malfunction occurs.

3 . The method of claim 2 , wherein detecting the malfunction further comprises detecting the expected time period until the failure occurs from the malfunction.

4 . The method according to claim 2 , wherein the operating data comprises output signals of the device, wherein the simulation data comprises output signals of the digital twin, and wherein each of the output signals of the digital twin corresponds to a respective output signal of the device; and

wherein each of the output signals of the device is measured by a sensor of the device used during operation of the device.

5 . The method according to claim 4 , wherein the operating data of the device comprises at least one additional output signal of the device, the additional output signal measured by an additional sensor of the device that is not required during operation of the device;

wherein the simulation data comprises at least one corresponding additional output signal of the digital twin; and

wherein the additional sensor of the device is selected from a plurality of potential additional sensors by means of the digital twin such that the additional output signal of the additional sensor optimizes detection of the malfunction.

6 . The method according to claim 2 , wherein the trained machine learning algorithm for detecting the malfunction determines a probability for the malfunction using the operating data of the device; and

wherein the trained machine learning algorithm recognizes the malfunction when the probability of the malfunction reaches a threshold value.

7 . The method according to claim 2 , wherein the machine learning algorithm is additionally trained with recorded operating data of the device; and/or

wherein the device is a tripod of the food industry.

8 . The method according to claim 7 , wherein the machine learning algorithm is trained with both the recorded operating data of the device and the simulation data, and wherein the device is a tripod of the food industry.

9 . The method according to claim 2 , wherein generating the simulation data comprises performing a plurality of simulation runs by the digital twin for each of the plurality of malfunctions.

10 . The method according to claim 9 , wherein the plurality of simulation runs are subject to stochastic variations.

11 . The method according to claim 2 , wherein training the machine learning algorithm on the simulation data and the plurality of malfunctions comprises using the plurality of malfunctions and a presence of no malfunction as ground truth.

12 . The method according to claim 2 , wherein training the machine learning algorithm on the simulation data and the plurality of malfunctions comprises using a predicted time until malfunction as ground truth.

13 . A system for determining an optimized production plan for production of one or more products using one or more production lines in the food industry or the beverage industry, the system being configured to:

optimize a changeover time of the production plan, provide the changeover time of the production plan to a user, and then implement the changeover time of the production plan, the changeover time comprising a product changeover, bottle changeover, type changeover, format changeover and/or pack changeover in the production of the products, wherein the system being configured to optimize the changeover time by:

breaking down the production plan into a plurality of contiguous units, each of the units corresponding to at least one product, processing step, or at least one device of a production line for producing a product of the production plan; and

arranging the units so that the changeover time has a minimum value; and

implementing optimized production costs of the production plan to be spent on processing the units, wherein the production costs are determined by retrofitting times and production times for the units, wherein the retrofitting times and production times are defined as product- and device-dependent, and wherein implementing the optimized production costs comprises:

assigning the units to a respective production line, and

determining an order in which the units are to be processed by the respective production line and processing the units in the determined order;

wherein the changeover time and/or the production cost additionally comprise a duration of a maintenance action for a detected malfunction of the device of the respective production line.

14 . The system of claim 13 , the system being further for automatically detecting the malfunction of the device in the food industry or the beverage industry, the system comprising:

a digital twin of the device, the digital twin configured to:

generate simulation data of the device, the simulation data generated for a plurality of malfunctions of the digital twin; and

a machine learning algorithm, the machine learning algorithm trained using the simulation data and the plurality of malfunctions of the digital twin, the machine learning algorithm configured to:

detect a malfunction of the device using operating data of the device; and

wherein the machine learning algorithm is executed by one of: a controller of the device, a processor connected to the device, and a cloud;

wherein detecting the malfunction comprises detecting an expected time period until a failure occurs that is caused by the malfunction.

15 . The system of claim 14 , wherein the detected malfunction of the device is an event, a failure, and/or a maintenance action to be taken, wherein:

the event comprises increased friction, increased vibration, imbalance, a wear parameter, or decreased lubricity,

the failure comprises an engine failure, a bearing failure, an axle failure, a crack, an incorrectly positioned bottle, a short circuit, or a stoppage of the device, and

the maintenance action to be taken comprises cleaning, lubricating, oiling, fixing, adjusting, replacing, or repairing; and/or

wherein detecting the malfunction further comprises detecting a component of the device on which the malfunction occurs.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (28)