	US 12,390,824 B2
	Monitoring method for an application plant and corresponding application plant
Tobias Winter, Bietigheim-Bissingen (DE); and Paul Thomä, Freiberg am Neckar (DE)
Appl. No. 17/609,429
Filed by Dürr Systems AG, Bietigheim-Bissingen (DE)
PCT Filed May 4, 2020, PCT No. PCT/EP2020/062242 § 371(c)(1), (2) Date Nov. 8, 2021, PCT Pub. No. WO2020/225175, PCT Pub. Date Nov. 12, 2020.
Claims priority of application No. 10 2019 112 099.3 (DE), filed on May 9, 2019.
Prior Publication US 2022/0219182 A1, Jul. 14, 2022
Int. Cl. B05B 12/00 (2018.01); G05B 19/418 (2006.01); G05B 23/02 (2006.01)

CPC B05B 12/006 (2013.01) [G05B 19/418 (2013.01); G05B 23/0221 (2013.01); G05B 23/024 (2013.01)]

12 Claims

1. A monitoring method for a painting plant for applying an application agent, comprising:

receiving first raw sensor data from a first sensor included in the painting plant representing a first operating variable of the painting plant;

receiving second raw sensor data from a second sensor included in the painting plant reflecting a second operating variable of the painting plant;

acquisition of first and second control signals for controlling the painting plant;

extracting features from the first raw sensor data and the second raw sensor data, the features comprising a reduced amount of data compared to the first raw sensor data and the second raw sensor data and outputting the features to a machine learning algorithm;

determining an operating state of the painting plant based on rules included in the machine learning algorithm, wherein the rules included in the machine learning algorithm are obtained by the following steps:

retrieving stored historical raw sensor data,

extracting features from the historical raw sensor data, and

determining the rules by means of the machine learning algorithm by evaluating the features determined from the historical raw sensor data and the associated labels;

wherein extracting features from the first raw sensor data and the second raw sensor data includes:

defining an observation period window for evaluating the first and/or second raw sensor data;

defining at least one comparison period within the observation period window;

subdividing the comparison period into several successive subsections;

calculating at least one statistical parameter for the first and/or second raw sensor data and/or further raw sensor data within the individual subsections, the statistical parameter being a component of the features;

determining the timing of changes of the first and/or second control signal within the observation time window; and

determining the at least one comparison period between two successive changes of the first and/or second control signals; and

controlling the painting plant based on the determined operating state of the painting plant including changing a setpoint value of at least one of the first and second control signals.