Method and apparatus for dynamic optimization
AssigneeVoyan Technology
InventorsMark A. Erickson, Pradeep Pandey, Sunil C. Shah, Thorkell Gudmundsson
Life of this patent
Estimating the term…
Abstract
The present invention is a method and apparatus for dynamic optimization of a dynamic physical system having a device under control and a controller along with a modeling system having an estimator and a constrained nonlinear program. The time varying trajectories of the dynamic physical system are parameterized and run through the modeling system in order to produce a set of optimal time varying setpoint trajectories to be used in the dynamic physical system. The optimal time varying setpoint trajectories optimize the physical system to produce the desired end result (or product) of the dynamic physical system.
Claims61
A dynamic optimizer comprising:
- a dynamic physical system, wherein said dynamic physical system runs at a first time scale;
- an estimator, wherein said estimator captures the dynamic and time varying behavior of said physical system and wherein said estimator runs at a second time scale;
- and
- a constrained program, wherein said constrained program runs at said second time scale and wherein said constrained program uses a cost function to evaluate the quality of a final output of said dynamic physical system and if necessary produces a time varying trajectory, said time varying trajectory to be used by said dynamic physical system in said first time scale.
A dynamic optimizer comprising:
- a dynamic physical system, wherein said dynamic physical system runs at a first time scale, said dynamic physical system having a device under control and a controller;
- an estimator wherein said estimator captures the dynamic and time varying behavior of said physical system and wherein said estimator runs at a second time scale;
- and
- a constrained quadratic program, wherein said constrained program runs at said second time scale and wherein said constrained quadratic program uses a cost function to evaluate the quality of a final output of said dynamic physical system and if necessary produces a reference parameter, said reference parameter to be used by said dynamic physical system in said first time scale.
A method for dynamic optimization comprising:
- creating a model of a dynamic physical system, wherein said physical system runs at a first time scale and said model of said dynamic physical system runs at a second time scale;
- creating an estimator for said model running at said second time scale, wherein said estimator captures the dynamic and time varying behavior of said physical system;
- and
- creating a constrained program running at said second time scale, said constrained program using a cost function to evaluate the quality of a final output of said dynamic physical system and if necessary producing a time varying trajectory to control said dynamic physical system running at said first time scale.
A computer readable medium containing executable instructions which, when executed in a processing system, cause said system to perform a method comprising:
- creating a model of a dynamic physical system, wherein said physical system runs at a first time scale and said model of said dynamic physical system runs at a second time scale;
- creating an estimator for said model running at said second time scale, wherein said estimator captures the dynamic and time varying behavior of said physical system;
- and
- creating a constrained program running at said second time scale, said constrained program using a cost function to evaluate the quality of a final output of said dynamic physical system and if necessary producing a time varying trajectory to control said dynamic physical system running at said first time scale.
A computer readable medium containing executable instructions which, when executed in a processing system, cause said system to perform a method comprising:
- a) parameterizing a first set of time varying trajectories, said first set of time varying trajectories including at least:
- a reference signal and an actuator signal;
- b) creating a vector valued function that captures the dynamic and time varying behavior of said physical system;
- c) creating an estimator;
- d) creating a cost function to evaluate the quality of an output of said physical system;
- e) setting said reference signal to a nominal reference trajectory;
- f) running said physical system with a most recent reference signal to produce a most recent output value;
- g) running said estimator with said most recent reference signal and said most recent output value from said physical system to produce an estimated output;
- h) evaluating said estimated output, wherein said estimated output is compared to a desired output such that:
- (1) if said estimated output is approximately equal to said desired output, then go to step 1);
- (2) if said estimated output is not approximately equal to the desired output, then go to step i);
- i) determining an upper bound and a lower bounds for said actuator signal such that said actuator signal does not saturate;
- j) running a constrained program to produce a new reference signal;
- k) evaluating said new reference signal wherein said new reference signal is compared to a previous reference signal such that:
- (1) if said new reference signal is approximately equal to said previous reference signal, then go to step 1);
- (2) if said new reference signal is not approximately equal to the previous reference signal, then go to step f);
- and
- l) stop.
A computer readable medium containing executable instructions which, when executed in a processing system, cause said system to perform a method comprising:
- a) parameterizing a first set of time varying trajectories, said first set of time varying trajectories including at least:
- a reference signal and an actuator signal;
- b) creating a vector valued function that captures the dynamic and time varying behavior of said physical system;
- c) creating an estimator;
- d) creating a cost function to evaluate the quality of an output of said physical system;
- e) setting said reference signal to a nominal reference trajectory;
- f) running said physical system with a most recent reference signal to produce a most recent output value;
- g) running said estimator with said most recent reference signal and said most recent output value from said physical system to produce an estimated output;
- h) evaluating said estimated output, wherein said estimated output is compared to a desired output such that:
- (1) if said estimated output is approximately equal to said desired output, then go to step l);
- (2) if said estimated output is not approximately equal to the desired output, then go to step i);
- i) determining an upper bound and a lower bounds for said actuator signal such that said actuator signal does not saturate;
- j) running a constrained program to produce a new reference signal and an estimated output based upon said new reference signal;
- k) evaluating said estimated output based upon said new reference signal wherein said estimated output based upon said new reference signal is compared to a previous estimated output such that:
- (1) if said estimated output based upon said new reference signal is approximately equal to said previous estimated output, then go to step l);
- (2) if said estimated output based upon said new reference signal is not approximately equal to the previous estimated output, then go to step f);
- and
- l) stop.
Dependencies are inferred from the claim text.