# US 10,481,581 B2 · Diagnosis facility and diagnostic method for monitoring performance of a technical plant

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,481,581 B2
- Document ID: US10481581B2
- Status: Granted
- Kind: B2
- Application number: 15/482,997
- Filed: 2017-04-10
- Published: 2017-10-26
- Granted: 2019-11-19
- Earliest claimed: 2016-04-12
- Assignees: Siemens Aktiengesellschaft
- Applicants: Siemens Aktiengesellschaft
- Inventors: Thomas Bierweiler; Daniel Labisch
- Web page: https://patentgrep.com/patent/US10481581B2

## CPC

- G05B 19/4063: Monitoring general control system (G05B19/4062 takes precedence)
- G05B 19/0425: {Safety, monitoring}
- G05B 19/41805: {characterised by assembly}
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G06F 30/3312: Timing analysis
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A diagnosis facility and diagnostic method for monitoring the performance of a technical plant with an automation system. In the performance of a plant with step sequences that run repeatedly, deviations of the time response in a cycle of the step sequence that is to be checked from the time response for fault-free cycles are detected and displayed by evaluation of the data set, using the self-organizing map. To this end, the durations of the execution of each step in the step sequence that is to be checked are determined and evaluated using a predetermined self-organizing map that has been learned using fault-free cycles. This type of evaluation has the advantage that the self-organizing map can be learned automatically and consequently hardly any knowledge of the respective process that is running on the plant is required for the diagnosis.

## Claims

### Claim 1

A diagnosis facility for monitoring performance of a technical plant with an automation system, comprising:
a memory in which at least one data set characterizing the performance of the technical plant is storable; and
an evaluation unit via which a diagnostic report on the performance of the technical plant is at least one of (i) determinable and (ii) displayable using the at least data set and a predetermined self-organizing map;
wherein the evaluation unit is configured to determine, by during the performance of the technical plant comprising step sequences that run repeatedly, a duration of an execution of each step of a cycle that is to be checked in a step sequence and to store said duration as a coordinate of the at least data set, the at least data set comprising an n-tuple having a predetermined number n of coordinates;
wherein the evaluation unit is further configured to detect and display a deviation of a time response in the cycle of the step sequence that is to be checked from a time response for fault-free cycles by evaluating the at least data set, using the self-organizing map; and
wherein corresponding n-tuples with predetermined durations for fault-free runs of the step sequence are stored on nodes of the self-organizing map.

### Claim 2 (depends on 1)

The diagnosis facility as claimed in claim 1, wherein the evaluation unit is further configured to determine a minimum Cartesian distance between the at least data set for the cycle that is to be checked and the nodes of the self-organizing map to detect a deviation in the time response.

### Claim 3 (depends on 2)

The diagnosis facility as claimed in claim 2, wherein the evaluation unit is further configured to display, as a diagnostic report, a deviation from the normal performance in the cycle that is to be checked in the step sequence when the Cartesian distance exceeds a predetermined threshold value.

### Claim 4 (depends on 1)

The diagnosis facility as claimed in claim 1, wherein the evaluation unit is further configured in order to predetermine the self-organizing map, to store in the memory, prior to a cycle that is to be checked, a plurality of data sets that characterize a fault-free performance of the plant, and utilized said plurality of data sets to calculate for the nodes, via a learning method, the n-tuples with durations of the execution of the steps for fault-free cycles and store said calculated n tuples on the nodes of the self-organizing map.

### Claim 5 (depends on 4)

The diagnosis facility as claimed in claim 4, wherein the evaluation unit is further configured, prior to a cycle that is to be re-checked, to store in the memory further data sets that characterize a fault-free performance of the plant, to determine a winner node for respective data sets, to calculate respective maximum distances between the further data sets and the winner node assigned thereto, and to store these, case increased by an added margin to avoid faulty diagnoses, as predetermined threshold values, each assigned to the winner node.

### Claim 6 (depends on 1)

The diagnosis facility as claimed in claim 1, wherein the evaluation unit is further configured, at a fork in a step sequence, due to which alternative steps of a step sequence are performable during a cycle, to set a value to “zero” as a duration for unperformed steps in a cycle that is to be checked and store said value for a data set coordinate provided for the unperformed step.

### Claim 7 (depends on 1)

The diagnosis facility as claimed in claim 1, wherein the evaluation unit is further configured, at a loop in a step sequence, due to which steps in a step sequence are performable a plurality of times in a cycle, during a learning of the self-organizing map using fault-free cycles for each execution of steps of a step sequence in a cycle, to set a coordinate of the n-tuple and to set a value “zero” as a duration of unperformed loop steps in a cycle of a sequence of steps that is to be checked and store said value for a coordinate of a data set that is assigned to an unperformed loop step.

### Claim 8 (depends on 1)

The diagnosis facility as claimed in claim 1, wherein at least the memory and the evaluation unit are implemented by software on a Remote Service computer for remote diagnosis of the technical plant.

### Claim 9

A diagnostic method for monitoring the performance of a technical plant, comprising:
storing at least one data set characterizing performance of the technical plant in a memory;
determining a diagnostic report on the performance of the technical plant using the at least one data set and a predetermined self-organizing map;
determining a duration of the execution of each step of a cycle that is to be checked in a step sequence during the performance of the technical plant with repeatedly run step sequences;
storing previously determined durations as a respective coordinate of the at least one data set during the performance of the technical plant with the repeatedly run step sequences, the at least one data set comprising an n-tuple with a predetermined number n of coordinates; and
detecting and displaying deviations of a time response in the cycle that is to be checked from fault-free cycles by evaluating the at least one data set during the performance of the technical plant with the repeatedly run step sequences, using the self-organizing map, n-tuples with the predetermined time responses for fault-free cycles of the step sequence being stored on the node of the self-organizing map.

### Claim 10

A non-transitory computer-readable medium encoded with a computer program executed by a computer that causes monitoring performance of a technical plant, comprising:
program code for storing at least one data set characterizing performance of the technical plant in a memory;
program code for determining a diagnostic report on the performance of the technical plant using the at least one data set and a predetermined self-organizing map;
program code for determining a duration of the execution of each step of a cycle that is to be checked in a step sequence during the performance of the technical plant with repeatedly run step sequences;
program code for storing previously determined durations as a respective coordinate of the at least one data set during the performance of the technical plant with the repeatedly run step sequences, the at least one data set comprising an n-tuple with a predetermined number n of coordinates; and
program code for detecting and displaying deviations of a time response in the cycle that is to be checked from fault-free cycles by evaluating the at least one data set during the performance of the technical plant with the repeatedly run step sequences, using the self-organizing map, n-tuples with the predetermined time responses for fault-free cycles of the step sequence being stored on the node of the self-organizing map.

### Claim 11 (depends on 10)

The non-transitory computer-readable medium of claim 10, wherein the computer program product comprises a storage medium.