help_imc.hlp

内模控制器（IMC）工具箱。包括参数整定、PID控制器参数转换等

This is the IMC-TUNE Help file
******************************

Online help is available for selected menu functions
and graphs. If no online help is available, use this file.
For online help click the Help button and then click anywhere 
within the frame of the help window and scroll the text 
using up-down arrow or PageUp-PageDown keys.
Click the upper left corner of this window to close it.

		FILE functions
		**************

NEW

Sets all variables and data files to defaults as if the IMC-TUNE
program was just started.

******************************************************************

LOAD

Press the space bar to see the choices for files to load.
Choose a file, then click on the LOAD window.  Click inside the 
rectangular box and type in the file name. Click on the OK button.

******************************************************************

SAVE

The directory above the rectangular box shows where the file will 
be saved. To save within the listed directory, click inside the 
rectangular box and type in the file name. Click on the OK button.

******************************************************************

CHANGE DIRECTORY

Click inside the rectangular box and delete the current directory 
name. Type in the new directory name, beginning with the drive 
(e.g.  a:\  b:\  or  c:\). Click on the OK button.

******************************************************************

EXIT

Exit IMC-TUNE.

******************************************************************

		EDIT functions
		**************


PROCESS

Allows the user to enter an uncertain process in the LaPlace 
form.

	    Polynomial 1
Process =  -------------- * exp(-T*s)
	    Polynomial 2

The coefficients of the polynomials and the deadtime T may be 
constants or may vary independently within specified ranges 
(see uncertainty limits).

Example of an uncertain process:

          K(-s+1)
P1 = ----------------------- exp(-T*s)
      (Tau1*s+1)*(Tau2*s+1)

with uncertain gain K, time constants Tau1 and Tau2 and deadtime T
would be entered as

       u;-1 1
P1 = ---------- * u
      u 1;u 1

in appropriate windows in the Process Edit screen.

Disturbance lag:

Enter coefficients [k1 k2 k3] of an assumed second order 
lag a disturbance passes through.

		      1
	Lag =  ---------------
		k1*s^2+k2*s+1

* default = [0 0 1]

******************************************************************

MODEL

Allows the user to enter a model of an uncertain process 
in the LaPlace form.

	    Polynomial 1
Model   =  -------------- * exp(-T*s)
	    Polynomial 2

The coefficients of the polynomials and the deadtime T have to be constants. 
Example of a model for an uncertain process:

          10(-s+1)
M1 = -------------------- exp(-5*s)
      (3s+1)(s^2+.5s+1)

would be entered as

       10;-1 1
M1 = ------------ * 5
      3 1;1 .5 1

in appropriate windows in the Model Edit screen.

In case of an optimal model calculation enter the model parameters
that you want to vary as "uncertain" the same way as the uncertain
process parameters.

         u
M1 = ---------- * u
      3 1;5 1

For this case, a gain and deadtime from specified ranges will be
chosen as model parameters to satisfy an optimality criterion
(see COMPUTE section, FIND OPTIMAL MODEL option) 

******************************************************************

INVERTIBLE PART OF THE MODEL

Allows the user to enter a part of a model in the LaPlace form which
the user wants to be inverted and become part of the IMC controller.


	      Polynomial 1
Inv. part =  -------------- 
	      Polynomial 2

The coefficients of the polynomials have to be constants. 
Example of an invertible part of a model:

              10
Inv1 = ------------------- 
        (3s+1)*(s^2+.5s+1)

would be entered as

           10
Inv1 = ------------
        3 1;1 .5 1

in appropriate windows in the Invertible Part of a Model Edit 
screen.
						       1
Since the realizable IMC controller is formed as Q = ------ * F,
where F is a filter                                   Inv1

        1
F = -----------
    (e*s+1)^r

The order of a filter "r" has to be entered in the frame within
the INVERTIBLE PART of the MODEL EDIT screen. The filter time
constant "e" can be either entered directly in the DEFAULT 
VARIABLES EDIT screen or can becomputed by choosing th eitem IMC 
CONTROLLER TUNING in the COMPUTE menu.

