US 5,477,444 A · Control system using an adaptive neural network for target and path optimization for a multivariable, nonlinear process

Abstract

A control system having four major components: a target optimizer, a path optimizer, a neural network adaptation controller and a neural network. In the target optimizer, the controlled variables are optimized to provide the most economically desirable outputs, subject to operating constraints. Various manipulated variable and disturbance values are provided for modeling purposes. The neural network receives as inputs a plurality of settings for each manipulated and disturbance variable. For target optimization all the neural network input values are set equal to produce a steady state controlled variable value. The entire process is repeated with differing manipulated variable values until an optimal solution develops. The resulting target controlled and manipulated variable values are provided to the path optimizer to allow the manipulated variables to be adjusted to obtain the target output. Various manipulated variable values are developed to model moves from current to desired values. In this case trend indicating values of the manipulated and disturbance variables are provided to produce time varying values of the controlled variables. The process is repeated until an optimal path is obtained, at which time the manipulated variable values are applied to the actual process. On a periodic basis all of the disturbance, manipulated and controlled variables are sampled to find areas where the training of the neural network is sparse or where high dynamic conditions are indicated. These values are added to the set of values used to train the neural network.

Claims46

Dependent claims

A control system for a process having a plurality of disturbance variables as process input conditions, a plurality of manipulated variables as process control conditions and a plurality of controlled variables as process output conditions, the control system comprising:

means for periodically determining the values of the disturbance variables;
means for optimizing the values of the controlled variables based on the determined disturbance variable values and providing target manipulated variable values associated with said optimized controlled variable values;
means receiving said target manipulated variable values for optimizing the movement from the existing manipulated variable values to said target manipulated variable values and providing new manipulated variable values;
and
a neural network coupled to said controlled variable optimizing means and said movement optimizing means and receiving manipulated variable values and disturbance variable values and producing the resulting controlled variable values representative of the process being controlled.

2 more elements

A method of controlling a process having a plurality of disturbance variables as process input conditions, a plurality of manipulated variables as process control conditions and a plurality of controlled variables as process output conditions, the method comprising the steps of:

periodically determining the values of the disturbance variables;
optimizing the values of the controlled variables based on the determined disturbance variable values and providing target manipulated variable values associated with said optimized controlled variable values;
and
optimizing the movement from the existing manipulated variable values to said target manipulated variable values and providing new manipulated variable values;
and
said steps of controlled variable optimizing and movement optimizing utilizing a neural network receiving manipulated variable values and disturbance variable values and producing the resulting controlled variable values representative of the process being controlled.

3 more elements

A system for performing a distillation process in an oil refinery having a plurality of disturbance variables as process input conditions, a plurality of manipulated variables as process control conditions and a plurality of controlled variables as process output conditions, the system comprising:

a distillation column having a means for receiving an input feed stream and providing overhead and bottom output streams;
means for cooling said overhead output stream;
means for controllably returning a portion of said cooled overhead output stream to said distillation column, said rate of controllably returning being a manipulated variable;
means for controllably maintaining the pressure of materials in said distillation column, said level of controllably maintaining the pressure being a manipulated variable;
means for controllably heating a portion of material in said distillation column, said rate of controllably heating being a manipulated variable;
means for periodically determining the values of the disturbance variables;
means for optimizing the values of the controlled variables based on the determined disturbance variable values and providing target manipulated variable values associated with said optimized controlled variable values;
means receiving said target manipulated variable values for optimizing the movement from the existing manipulated variable values to said target manipulated variable values and providing new manipulated variable values;
and
a neural network coupled to said controlled variable optimizing means and said movement optimizing means and receiving manipulated variable values and disturbance variable values and producing the resulting controlled variable values representative of the distillation process being controlled.

7 more elements

Dependencies are inferred from the claim text.

Control system using an adaptive neural network for target and path optimization for a multivariable, nonlinear process

Life of this patent

Control system using an adaptive neural network for target and path optimization for a multivariable, nonlinear process

Life of this patent

Abstract

Claims46