Optimization of a plurality of multiple-fuel fired boilers using iterated linear programming

• US 5,159,562 A
• Filed: 03/26/1992
• Issued: 10/27/1992
• Est. Priority Date: 07/31/1990
• Status: Expired due to Fees

• Current Assignee: Westinghouse Process Control, Inc. (Emerson Electric Company)
• Original Assignee: Westinghouse Electric Corporation (Paramount Global (f/k/a ViacomCBS Inc.))
• Inventors: Allen, Thomas, Huff, Frederick G., Novacek, John J., Putman, Richard E. J.
• Primary Examiner(s): Teska, Kevin J.
• Application Number:
  US07/859,250
• Time in Patent Office: 215 Days
• Field of Search: 364/492, 364/493, 364/494, 364/495, 364/161, 364/165, 60/39.3, 60/39.55, 60/463, 290/7, 236/14
• US Class Current: 700/288
• CPC Class Codes: G05B 13/024 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: Y02P 90/845 Inventory and reporting systems for greenhouse gases [GHG]

Abstract

A linear programming matrix is used to represent a complex process with many interrelated process variables. Assumptions are initially made regarding the values of the matrix elements, based on current or expected operation of the process. After the linear programming matrix is solved, the values of the coefficients are modified according to a test strategy and the linear programming matrix is again solved in an effort to obtain a better result. This process is repeated until solutions of matrix converge and the amount of modification of the matrix is minimal. One test strategy that can be used is SIMPLEX self-directing evolutionary operation. Another test strategy uses progressively more accurate linear approximations of nonlinear relationships in the linear programming matrix.

Claims

1. A method for optimizing control of a process having interdependent operating conditions determined by a control unit, the operating conditions including controlled supply of a substance, said method comprising the steps of:

(a) defining relationships in a linear programming matrix between the interdependent operating conditions, all constraints on the process sand a process variable for optimization;

(b) assigning initial values to matrix elements in the linear programming matrix;

(c) executing a computer program to solve the linear programming matrix;

(d) modifying selected matrix elements representing a set of the interdependent operating conditions according to a test strategy and adjusting any unselected matrix elements that require change due to said modifying;

(e) executing the computer program to produce a solution of the linear programming matrix after said modifying in step (d);

(f) repeating steps (d) and (e) for each test defined by the test strategy until convergence of the solution of the linear programming matrix resulting in optimized operating conditions; and

(g) adjusting the controlled supply of the substance using the control unit to establish the optimized operating conditions indicated by the solution of the linear programming matrix resulting at the convergence in step (f).

2. A method as recited in claim 1,

wherein said defining in step (a) includes defining energy balance equations for a cogeneration system producing steam and electrical power, and

wherein said modifying in step (d) comprises the steps of:

(d1) calculating, from the solution produced in step (c), first values of matrix elements representing a first steady state model of the cogeneration system;

(d2) calculating, from a modification, according to the test strategy, of the solution produced in step (c), a second values of the matrix elements representing a second steady state model of the cogeneration system; and

(d3) assigning the first values to the matrix elements representing steady state operation of the cogeneration system and assigning rate of change values, determined in dependence upon differences between the first and second values calculated in steps (d1) and (d2), to matrix elements representing effects of changes from conditions in steady state operation of the cogeneration system.

3. A method as recited in claim 1, wherein the interdependent operating conditions include at least five variable operating conditions,

wherein said defining in step (a) includes defining relationships among the variable operating conditions, and

wherein said modifying in step (d) uses a simplex self-directing evolutionary operation test strategy.

4. A method as recited in claim 3, wherein the interdependent operating conditions include at least 10 variable operating conditions, and

wherein said defining in step (a) includes defining the relationship among the variable operating conditions.

5. A method as recited in claim 3, wherein the process optimized by said method produces steam from a plurality of multiple-fuel fired boilers, and

wherein said defining in step (a) defines the relationships between fuel flow and steam produced for each fuel in each boiler.

