Model-Free Adaptive (MFA) Control with Intelligent Engine and Loop Inspection

Abstract

A method and apparatus for intelligently controlling continuous process variables. An automatic controller comprises an Intelligent Engine mechanism and a number of Model-Free Adaptive (MFA) controllers, each of which is suitable to control a process with specific behaviors. The Intelligent Engine can automatically select the appropriate MFA controller and its parameters so that the automatic controller can be easily used by people with limited control experience and those who do not have the time to commission, tune, and maintain automatic controllers.

1 . A controller for controlling a single-input-single-output process having a process output which is controlled by the controller output u(t) applied to an actuator whose output is the input of said process, said controller including a neural network comprising:
 a) an error input representative of the difference between a predetermined setpoint and said process output;   b) a normalization unit for normalizing said error input to a predetermined range of values;   c) a scaling function for scaling said normalized error input to produce a value E 1  of the form   
       
         
           
             
               
                 
                   E 
                   1 
                 
                 = 
                 
                   
                     
                       K 
                       cx 
                     
                     
                       T 
                       cx 
                     
                   
                    
                   
                     N 
                      
                     
                       ( 
                       
                         e 
                          
                         
                           ( 
                           t 
                           ) 
                         
                       
                       ) 
                     
                   
                 
               
               , 
             
           
         
       
       in which K cx  is the controller gain; T cx  is the user-selected time constant of said process; N(·) is the normalization function of said normalization unit; and e(t) is the value of said error input at any given time;
 d) a layer of input neurons having as their inputs successively time-delayed values of E 1 ; 
 e) a layer of hidden neurons each having as its output the sum of individually weighted ones of said successively time-delayed values of E 1 ; 
 f) an output neuron having as its output the sum of a first function of the individually weighted outputs of said hidden neurons; 
 g) a control signal u(t) which is at least in part the denormalized value of a second function of the output of said output neuron; and 
 h) an intelligent engine arranged to automatically enter the said controller gain K cx  and time constant T cx . 
 
     
     
         2 . The controller of  claim 1 , in which said control signal u(t) is the sum of said denormalized value and the value K cx e(t). 
     
     
         3 . The controller of  claim 1 , in which said first and second functions are both of the form 
       
         
           
             
               
                 
                   f 
                    
                   
                     ( 
                     x 
                     ) 
                   
                 
                 = 
                 0 
               
               , 
               
                 
                   if 
                    
                   
                       
                   
                    
                   x 
                 
                 < 
                 
                   - 
                   
                     b 
                     a 
                   
                 
               
             
           
         
         
           
             
               
                 
                   f 
                    
                   
                     ( 
                     x 
                     ) 
                   
                 
                 = 
                 
                   ax 
                   + 
                   b 
                 
               
               , 
               
                 
                   if 
                   - 
                   
                     b 
                     a 
                   
                 
                 ≤ 
                 x 
                 ≤ 
                 
                   b 
                   a 
                 
               
             
           
         
         
           
             
               
                 
                   f 
                    
                   
                     ( 
                     x 
                     ) 
                   
                 
                 = 
                 1 
               
               , 
               
                 
                   if 
                    
                   
                       
                   
                    
                   x 
                 
                 > 
                 
                   b 
                   a 
                 
               
             
           
         
       
       wherein a is an arbitrary constant and b=1/2. 
     
     
         4 . The controller of  claim 1 , in which said error e(t) is calculated differently based on the acting type of the process in the form
     e ( t )= r ( t )− y ( t ), if direct acting
       e ( t )=−[ r ( t )− y ( t )]. if reverse acting
   
       wherein the said intelligent engine is arranged to automatically enter the information for the process acting type. 
     
     
         5 . A method of inspecting the functionality of a control system using automatically selected Model-Free Adaptive (MFA) controllers, comprising:
 a) establishing a database to save the selected MFA controller and related configuration information of that controller;   b) checking all configuration information and inputting any necessary changes;   c) if all configurations are accurate, individually verifying that the control loop signals of setpoint (SP), process variable (PV), controller output (OP), and output tracking variable (OTV) are correct, and that the controller is in manual mode; and   d) if all of said signals are verified to be correct, indicating to the user that the control system is ready to launch.   
     
     
         6 . The method of  claim 5 , in which the Model-Free Adaptive (MFA) controller is a general-purpose controller including a P (proportional) controller, a PI (proportional-integral) controller, a PD (proportional-derivative) controller, and a PID (proportional-integral-derivative) controller.