Method and system for updating tuning parameters of a controller

Patent metadata

Field	Value
Publication number	US-9170573-B2
Application number	US-201314108899-A
Country	US
Kind code	B2
Filing date	Dec 17, 2013
Priority date	Sep 24, 2009
Publication date	Oct 27, 2015
Grant date	Oct 27, 2015

Abstract

A method and system for updating tuning parameters associated with a controller without repetitive compilation of a controller code. The controller code represents an algorithm associated with the controller and can be compiled separately from a data set representing a solution for an optimization problem and also from a data set representing parameters required for prediction. The algorithm can be implemented in a programming language code suitable for implementation on an embedded platform or other types of computer platforms. The data sets can be represented in a specified data structure and the variables associated with the data structure can be declared in the controller template code. The variables can be updated independently without varying the compiled code associated with the controller algorithm that is referring to the variables. The controller can also be updated while the controller actively performs online. Such an approach enables repetitive tuning of the controller without repetitive compilation of the code representing the controller algorithm.

First claim

What is claimed is:

1. A method for updating tuning parameters of a controller using a development station that is separate from but in operative communication with the controller, said method comprising:
compiling a set of controller instructions to generate a compiled controller code, the compiled controller code including a control algorithm to control a dynamic operating process, the complied controller code comprising one or more references to a pre-allocated memory space that stores a data set that is not compiled, the data set comprising one or more tuning parameters for tuning the control algorithm of the compiled controller code;
generating the data set using the development station;
transferring the generated data set from the development station to the pre-allocated memory space;
controlling the dynamic operating process using the control algorithm of the compiled controller code referencing the one or more tuning parameters of the data set;
updating one or more of the tuning parameters of the data set using the development station during a tuning process of the controller and transferring one or more of the updated tuning parameters from the development station to the pre-allocated memory space without altering said compiled controller code in order to tune said control algorithm of the controller without performing a recompilation of the compiled controller code; and
controlling the dynamic operating process using the control algorithm of the compiled controller code referencing the one or more updated tuning parameters.

2. The method of claim 1 , wherein the data set has a predefined data structure.

3. The method of claim 2 , wherein the data set further has one or more declared variables.

4. The method of claim 1 , wherein the data set includes one or more variables, and the compiled controller code includes a variable space, wherein the one or more variables of the data set are directly or indirectly accessible via the compiled controller code variable space.

5. The method of claim 1 , wherein the data set includes one or more variables that influence the control algorithm of the compiled controller code.

6. The method of claim 5 , wherein the one or more variables of the data set define, at least in part, one or more of the tuning parameters, a state space and/or one or more parameter space partitions for the compiled controller code.

7. The method of claim 1 , wherein said controller comprises a MPC-based controller.

8. The method of claim 7 , wherein said MPC-based controller comprises an explicit MPC-based controller.

9. The method of claim 1 , wherein said controller is an embedded controller.

10. The method of claim 1 , further comprising repeatedly updating using the development station one or more of the tuning parameters of the data set during the tuning process of the controller without halting operation of the compiled controller code.

11. The method of claim 10 , wherein the data set is generated using an offline optimization process using the development station.

12. The method of claim 10 , wherein repeatedly updating one or more of the tuning parameters of the data set during the tuning process of the controller comprises configuring a data loader and a first memory unit, wherein said data loader copies the data set from said first memory unit to the pre-allocated memory space.

13. A method for updating tuning parameters of a controller using a development station that is separate from but in operative communication with the controller, said method comprising:
loading a compiled controller code onto the controller for execution by the controller, the compiled controller code including a control algorithm to control a dynamic operating process of a system, the complied controller code declaring a number of variables that are associated with a data set, wherein the data set is not compiled;
generating the data set using a development station, the data set including one or more tuning parameters for tuning the control algorithm that can be accessed through the number of declared variables of the compiled controller code; and
updating at least one of the one or more tuning parameters of the data set during a tuning process using the development station without altering said compiled controller code in order to tune said control algorithm of the controller without performing a recompilation of the compiled controller code.

14. The method of claim 13 further comprising loading the data set onto the controller such that the compiled controller code has direct accesses to one or more of the tuning parameters through the number of declared variables of the compiled controller code.

15. The method of claim 13 , wherein the data set has a predefined data structure.

16. The method of claim 13 , wherein the data set has one or more variables that correspond to one or more of the declared variables of the compiled controller code.

17. The method of claim 13 , wherein the data set includes one or more variables, wherein the one or more variables of the data set are directly or indirectly accessible via the declared variables of the complied controller code.

18. The method of claim 13 , wherein said controller comprises a MPC-based controller.

19. A controller for controlling a dynamic operating process, the controller comprising:
a memory for storing a compiled controller code, wherein the complied controller code includes one or more references to a pre-allocated memory space that stores a data set that is not compiled, the data set transferred to the pre-allocated memory space from an external source via a communication input of the controller, the data set comprising one or more tuning parameters for tuning the controller;
a processor coupled to the memory for executing the compiled controller code, including referencing the non-compiled data set stored in the pre-allocated memory space of the memory; and
the memory repeatedly receiving one or more updates to the data set from the external source during a tuning process of the controller without altering said compiled controller code including performing a recompilation of the compiled controller code, whereupon the compiled controller code references the updated non-compiled data set stored in the pre-allocated memory space of the memory.

20. The controller of claim 19 , wherein the controller provides one or more outputs to the dynamic operating process.

Assignees

• Honeywell Int Inc

Inventors

• Kihas Dejan

Classifications

• G05B13/048 Primary

  using a predictor · CPC title

• G05B13/042 Primary

  in which a parameter or coefficient is automatically adjusted to optimise the performance · CPC title

• G05B17/02

  electric · CPC title