Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
This lecture was delivered by Dr. Sakal Japendu for Process Control course at Ambedkar University, Delhi. It includes: Empirical, Model, Identification, Design, Implement, Graphical, Statistical, Calculations, Modeling, Process
Typology: Slides
1 / 35
This page cannot be seen from the preview
Don't miss anything!
When I complete this chapter, I want to be
able to do the following.
Design and implement a good experiment
-^
Perform the graphical calculations
-^
Perform the statistical calculations
-^
Combine fundamental and empiricalmodelling for chemical process systems
docsity.com
Outline of the lesson.
Experimental design for model building
-^
Process reaction curve (graphical)
-^
Statistical parameter estimation
-^
Workshop
docsity.com
Experimental DesignPlant Experimentation
Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
Alternativedata A priori knowledge
Not justprocesscontrol
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic EvaluationModel Verification
Start Complete
Looks very general; it is!However, we still need tounderstand the process! •^
Changing the temperature 10 K in a ethane pyrolysisreactor is allowed.
Changing the temperature in a ?? Reactor would killthe micro-organisms
T A
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start
EMPIRICAL MODEL BUILDING PROCEDUREComplete
-^
Gain, time constant, dead time ...
-^
Does the model fit the data usedto evaluate the parameters?
-^
Does the model fit a new set ofdata not used in parameterestimation.
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start
EMPIRICAL MODEL BUILDING PROCEDUREComplete
What our goal?We seek models good enough forcontrol design, controller tuning,and process design.
-^
How do we know?We’ll have to trust the book andinstructor for now. But, we willcheck often in the future!
docsity.com
45 35 25 15 5 - input variable in deviation (% open)
15 11 7 3 -1 -
output variable in deviation (K)
0
10
20
30
40
time (min)
Process reaction curve - Method I
δ
= maximum slope
θ
igure
shown in f
S
K
p^ =
τ θ
δ
/
/
Data is plotted in deviation variables
docsity.com
45 35 25 15 5 - input variable in deviation (% open)
15 11 7 3 -1 -
output variable in deviation (K)
0
10
20
30
40
time (min)
Process reaction curve - Method II
δ
τ
τ θ
δ −
%
%
%
63
28
63
5 1^ t
t
t
p
t63%
t28%
Data is plotted in deviation variables
docsity.com
Recommended
Process reaction curve - Methods I and II
The same experiment in either method!Method I
Developed first
-^
Prone to errorsbecause of evaluationof maximum slope
Method II
Developed in 1960’s
-^
Simple calculations
docsity.com
Process reaction curve
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
45 35 25 15 5 - input variable in deviation (% open)
15 11 7 3 -1 -
output variable in deviation (K)
0
10
20
30
40
time (min)
Is this a well designed
experiment?
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
45 35 25 15 5 - input variable, % open
15 11 7 3 -1 -
output variable, degrees C
0
10
20
30
40
time
Process reaction curve
Should we use this data?
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
45 35 25 15 5 - input variable, % open
15 11 7 3 -1 -
output variable, degrees C
0
10
20
30
40
time
Process reaction curve
The output must be “moved”enough. Rule of thumb:
Signal/noise > 5
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
Process reaction curve
45 35 25 15 5 - input variable, % open
10 6 2 -2 -6 -
0
20
40
60
80
time
Output did notreturn close to theinitial value,although inputreturned to initialvalue
This is a good experimental design; it checksfor disturbances
docsity.com
Experimental DesignPlant Experimentation Determine Model Structure
Parameter EstimationDiagnostic Evaluation
Model Verification
Start Complete
Process reaction curve
45 35 25 15 5 - input variable, % open
15 11 7 3 -1 -
output variable, degrees C
0
10
20
30
40
time
Plot measured vs predicted measured
predicted
docsity.com