6. A method as recited in claim 5 wherein said plurality of multiple-fuel fired boilers include more than two boilers each capable of receiving more than two fuels, in a single steam producing system, and

wherein said defining in step (a) defines coefficients of an equation relating fuel supplied to steam produced for each of the fuels in each of the boilers.

7. A method as recited in claim 1, wherein the process optimized by said method produces steam from a plurality of boilers, each boiler capable of receiving a plurality of fuels,

wherein said defining in step (a) includes defining a sum of each of the plurality of fuels times a heating value coefficient as equal to the steam produced by each of the boilers times an efficiency coefficient, and

wherein said modifying in step (d) uses a simplex self-directing evolutionary operation test strategy.

8. A method as recited in claim 7,

wherein said defining in step (a) allows for variance in the heating value coefficient of each of the fuels in each of the boilers in dependence upon load on each of the boilers, and

wherein said modifying in step (d) includes adjusting each heating value coefficient in dependence upon changes in the load on each of the boilers.

9. A method as recited in claim 8 wherein the heating value coefficients have a non-linear relationship with load on each of the boilers,

wherein said modifying in step (d) comprises the steps of:

(d1) determining a base case load for each boiler;

(d2) generating test cases with different loads for the boilers according to the simplex self-directing evolutionary operation test strategy and the base case load for each boiler determined in step (d1); and

(d3) adjusting each heating value coefficient by determining a value for each fuel/boiler combination in each of the test cases, corresponding to each of the loads, and

wherein said repeating in step (f) is performed for each of the test cases.

10. An apparatus for controlling a process having interdependent operating conditions, the independent operating conditions including controlled supply of a substance, said apparatus comprising:

input means for defining relationships between the operating conditions, all constraints on the process and a process variable to be optmized, in a linear programming matrix;

computer means for assigning initial values to matrix elements in the linear programming matrix, for executing a computer program to solve the linear programming matrix and for modifying selected matrix elements representing a set of the operating conditions according to a test strategy and adjusting any matrix elements that require change due to the modifying until convergence of the solution of the linear programming matrix resulting in optimized operating conditions; and

control means for controlling the supply of the substance to establish the optmized operating conditions indicated by the solution of the linear programming matrix resulting at the convergence of the solution of the linear programming matrix.

11. A method for optimizing production of steam from at least three boilers, each boiler capable of receiving at least three fuels in a fuel mixture determined by a control unit, said method comprising the steps of:

(a) defining relationships in a linear programming matrix between all constraints on the system, a process variable for optimization, each fuel flow times a corresponding heating value coefficient and the steam produced by each of the boilers times a corresponding efficiency coefficient;

(b) assigning initial values to matrix elements in the linear programming matrix;

(c) executing a computer program to solve the linear programming matrix;

(d) modifying load distribution on the boilers according to a simplex self-directing evolutionary operation test strategy and adjusting each heating value and efficiency coefficient as necessary due to nonlinearity between fuel flow and steam produced;

(e) executing the computer program to produce a solution of the linear programming matrix after said modifying in step (d);

(f) repeating steps (d) and (e) for each test defined by the simplex self-directing evolutionary operation test strategy until convergence of the solution of the linear programming matrix; and

(g) adjusting the fuel mixture using the control unit of at least one of the boilers to establish the operating conditions indicated by the solution of the linear programming matrix resulting at the convergence in step (f).

12. A method as recited in claim 11,

wherein said defining in step (a) comprises defining the heating value coefficient for each fuel supplied to a corresponding boiler as a function of load on the corresponding boiler;

wherein said assigning in step (b) comprises assigning to the heating value coefficient for each fuel/boiler combination a maximum value of the function defined in step (a) for the heating value coefficient;

wherein said modifying in step (d) comprises the steps of;

(d1) determining a base case load for each boiler;

(d2) generating test cases with test loads for the boilers determined according to the simplex self-directing evolutionary operation test strategy and the base case load determined in step (d1) for each boiler;

(d3) adjusting each heating value coefficient by determining a value of the function, corresponding to each fuel/boiler combination, for a corresponding one of the test loads, and

wherein said repeating in step (f) is performed for each of the test cases.