Predicting early warnings of an operating mode of equipment in industry plants

Patent metadata

Field	Value
Publication number	US-11500368-B2
Application number	US-202017060688-A
Country	US
Kind code	B2
Filing date	Oct 1, 2020
Priority date	May 21, 2020
Publication date	Nov 15, 2022
Grant date	Nov 15, 2022

Abstract

Currently solutions for early detection of failures in manufacturing utilize predefined threshold levels of the process variables associated with equipment in manufacturing unit/industry plants. The pre-defined threshold and levels thereof are compared with the real values obtained from the manufacturing unit to check behavior of process variables (also referred as ‘process parameters’) and thus are prone to error. The present disclosure provides systems and method for predicting early warning of operating mode of equipment operating in industry plants which is based on transforming conditions on process parameters into conditions on corresponding fuzzy indices based on their thresholds. The fuzzy indices (concordance index, discordance index) of individual conditions are combined into a composite fuzzy index (composite index or degree of credibility) that describes the failure scenario in the process parameter space. A fuzzy logic-based detection is useful for detecting a failure mode early and providing alerts to operators for necessary action.

First claim

What is claimed is:

1. A processor implemented method for predicting early warnings of an operating mode of equipment in an industry plant, the method comprising:
learning, via one or more hardware processors, historical pattern based on implementation of at least one of artificial intelligence and machine learning models, wherein the historical pattern is related to operating modes and a failure prediction of historical operating modes associated with the equipment operating in the industry plant;
setting in a memory, via the one or more hardware processors, a pre-defined threshold for each process that is run on the equipment, based on the learned historical pattern by at least one of artificial intelligence and machine learning models;
obtaining, via the one or more hardware processors, time-series data from sensors associated with one or more equipment operating in the industry plant, wherein the time-series data pertains to one or more process parameters associated with one or more processes running on the equipment operating in the industry plant;
pre-processing, via the one or more hardware processors, the obtained time-series data to obtain normalized time-series data, wherein the pre-processing is applied on the obtained time-series data to remove a noise, high negative values and any unusual readings and impute missing data;
deriving, via the one or more hardware processors, one or more limits from at least one of the normalized time-series data and a data sheet obtained that is specific to the industry plant;
computing, via the one or more hardware processors, a concordance index (CI) for the one or more process parameters obtained from the sensors using a first subset of the one or more derived limits, wherein the CI is a fuzzy index defined such that the CI is “0” beyond one or more strict limits and the CI is “1” within one or more indifference limits;
computing, via the one or more hardware processors, a discordance index (DI) for the one or more process parameters using a second subset of the one or more derived limits, wherein the discordance index is a fuzzy index defined such that the DI is “0” within the one or more strict limits and the DI is “1” beyond one or more veto limits;
determining, via the one or more hardware processors, a weight of the one or more process parameters based on training of data-based models, wherein the data-based models are built using historical data learnt over a time period;
computing, via the one or more hardware processors, a degree of credibility (DOC) for the one or more process parameters based on a computed global concordance index and the computed discordance index for each of the one or more process parameters;
predicting, via the one or more hardware processors, an early warning specific to at least one operating mode of the one or more equipment operating in the industry plant, based on a value of the DOC of the one or more process parameters and the set pre-defined threshold, wherein the early warning corresponds to predicting a failure in advance, that is likely to occur in (i) one or more processes running in equipment, and (ii) in the behavior of the equipment operating in the industry plant;
suggesting, via the one or more hardware processors, corrective actions to be taken to recover from the predicted failure; and
driving, via the one or more hardware processors, inputs corresponding to the corrective actions based on the learned historical pattern.

2. The processor implemented method as claimed in claim 1 , wherein the one or more derived limits comprise the one or more veto limits, the one or more strict limits, and the one or more indifference limits.

3. The processor implemented method as claimed in claim 1 , wherein the one or more derived limits are specific to the one or more process parameters of one or more processes being run in the industry plant.

4. The processor implemented method as claimed in claim 1 , wherein the first subset of the one or more derived limits comprises the one or more strict limits and the one or more indifference limits.

5. The processor implemented method as claimed in claim 1 , wherein the second subset of the one or more derived limits comprises the one or more strict limits and the one or more veto limits.

6. The processor implemented method as claimed in claim 1 , wherein the computed global concordance index is based on the computed concordance index (CI) and the determined weight of the one or more process parameters.

7. The processor implemented method as claimed in claim 1 , wherein the early warning is predicted based on a comparison of (i) the value of the DOC of the one or more process parameters and (ii) the set pre-defined threshold.

8. The processor implemented method as claimed in claim 1 , wherein the at least one operating mode comprises a failure mode or a normal operation mode.

9. A system for predicting early warnings of an operating mode of equipment in an industry plant, the system comprising:
a memory storing instructions;
one or more communication interfaces; and
one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to:
learn historical pattern based on implementation of at least one of artificial intelligence and machine learning models, wherein the historical pattern is related to operating modes and a failure prediction of historical operating modes associated with the equipment operating in the industry plant;
set in the memory a pre-defined threshold for each process that is run on the equipment, based on the learned historical pattern by at least one of artificial intelligence or machine learning models;
obtain time-series data from sensors associated with one or more equipment operating in the industry plant, wherein the time-series data pertains to one or more process parameters associated with one or more processes running on the equipment operating in the industry plant;
pre-process the obtained time-series data to obtain normalized time-series data wherein the pre-processing is applied on the obtained time-series data to remove a noise, one of high negative values or any unusual readings, and impute missing data;
derive one or more limits from at least one of the normalized time-series data and a data sheet obtained that is specific to the industry plant;
compute a concordance index (CI) for one or more process parameters obtained from the sensors using a first subset of the one or more derived limits, wherein the CI is a fuzzy index defined such that the CI is “0” beyond one or more strict limits and the CI is “1” within one or more indifference limits;
compute a discordance index (DI) for the one or more process parameters using a second subset of the one or more derived limits, wherein the DI is a fuzzy index defined such that the DI is “0” within the one or more strict limits and the DI is “1” beyond one or more veto limits;
determine a weight of the one or more process parameters based on training of data-based models, wherein the data-based models are built using historical data learnt over a time period;
compute a degree of credibility (DOC) for the one or more process parameters based on a computed global concordance index and the computed discordance index for each of the one or more process parameters;
predict an early warning specific to at least one operating mode of the one or more equipment operating in the industry plant based on a value of the DOC of the one or more process parameters and the set pre-defined threshold, wherein the early warning corresponds to predicting a failure in advance, that is likely to occur in (i) one or more processes runni

Assignees

• Tata Consultancy Services Ltd

Inventors

• Gaduparthi Trinath
• Basavarsu Purushottham Gautham
• Tambe Yogesh
• Nirgudkar Himanshu

Classifications

• G05B23/0235

  based on a comparison with predetermined threshold or range, e.g. "classical methods", carried out during normal operation; threshold adaptation or choice; when or how to compare with the threshold · CPC title

• G05B23/0205

  by means of a monitoring system capable of detecting and responding to faults · CPC title

• G05B23/0254

  based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks · CPC title

• G05B23/0278 Primary

  Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA · CPC title

• G05B23/0221

  Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods · CPC title