# US 2023/0053545 A1 · Industrial Plant Monitoring > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2023/0053545 A1 - Document ID: US20230053545A1 - Status: Published application - Kind: A1 - Application number: 17/796,698 - Filed: 2021-01-29 - Published: 2023-02-23 - Earliest claimed: 2020-02-04 - Assignees: BASF SE - Applicants: BASF SE - Inventors: Nikolaos Fanidakis; Claus-Juergen Neumann; Benjamin Priese; Frank Strohmaier; Norman Volkert; Thomas Christ; Torsten Norbert Kneitz; Alexander Kubisch - Web page: https://patentgrep.com/patent/US20230053545A1 ## CPC - G05B 19/41875: {characterised by quality surveillance of production} - 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} - G05B 23/0281: {Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis} - G05B 2219/37591: Plant characteristics - Y02P 80/30: Reducing waste in manufacturing processes; Calculations of released waste quantities - Y02P 90/80: Management or planning - 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) - Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS ## Abstract The present teachings relate to a method comprising a plurality of sensors, and one or more functionally connected processing units, the method comprising: providing, at any of the one or more processing units, time-series residual data of a sensor object; the sensor object being a group of at least some of the sensors from the plurality of sensors, and wherein the residual data comprises, for each of the sensors of the sensor object, a residue signal which is a difference between the sensor's measured output and the sensor's expected output, monitoring, via any of the one or more processing units, a level signal; wherein the level signal is indicative of a collective time-based variation of the time-series residual data, monitoring, via any of the one or more processing units, an association signal; wherein the association signal is indicative of the variation and/or association structure of the time-series residual data, generating, via any of the one or more processing units, an anomaly event signal when at a given time a value of the level signal and/or a value of the association signal changes from an expected value of the respective signal at or around that time. The present teachings also relate to a monitoring and/or control system for a plant comprising a plurality of sensors, wherein the system comprises one or more processing units configured to perform the method steps of any of the steps herein disclosed, and a computer software product. ## Claims ### Claim 1 A method for monitoring a plant comprising a plurality of sensors, and one or more functionally connected processing units, the method comprising: providing, at any of the one or more processing units, time-series residual data of a sensor object; the sensor object being a group of at least some of the sensors from the plurality of sensors, and wherein the residual data comprises, for each of the sensors of the sensor object, a residue signal which is a difference between the sensor's measured output and the sensor's expected output, monitoring, via any of the one or more processing units, a level signal; wherein the level signal is indicative of a collective time-based variation of the time-series residual data, monitoring, via any of the one or more processing units, an association signal; wherein the association signal is indicative of the variation and/or association structure of the time-series residual data, generating, via any of the one or more processing units, an anomaly event signal when at a given time a value of the level signal and/or a value of the association signal changes from an expected value of the respective signal at or around that time. ### Claim 2 (depends on 1) The method according to claim 1, wherein the any of the respective expected value is provided as a corresponding expected value limit specifying for a given time a plurality of expected values, as a range or as discrete values, that the corresponding signal may validly have without the anomaly event being generated. ### Claim 3 (depends on 1) The method according to claim 1, wherein the method further comprises: determining, in response to the anomaly event signal, at least one root cause of the anomaly, by performing any one or more of the: checking for which of the sensors in the sensor object the sensor's measured output changed from the sensor's expected output at or around the same time as the occurrence of the anomaly event; analyzing the time series residual signal of each sensor within the sensor object to determine one or more main drivers or most dominant contributors to the level signal value; analyzing the time series residual signal of each sensor within the sensor object to determine one or more main drivers or most dominant contributors to the association signal value; and analyzing covariance of time series residual signals of each pair combination of the sensor residual signals within the sensor object to determine one or more main drivers or most dominant contributors to the association signal value. ### Claim 4 (depends on 1) The method according to claim 1, wherein the method further comprises: determining, in response to the anomaly event signal, state of health of at least one equipment related to the sensor object. ### Claim 5 (depends on 2) The method according to claim 2, wherein any of the expected value or the expected limit value is provided by a sensor object model that is at least partially a predictive model trained using historical residual data of the sensor object. ### Claim 6 (depends on 5) The method according to claim 5, wherein one or more covariate signals are provided as input to the sensor object model, each covariate signal being a signal representing a parameter upon which at least one of the residual signals are dependent upon. ### Claim 7 (depends on 1) The method according to claim 1, wherein at least one sensor expected output is provided by an expected state model that is a predictive model trained using historical time-series output data of the respective sensor. ### Claim 8 (depends on 1) The method according to claim 1, wherein the sensor object is provided by at least partially automatically grouping the at least some of the sensors using at least one data-centric algorithm. ### Claim 9 (depends on 1) The method according to claim 1, wherein the sensor object is at least partially automatically generated by any of the one or more processing units using at least one self-organizing map. ### Claim 10 (depends on 7) The method according to claim 7, wherein the expected state model is selected automatically by the processing unit by analyzing a plurality of different predictive model types, and selecting the model type as the expected state model which provides the lowest error between: the output of that model when trained with a specific training window of the historical time-series data; and the actual historical sensor output within a specific time-window of the historical time-series data. ### Claim 11 (depends on 1) The method according to claim 1, wherein the level signal value is generated using a distance estimator indicating the time at which and the amount by which the time-series residual data deviates from its normal or expected or mean state. ### Claim 12 (depends on 1) The method according to claim 1, wherein the association signal value is generated using a statistical measure of multivariate dependencies in the residual data, or to measure at a given time the dispersion of the time-series residual data. ### Claim 13 (depends on 1) The method according to claim 1, wherein method comprises: detecting, via any of the one or more processing units, a drift in the output signal a sensor; wherein the sensor is among the at least some of the sensors, and wherein the drift is computed from historical time series data of the sensor, the historical data of the sensor being from a time period that is at least 1 week long, and wherein the drift is detected by computing the strength, smoothness and currentness of the historical data of the sensor. ### Claim 14 (depends on 1) A monitoring and/or control system for a plant comprising a plurality of sensors, wherein the system comprises one or more processing units configured to perform the method steps of claim 1. ### Claim 15 (depends on 1) A computer program comprising instructions which, when executed by a processing unit of a plant monitoring and/or control system functionally connected to a plurality of sensors, cause the system to carry out the method steps of claim 1.