US 2021/0080916 A1Published application

Feedback loop driven end-to-end state control of complex data-analytic systems

AssigneeAccenture Global Solutions Limited

InventorsJaeyoung Christopher Kang, Jürgen Albert Weichenberger, Teresa Sheausan Tung, William R. Gatehouse, Tiffany Cecilia Dharma, Jan Andre Nicholls

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

A system provides feedback driven end-to-end state control of a data model. A data model may be used to model the behavior of a petrochemical refinery to predict future events. The system may be used to ensure proper operation of the data model. Contingency data models may be executed when a failure is detected. Further, when the system detects accuracy that is out of tolerance, the system may initiate retraining of the data model being currently used.

Claims20

Dependent claims

A product comprising:

a machine-readable medium other than a transitory signal;
instructions stored on the machine-readable medium, the instructions, when executed, configured to cause a processor to:
access a dependency matrix to determine a mandatory execution criterion for a primary data model, the primary data model, when executed, configured to
cause the processor to:
receive current sensor data regarding a current operational state of a petrochemical refinery;
responsive to the current sensor data, predict a probability of a foaming event occurring within a pre-defined time-window;
and
cause an injection of an anti-foaming agent into the petrochemical refinery when the probability exceeds a threshold;
responsive to the dependency matrix, determine whether the mandatory execution criterion for the primary model is met;
when the mandatory execution criterion is not met:
cancel an execution of the primary model;
generate a prompt indicating that execution of the primary model has been cancelled;
execute a contingency data model, the contingency data model configured to
receive the current sensor data regarding the current operational state of the petrochemical refinery;
responsive to the current sensor data, predict a future operational state of the petrochemical refinery to determine the probability of the foaming event occurring within the pre-defined time-window;
cause the injection of the anti-foaming agent into the petrochemical refinery when the probability exceeds the threshold;
and
generate contingency output based on the future operational state;
pass the contingency output to accuracy analysis circuitry;
receive an indication of whether to retrain the contingency data model;
when the accuracy analysis circuitry determines to retrain the contingency data model:
retrain the contingency by adjusting an operational parameter for the contingency data model;
and
after retraining, re-launch the contingency data model;
and
when mandatory execution criterion is met, execute the primary data model.

23 more elements

A method comprising:

accessing a dependency matrix to determine a mandatory execution criterion for a primary data model, where executing the primary data model includes:
receiving current sensor data regarding a current operational state of a petrochemical refinery system;
responsive to the current sensor data, predicting a probability of a foaming event occurring within a pre-defined time-window;
and
causing an injection of an anti-foaming agent into the petrochemical refinery when the probability exceeds a threshold;
responsive to the dependency matrix, determining whether the mandatory execution criterion for the primary model is met;
when the mandatory execution criterion is not met:
cancelling an execution of the primary model;
generating a prompt indicating that execution of the primary model has been cancelled;
executing a contingency data model by:
receiving the current sensor data regarding the current operational state of the petrochemical refinery;
responsive to the current sensor data, predicting a future operational state of the petrochemical refinery to determine the probability of the foaming event occurring within the pre-defined time-window;
causing the injection of the anti-foaming agent into the petrochemical refinery when the probability exceeds the threshold;
and
generating contingency output based on the future operational state;
passing the contingency output to accuracy analysis circuitry;
receiving an indication of whether to retrain the contingency data model;
when the accuracy analysis circuitry determines to retrain the contingency data model:
retraining the contingency by adjusting an operational parameter for the contingency data model;
and
after retraining, re-launch the contingency data model;
and
when mandatory execution criterion is met, executing the primary data model.

20 more elements

A system comprising:

a validation memory comprising a dependency matrix for a primary data model, the dependency matrix including an indicator of a mandatory execution criterion for the primary data model;
and
model execution circuitry in data communication with the validation memory, the model execution circuitry configured to:
access the dependency matrix to determine the mandatory execution criterion for a primary data model, the primary data model, when executed, configured to cause the processor to:
receive current sensor data regarding a current operational state of a petrochemical refinery;
responsive to the current sensor data, predict a probability of a foaming event occurring within a pre-defined time-window;
and
cause an injection of an anti-foaming agent into the petrochemical refinery when the probability exceeds a threshold;
responsive to the dependency matrix, determine whether the mandatory execution criterion for the primary model is met;
when the mandatory execution criterion is not met:
cancel an execution of the primary model;
generate a prompt indicating that execution of the primary model has been cancelled;
execute a contingency data model, the contingency data model configured to
receive the current sensor data regarding the current operational state of the petrochemical refinery;
responsive to the current sensor data, predict a future operational state of the petrochemical refinery to determine the probability of the foaming event occurring within the pre-defined time-window;
cause the injection of the anti-foaming agent into the petrochemical refinery when the probability exceeds the threshold;
and
generate contingency output based on the future operational state;
pass the contingency output to accuracy analysis circuitry;
receive an indication of whether to retrain the contingency data model;
when the accuracy analysis circuitry determines to retrain the contingency data model:
retrain the contingency by adjusting an operational parameter for the contingency data model;
and
after retraining, re-launch the contingency data model;
and
when mandatory execution criterion is met, execute the primary data model.

23 more elements

Dependencies are inferred from the claim text.

Facts

Application: 17/104,382Filed 25 Nov 2020
Priority: 27 Jul 2016
Published: 18 Mar 2021
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

G05B 17/02
Systems involving the use of models or simulators of said systems: electric
B01D 19/0063
Degasification of liquids: regulation, control including valves and floats
B01D 19/04
Foam dispersion or prevention: by addition of chemical substances
G05B 13/021
Not using a model or a simulator of the controlled system: in which a variable is automatically adjusted to optimise the performance

Family

Application62/367,454Us provisional application · filed 27 Jul 2016
Application15/445,077Division · filed 28 Feb 2017
Application17/104,382Division
US 10,871,753 B2Related patent
US 11,846,921 B2Related patent
US 2018/0032038 A1Published application
US 2021/0080916 A1This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key