CLOUD-LEVEL CONTROL LOOP TUNING ANALYTICS

• US 20150277399A1
• Filed: 10/27/2014
• Published: 10/01/2015
• Est. Priority Date: 03/26/2014
• Status: Active Grant

• Current Assignee: Rockwell Automation Technologies Incorporated (Rockwell Automation, Inc.)
• Original Assignee: Rockwell Automation Technologies Incorporated (Rockwell Automation, Inc.)
• Inventors: Maturana, Francisco P., Asenjo, Juan L.
• Granted Patent: US 9,971,317 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: Y02P 90/80 Management or planning

Abstract

A control loop tuning system executing on a cloud platform facilitate remote control system analysis and generation of suitable controller gains for a given closed-loop control application. The system leverages cloud-side analytics and a gain correlation model generated based on historical data collected from the industrial control system and maintained on cloud storage. The gain correlation model creates a virtual association between controller gains and process variables based on operational and configuration data collected from the industrial control system. The system then applies iterative analytics to the model to converge on a set of controller gains determined to satisfy an optimization criterion. The recommended controller gains are then provided to a client device for review and implementation in the real system controller.

Claims

1. A system for tuning controller gain values on a cloud platform, comprising:

a memory that stores computer-executable components;

a processor, operatively coupled to the memory, that executes the computer-executable components, the computer-executable components comprising;

a system interface component configured to collect industrial data from a set of devices comprising an industrial control system and store the industrial data on a cloud platform;

a gain modeling component configured to generate a gain correlation model based on analysis of the industrial data and store the gain correlation model on the cloud platform, wherein the gain correlation model defines at least one correlation between controller gain values and process variables of the industrial control system; and

a correlation analytics component configured to determine at least one controller gain value for the industrial control system based on analysis of the gain correlation model.

2. The system of claim 1, further comprising a client interface component configured to send the at least one controller gain value to a client device communicatively connected to the cloud platform.

3. The system of claim 1, wherein the gain modeling component is configured to generate the gain correlation model based on multi-enterprise industrial data collected from multiple industrial control systems associated with multiple industrial enterprises.

4. The system of claim 3, wherein the gain modeling component is configured to generate the gain correlation model for the industrial control system based on a subset of the multi-enterprise industrial data determined to correspond to one or more other industrial control systems having a similarity to the industrial control system.

5. The system of claim 1, wherein the correlation analytics component is configured to simulate, based on the gain correlation model, a process variable response using at least one test value of the at least one controller gain.

6. The system of claim 5, wherein the correlation analytics component is configured to modify the at least one test value as a function of the process variable response based on one or more gain tuning rules to yield at least one modified test value, and to re-simulate the process variable response for the at least one modified test value to yield a new process variable response.

7. The system of claim 6, wherein the correlation analytics component is configured to iteratively modify the at least one test value to yield the at least one modified test value and re-simulate the process variable response for the at least one modified test value, and to select, as the at least one controller gain value, the at least one modified test value in response to a determination that the new process variable response for the at least one modified test value satisfies a defined criterion.

8. The system of claim 1, wherein the system interface component is configured to receive at least a portion of the industrial data as a data packet from a cloud agent device associated with one or more of the set of devices, wherein the data packet comprises header information identifying at least one of an owner of the industrial control system, a site at which the industrial control system is located, a priority of the portion of the industrial data, a message type of the data packet, or a process identifier specifying a type of processing to be performed on the portion of the industrial data.

9. The system of claim 1, wherein the system data comprises at least one of monitored process variable values for the industrial control system, position data or velocity data for a motion device, configuration information for one or more of the set of devices, device alarm data, system alarm data, machine cycle time data, a key performance indicator for the industrial control system, or firmware information for at least one of the set of devices.

10. A method for tuning an industrial control loop, comprising:

receiving, at a cloud platform by a system comprising at least one processor, industrial data from one or more industrial devices of an industrial control system;

creating, by the system, a gain correlation model based on analysis of the industrial data and storing the gain correlation model on cloud storage of the cloud platform, wherein the gain correlation model encodes at least one correlation between controller gain values and process variables of the industrial control system; and

determining, by the system, at least one controller gain value for the industrial control system based on an analysis of the gain correlation model.

11. The method of claim 10, further comprising sending the at least one controller gain value to a specified client device that has authorization to access the cloud platform.

12. The method of claim 10, wherein the creating comprises creating the gain correlation model based on multi-enterprise industrial data collected from multiple industrial control systems associated with multiple industrial enterprises.

13. The method of claim 12, wherein the creating comprises creating the gain correlation model for the industrial control system based on a selected subset of the multi-enterprise data corresponding to at least one other industrial control system determined to be similar to the industrial control system.

14. The method of claim 10, wherein the determining comprises simulating a process variable response based on the correlation gain model using at least one test value of the at least one controller gain.

15. The method of claim 14, wherein the determining further comprises:

modifying the at least one test value as a function of the process variable response based on one or more gain tuning rules resulting in at least one modified test value; and

repeating the simulating using the at least one modified test value resulting in a new process variable response.

16. The method of claim 15, further comprising:

repeating the modifying and the repeating until the new process variable response satisfies a defined criterion; and

setting, as the at least one controller value, the at least one modified test value that resulted in the new process variable response that satisfied the defined criterion.

17. The method of claim 10, wherein the receiving comprises:

receiving at least a portion of the industrial data as a data packet from a cloud agent device associated with one or more of the set of devices; and

processing the data packet in accordance with header information included in the data packet,

wherein the header information specifies at least one of an owner of the industrial control system, a site at which the industrial control system is located, a priority of the portion of the industrial data, a message type of the data packet, or a process identifier specifying a type of processing to be performed on the portion of the industrial data.

18. A non-transitory computer-readable medium having stored thereon instructions that, in response to execution, cause a system comprising a processor to perform operations, the operations comprising:

receiving, via a cloud platform, industrial data from one or more industrial devices of an industrial control system;

generating a gain correlation model based on analysis of the industrial data and storing the gain correlation model on cloud storage of the cloud platform, wherein the gain correlation model defines at least one correlation between controller gain values and process variables of the industrial control system; and

generating at least one controller gain value for the industrial control system based on an analysis of the gain correlation model.

19. The non-transitory computer-readable medium of claim 18, further comprising initiating a sending of the at least one controller gain value to a specified client device.

20. The non-transitory computer-readable medium of claim 18, wherein the generating the at least one controller gain value comprises:

simulating a process variable response based on the correlation gain model using at least one test value of the at least one controller gain;

modifying the at least one test value as a function of the process variable response based on one or more gain tuning rules to yield at least one modified test value;

repeating the simulating using the at least one modified test value to yield a new process variable response until the new process variable response satisfies a defined criterion; and

setting, as the at least one controller value, the at least one modified test value that resulted in the new process variable response that satisfied the defined criterion.