Method for computer-aided processing of state messages in an automation installation

Patent metadata

Field	Value
Publication number	US-11243519-B2
Application number	US-202016861385-A
Country	US
Kind code	B2
Filing date	Apr 29, 2020
Priority date	May 2, 2019
Publication date	Feb 8, 2022
Grant date	Feb 8, 2022

Abstract

A method for computer-aided processing of state messages in an automation installation, wherein state messages are generated by components and detected with their generation points in time, where causative states present at the generation point of the state message or beforehand in other components are determined for a multiplicity of state messages of a respective component and the current state in the generated state message, where the propagation time between occurrence of the respective causative state and the generation point of the state message is calculated for each causative state, where groups are formed from the causative states, where in a respective group all causative states have at least the common feature that they were determined for the same current state in the respective component, and where at least one statistical parameter is determined from the propagation times which belong to the causative states of the same group and stored.

First claim

What is claimed is:

1. A method for computer-aided processing of state messages in an automation installation, the state messages being generated by a multiplicity of components during performance of an automated process in the automation installation and being detected with their generation points in time, a state message being generated by a respective component upon a change from a preceding to a new state and indicating the new state, a pattern description being provided for a respective component of at least one portion of components of the automation installation, said pattern description indicating for at least one state in the respective component in at least one causative state each allocated to a corresponding state in the respective component, each causative state belonging to a different component than the respective component and can be a cause of that state in the respective component to which the causative state is allocated, the method for the respective component comprising:
a) determining each causative state allocated to a current state which is present at a generation point in time of the state message in other components for a multiplicity of state messages generated by the respective component, based on the pattern description for the current state in the generated state message, and determining those causative states allocated to the current state from which a change was effected upon a last state change which occurred in a respective other component before the generation point in time of the state message, a propagation time between an occurrence of a respective causative state and the generation point in time of the state message being calculated for each causative state;
b) forming groups from the causative states which were determined for the respective component in step a), in a respective group all causative states having at least a common feature which said causative states determined for the same current state in the respective component; and
c) determining at least one statistical parameter from the propagation times which belong to the causative states of the same group and storing said determined at least one statistical parameter.

2. The method as claimed in claim 1 , further comprising for the respective component:
d) determining each causative state allocated to the current state from which a change was effected upon the last state change which occurred in another respective component before the generation point in time of the state message for a plurality of state messages generated by the respective component, based on the pattern description for the current state in the generated state message, the further propagation time between an end of the respective causative state and the generation point in time of the state message being calculated for each causative state;
e) forming further groups from the causative states which were determined for the respective component in step d), in a respective further group all causative states having at least a common feature that said causative states were determined for the same current state in the respective component; and
f) determining at least one statistical parameter from the further propagation times which belong to the causative states of the same further group.

3. The method as claimed in claim 1 , wherein the at least one statistical parameter comprises a minimum value and a maximum value of the propagation times in a respective group.

4. The method as claimed in claim 2 , wherein the at least one statistical parameter comprises a minimum value and a maximum value of the propagation times in a respective group.

5. The method as claimed in claim 1 , wherein the at least one statistical parameter comprises a frequency distribution of the propagation times in a respective group.

6. The method as claimed in claim 5 , wherein at least one of (i) the at least one statistical parameter comprises a first quantile value, in accordance with which the propagation times of a predetermined percentage proportion of the frequency distribution lie below a first quantile value and (ii) the at least one statistical parameter comprises a second quantile value, in accordance with which the propagation times of a predefined percentage proportion of the frequency distribution lie above a second quantile value.

7. The method as claimed in claim 5 , wherein the frequency distribution is approximated with a gaussian distribution and a mean value and a standard deviation of the gaussian distribution are determined as statistical parameters.

8. The method as claimed in claim 6 , wherein the frequency distribution is approximated with a gaussian distribution and a mean value and a standard deviation of the gaussian distribution are determined as statistical parameters.

9. The method as claimed in claim 1 , wherein during at least one of step a) and step d) only such causative states are determined which precede the generation point in time of the state message by less than a predetermined time threshold.

10. The method as claimed in claim 1 , wherein up to a predefined point in time after a beginning of an automated process, the at least one statistical parameter determined in at least one of step c) and step f) is combined with statistical parameters which were previously determined for a different automated process of the automation installation;
wherein the combination of the at least one statistical parameter determined in at least one of said step c) and step f) and statistical parameters which were previously determined for the different automated process of the automation installation is stored; and
wherein the combination of the at least one statistical parameter determined in at least one of said step c) and step f) and statistical parameters which were previously determined for the different automated process of the automation installation comprises a weighted sum and a weighting of the previously determined statistical parameters in the combination of the at least one statistical parameter determined in at least one of said step c) and step f) and statistical parameters which were previously determined for the different automated process of the automation installation decreases with increasing reduction of a temporal distance with respect to the predefined point in time.

11. The method as claimed in claim 1 , wherein a temporal profile of state changes in the components of the automation installation, before performing at least one of step a) and d), is pre-processed such that states which conceal whether a state present before a last state change is still present in a corresponding component are removed from the temporal profile.

12. The method as claimed in claim 1 , wherein the automation installation is an installation for at least one of producing a product and processing the product.

13. The method as claimed in claim 1 , wherein the installation for at least one of producing a product and processing the product comprises at least one of a filling and packaging installation.

14. A device for computer-aided processing of state messages in an automation installation, comprising:
a processor including memory;
wherein the state messages are generated by a plurality of components during performance of an automated process in the automation installation and are detected with generation points in time of the state messages;
wherein a state message is generated by a respective component upon a change from a preceding state to a new state and indicates the new state;
wherein a pattern description is provided for a respective component of at least one portion of the component

Assignees

• Siemens Ag

Inventors

• Gross Ralf
• Klos Hans-Henning
• Kohler Benjamin
• Riedl Wolfgang
• Schnittger Jens

Classifications

• G05B19/4183

  characterised by data acquisition, e.g. workpiece identification · CPC title

• G06F17/18

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G05B23/0218 Primary

  characterised by the fault detection method dealing with either existing or incipient faults · CPC title

• G05B23/0243 Primary

  model based detection method, e.g. first-principles knowledge model · CPC title

• G05B2219/24065

  Real time diagnostics · CPC title