In case of an optimal model calculation enter the model parameters
that you want to vary as "uncertain" the same way as the uncertain
process parameters (See EDIT MODEL). This is important especially
in case optimal gain of a model is to be found.

******************************************************************

UNCERTAINTY LIMITS

Enter two vectors - upper bound of all uncertain process parameters
                  - lower bound of all uncertain process parameters
in appropriate windows. Leave a space between numbers.

Example:

Upp. bound = 11 4 6
Low. bound =  9 2 4    

In case of the uncertainty bounds calculation enter also
a nominal plant, a vector of numbers that will be assigned to
the uncertain parameters. The uncertainty bounds will then be
calculated given a control system and assuming the nominal 
plant be mid-range parameters with the same percentage 
deviations around each parameter.

Example:

Nominal plant = 10 3 5 

******************************************************************

MODEL RANGE LIMITS

Enter three vectors 
		- nominal model
		- upper bound of all uncertain model parameters
                - lower bound of all uncertain model parameters
in appropriate windows. Leave a space between numbers.

Example:

Nom. model = 10 3 5
Upp. bound = 11 4 6
Low. bound =  9 2 4    

Upper and lower bounds will be used only if FIND OPTIMAL MODEL
in the COMPUTE section is chosen. The current nominal model 
parameters will be used in all other calculations.

******************************************************************

DEFAULT VARIABLES

Allows the user to enter initial values of all parameters 
associated with the design problem.

Epsilon: 
	the filter time constant (a scalar for 1-degree-of-freedom
	IMC controller, a 1x2 vector for a 2-degree-of-freedom IMC
	controller)
	Epsilon(1) is a filter time constant for the "q*" portion
	of the 2-degree-of-freedom IMC controller
	Epsilon(2) is a filter time constant for the "qqd" portion
	of the 2-degree-of-freedom IMC controller
        * default = 0

Maximum peak:
	the maximum peak (a scalar for 1-degree-of-freedom
	IMC controller, a 1x2 vector for a 2-degree-of-freedom 
	IMC controller) of possible frequency responses over all
	plants within the specified uncertainty set
	Mp(1) is the maximum peak for the complementary 
	sensitivity function
	Mp(2) is the maximum peak for the sensitivity function
        * default = 1.05

	!!! Warning !!! The user should check whether the Mp  
	he/she specified is achievable for the 1-degree-of-
	freedom IMC controller tuning. If not, increase the Mp
	specification or the resulting filter time constant will 
	approach infinity.

Noise amplification factor:
	a number that specifies maximum allowable amplification
	of noise by the IMC controller
	* default = 20

Worst case frequency:
	frequency at which the maximum peak was found
	* default = 0

Frequency range:
	a vector of two integers that specifies the range of 
	frequencies for which the upper/lower bound of frequency
	responses will be computed
	* default = -1 1
	(i.e. the frequency range is [10^(-1) 10^(1)])

Number of points per decade:
	for which the maximum magnitude of frequency responses
	will be computed
	* default = 10
	(i.e. for the above frequency range it would mean to 
	compute 20 points for both upper and lower bound)

Accuracy:
	A number taken as a condition to stop some algorithms when
	desired accuracy is achieved
	* default = 0.005

Infinity:
	A number to be considered "very large" or "infinity"
	* default = 10^5

Control structure filename:
	These are four built-in control structures:

	imc1com -  1-degree-of-freedom complementary sensitivity
		   function
	imc2com -  2-degree-of-freedom complementary sensitivity
		   function
	imc1sen -  1-degree-of-freedom sensitivity function
	imc2sen -  2-degree-of-freedom sensitivity function
	imc1msf -  model state feedback

	The user can define a control structure and save it in
	an accessible directory under "user1.m" or "user2.m".
	The user-defined control structures can be used only for 
	upper or lower bound of frequency responses computations 
	not for tuning the control systems.

	user1   -  user-defined control structure for a 1-degree
		   of-freedom control system (graphs shown
		   under the complementary sensitivity function
		   option)
	user2   -  user-defined control structure for a 2-degree
		   of-freedom control system (graphs shown
		   under the complementary sensitivity function
		   option)

	* default = imc1com


******************************************************************