Online frequently derived measurements for process monitoring, control and optimization
AssigneeABB Schweiz AG
InventorsChangYuan Liu, Donald Ray Stanley, III, Andrew Daniel Steiger
Life of this patent
Estimating the term…
Abstract
A system, method, and/or apparatus is provided for production processes in which online frequently derived measurements are determined use existing online and/or offline reference measurements and real-time and/or historical data to model process variables and/or product properties to achieve enhanced production goals.
Claims22
A method of frequently deriving online measurements associated with a production process for a product, the method comprising:
- obtaining a plurality of process variables associated with the production process and product properties associated with the product;
- determining a plurality of intermediate variables based on the plurality of process variables and/or product properties;
- obtaining one or more reference measurements from the production process and/or the product that are associated with the plurality of intermediate variables, wherein each of the plurality of intermediate variables are determined by combining one or more process variables from process data and one or more properties of the product from the one or more reference measurements;
- modelling a relationship between the one or more reference measurements and the plurality of intermediate variables;
- repeatedly deriving the online measurements based on the modelled relationship;
- controlling the production process using the derived online measurements;
- determining at least one of the one or more reference measurements is valid;
- and
- auto-tracking at least one of the derived online measurements to the at least one valid reference measurement, wherein the auto-tracking is performed using a regression function to determine and correct a relationship between the derived online measurements to the at least one valid reference measurement.
A computer system configured to frequently derive online measurements associated with a production process for a product, the computer system including at least one hardware computer processor that operates during the production process to:
- obtain a plurality of process variables associated with the production process and a plurality of product properties associated with the product;
- determine a plurality of intermediate variables based on the plurality of process variables and/or the plurality of product properties;
- obtain one or more reference measurements from the production process and/or the product that are associated with the plurality of intermediate variables, wherein each of the plurality of intermediate variables are determined by combining one or more process variables from process data and one or more properties of the product from the one or more reference measurements;
- model a relationship between the one or more reference measurements and the plurality of intermediate variables;
- repeatedly derive online measurements based on the modelled relationship;
- controlling the production process using the derived online measurements;
- determine at least one of the one or more reference measurements is valid;
- and
- auto-track at least one of the derived online measurements to the one or more valid reference measurements, wherein the auto-tracking is performed using a regression function to determine and correct a relationship between the derived online measurements to the at least one valid reference measurement.
Dependencies are inferred from the claim text.