Two-phase method for real time process control

• US 5,329,443 A
• Filed: 06/16/1992
• Issued: 07/12/1994
• Est. Priority Date: 06/16/1992
• Status: Expired due to Fees

• Current Assignee: Praxair Technology Incorporated (Linde Plc)
• Original Assignee: Praxair Technology Incorporated (Linde Plc)
• Inventors: Karwan, Mark H., Jordan, Michael D., Bonaquist, Dante P.
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): TOUSI, CAMERON
• Application Number:
  US07/899,477
• Time in Patent Office: 756 Days
• Field of Search: 364/153, 364/148, 364/149, 364/150, 364/151, 364/152, 364/154-156, 364/158, 364/172
• US Class Current: 700/33
• CPC Class Codes: F25J 2290/10 Mathematical formulae, modeling, plot or curves; Design methods
• CPC Class Codes: F25J 3/044 using a single pressure main column system only F25J3/0446, F25J3/04624, F25J3/04636 take precedence
• CPC Class Codes: F25J 3/04848 Control strategy, e.g. advanced process control or dynamic modeling
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 13/048 using a predictor

Abstract

A method of process control using linear programming models to implement the control of one or more controlled variables to maintain a desired target value over a given time horizon. The controlled variables are controlled by a process controller which adjusts manipulated variables in real time. The method uses two linear programming models with the second model computed from the solution of the first in which the absolute value of the deviation for each controlled variable from its target value over a specified time horizon is computed. This permits the controlled variables to be adjusted with the changes in the manipulated variables held to a minimum over the time horizon under control.

Claims

1. A method for controlling a process for separating air into high purity gas streams having one or more manipulated variables in which one manipulated variable is an air flow stream and one or more controlled variables with one of the controlled variables being the degree of purity of a discharge gas stream and having target values which depend upon the adjusted value of said manipulated variables, with the process controlled in real time using a process controller under the operation of a computer comprising the steps of:

establishing a first performance index to compute the absolute value of the deviation for each controlled variable in the process from its target value over a specified time horizon;

generating a first linear programming model the solution of which will minimize said first performance index;

solving said first linear programming model;

establishing a second performance index to compute the absolute change in the value of each manipulated variable from its previous value over a specified time interval;

generating a second linear programming model the solution of which will minimize said second performance index;

incorporating at least one dynamic constraint in said second linear programming model computed from the solution of said first linear programming model and being equal to a value above zero and of no greater than the value of the solution of said first linear programming model plus a predetermined amount;

solving said second linear programming model with said dynamic constraint; and

adjusting the manipulated variables in response to the solution of said second linear programming model to drive said controlled variables toward their target values and thereby causing the air to be separated into a high purity gas stream.

2. A method defined in claim 1 further comprising:

using said second performance index for defining an objective function for said first performance index as follows;

space="preserve" listing-type="tabular">______________________________________ Minimize ##STR6## Minimize ##STR7## wherein;

α

.sup.p .sub.t+n = positive deviation of controlled variable p from target at time t + n;

β

.sup.p .sub.t+n = negative deviation of controlled variable p from target at time t + n;

s.sup.p .sub.t+n = weight on deviation of controlled variable p from target at time t + n;

u.sup.r .sub.t+ n = positive change in manipulated variable r at time t + n;

v.sup.r .sub.t+n = negative change in manipulated variable r at time t + n; and

w.sup.r .sub.t+n = weight on changes in manipulated variable r at time t + n ε

>

0. ______________________________________

3. A method as defined in claim 2 wherein said predetermined value is a fraction of the value of the solution of the first phase linear programming value.

4. A method as defined in claim 2 wherein said predetermined value is computed as follows:

space="preserve" listing-type="equation">Δ

IAE.sup.P =(N)Δ

.sup.P (

     119)

Where;

N number of time intervals considered in the controller time horizon

Δ

^P increase in the contribution to phase one performance index from deviations from the target value of controlled variable p per time interval allowed for phase two solution.

5. A method as defined in 3 wherein said specified time horizon covers a predetermined number of time intervals.

6. A method as defined in claim 5 further comprising implementing only the manipulated variable charge corresponding to the first element in the time sequence of manipulated variable changes derived from the solution of said second linear programming model.

7. A method as defined in claim 6 further comprising measuring the current state of the process . defined by the most recent values for all variables including controlled and manipulated variables.

8. A method as defined in claim 7 wherein external unmeasured disturbances in each time interval will cause the state of the process to be modified.

9. A method as defined in claim 8 wherein the steps of solving said first linear programming model and said second linear programming model are repeated each time interval based on the updated state of the process.

10. A method as defined in claim 9 wherein the process being controlled is a process for separating air into high purity gas streams with one of the controlled variables being the degree of purity of a discharged gas stream and wherein one of the manipulated variables is the air flow feed stream.