Process control system using spatially dependent data for controlling a web-based process

• US 7,882,049 B2
• Filed: 08/19/2008
• Issued: 02/01/2011
• Est. Priority Date: 05/13/2005
• Status: Active Grant

• Current Assignee: Rockwell Automation Technologies Incorporated (Allen-Bradley Co., Inc.)
• Original Assignee: Rockwell Automation Technologies Incorporated (Allen-Bradley Co., Inc.)
• Inventors: Collette, L. Paul III, Johnson, W. Douglas
• Primary Examiner(s): Holmes; Michael B.
• Assistant Examiner(s): Kennedy; Adrian L
• Application Number:
  US12/194,440
• Publication Number:
  US 20080300709A1
• Time in Patent Office: 896 Days
• Field of Search: None
• US Class Current: 706/15
• CPC Class Codes: G05B 13/027 using neural networks only
• CPC Class Codes: G05B 2219/32193 Ann, neural base quality management
• CPC Class Codes: G05B 2219/33021 Connect plural macrocircuits, neural network modules in a larger network
• CPC Class Codes: G05B 2219/33027 Artificial neural network controller
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 3/105 Shells for specifying net layout
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

System and method for controlling a process with spatially dependent conditions for producing a product with spatially dependent properties, e.g., a web/sheet-based process for producing a web/sheet-based product. Input data comprising a plurality of input data sets are provided to a neural network (analog or computer-based), each data set comprising values for one or more input parameters, each comprising a respective process condition or product property. The input data preserve spatial relationships of the input data. The neural network generates output data in accordance with the input data, the output data comprising a plurality of output data sets, each comprising values for one or more output parameters, each comprising a predicted process condition or product property. The output data preserve spatial relationships of the output data, which correspond to the spatial relationships of the input data. The output data are useable by a controller or operator to control the process.

Claims

1. A method for controlling a process with spatially dependent conditions for producing a product with spatially dependent properties, comprising:

synthesizing, via a machine learning method, one or more process conditions at each of one or more spatial positions in a production line of a process or synthesizing one or more product properties at each of one or more spatial positions on the product to generate input data, wherein synthesizing comprises generating additional input data using nonlinear models based on historical spatial relationship data;

providing the input data to a process control system, wherein the input data comprise a plurality of input data sets, each input data set comprising values for a set of one or more input parameters; and

generating output data in accordance with the input data, wherein the output data comprise a plurality of output data sets, each output data set comprising values for a set of one or more output parameters, each output parameter comprising a predicted process condition or product property, wherein the output data preserve spatial relationships of the output data corresponding to the spatial relationships of the input data, and wherein the output data are useable by the process control system or an operator to control the process.

2. The method of claim 1, wherein providing the input data comprises:

measuring one or more process conditions at each of one or more spatial positions in a production line of the process to generate the input data;

or

measuring one or more product properties at each of one or more spatial positions on the product to generate the input data.

3. The method of claim 1, comprising training the process control system using synthesized process condition data having spatial relationships corresponding to the spatial relationships of the input data.

4. The method of claim 1, comprising pre-processing the input data prior to providing the input data to the process control system.

5. The method of claim 4, wherein pre-processing the input data comprises one or more of:

removing and/or replacing unusable or invalid data with valid data;

or

putting the data into a more usable form or format.

6. The method of claim 1, wherein the process control system comprises a plurality of sub-networks.

7. The method of claim 6, wherein each of the plurality of sub-networks corresponds to a respective data set in the plurality of data sets and each input parameter in each data set corresponds to a respective input node of a sub-network.

8. The method of claim 1, wherein synthesizing comprises one or more of:

interpolating measured data;

or

extrapolating measured data.

9. The method of claim 1, comprising:

providing the output data to an operator; and

controlling an actuator array via manual input by the operator to change a controllable process state in accordance with the output data, wherein the actuator array comprises a plurality of actuators, and wherein the plurality of actuators have spatial relationships corresponding to the spatial relationships of the input data.

10. The method of claim 1, comprising:

providing the output data to a controller;

controlling an actuator array via the controller to change a controllable process state in accordance with the controller output data, wherein the actuator array comprises a plurality of actuators, and wherein the plurality of actuators have spatial relationships corresponding to the spatial relationships of the input data.

11. The method of claim 10, comprising performing the steps of the method in an iterative manner to control production of the product with desired properties.

12. The method of claim 11, comprising periodically updating the process control system based on measured process conditions and product properties.

13. The method of claim 1,

wherein each input data set comprises position information for the set of input parameters, and wherein the position information indicates the spatial relationships of the input data;

and

wherein each output data set comprises position information for the one or more output parameters.

14. The method of claim 1,

wherein the input data are provided to inputs of the process control system in a manner that preserves the spatial relationships of the input data;

wherein the output data are provided by outputs of the process control system in a manner that preserves spatial relationships of the output data; and

wherein the spatial relationships of the output data correspond to the spatial relationships of the input data.

15. A method for controlling a web-based manufacturing process with spatially dependent conditions for producing a web product with spatially dependent properties, comprising:

synthesizing, via a machine learning method, one or more process conditions at each of one or more spatial positions in a production line of a web-based manufacturing process or synthesizing one or more web product properties at each of one or more spatial positions on the web product to generate input data, wherein synthesizing comprises generating additional input data using nonlinear models based on historical spatial relationship data;

providing the input data to a process control system, wherein the input data comprise a plurality of input data sets, each input data set comprising values for a set of one or more input parameters; and

generating output data in accordance with the input data, wherein the output data comprise a plurality of output data sets, each output data set comprising values for a set of one or more output parameters, each output parameter comprising a predicted process condition or web product property, wherein the output data preserve spatial relationships of the output data corresponding to the spatial relationships of the input data, and wherein the output data are useable by the process control system or an operator to control the web-based manufacturing process.

16. The method of claim 15, wherein providing the input data comprises:

measuring one or more process conditions at each of one or more spatial positions in a production line of the web-based manufacturing process to generate the input data;

or

measuring one or more web product properties at each of one or more spatial positions on the web product to generate the input data.

17. The method of claim 15, comprising training the process control system using synthesized process condition data having spatial relationships corresponding to the spatial relationships of the input data.

18. The method of claim 15, comprising pre-processing the input data prior to providing the input data to the process control system.

19. The method of claim 18, wherein pre-processing the input data comprises one or more of:

removing and/or replacing unusable or invalid data with valid data;

or

putting the data into a more usable form or format.

20. A system for controlling a process with spatially dependent conditions for producing a product with spatially dependent properties, comprising:

a processor; and

a memory coupled to the processor that stores program instructions executable by the processor to implement a process control system, wherein the process control system comprises;

a plurality of inputs, operable to receive input data, wherein the input data comprise a plurality of input data sets, each comprising values for a set of one or more input parameters, and wherein the input data are provided by one or more of;

synthesizing, via a machine learning method, one or more process conditions at each of one or more spatial positions in a production line of the process to generate the input data, or synthesizing one or more product properties at each of one or more spatial positions on the product to generate the input data, wherein synthesizing comprises generating additional input data using nonlinear models based on historical spatial relationship data;

wherein the process control system is operable to generate output data in accordance with the input data, wherein the output data comprise a plurality of output data sets, each output data set comprising values for a set of one or more output parameters, each output parameter comprising a predicted process condition or product property, wherein the output data preserve spatial relationships of the output data corresponding to the spatial relationships of the input data.