Lect. 26 CHE 185 – PID ENHANCEMENTS

Report
CHE 185 – PROCESS
CONTROL AND DYNAMICS
PID ENHANCEMENTS
LIMITATIONS OF CONVENTIAL PID
CONTROLLERS
• THE PERFORMANCE OF PID CONTROLLERS
CAN BE SUBSTANTIALLY LIMITED BY:
– PROCESS NONLINEARITY
– MEASUREMENT DEADTIME
– PROCESS CONSTRAINTS
• THERE ARE SEVERAL APPROACHES FOR
PID CONTROLLERS TO HANDLE EACH OF
THESE PROBLEMS
INFERENTIAL CONTROL
• USES EASILY MEASURED PROCESS
VARIABLES (T, P, F) TO INFER MORE
DIFFICULT TO MEASURE QUANTITIES SUCH
AS COMPOSITIONS AND MOLECULAR
WEIGHT.
• CAN SUBSTANTIALLY REDUCE ANALYZER
DELAY.
• CAN BE MUCH LESS EXPENSIVE IN TERMS
OF CAPITAL AND OPERATING COSTS.
• CAN PROVIDE MEASUREMENTS THAT ARE
NOT AVAILABLE ANY OTHER WAY
INFERENTIAL CONTROL
• EXAMPLES OF VARIABLES THAT ARE NOT
EASY TO MEASURE DIRECTLY
–
–
–
–
–
DENSITY
VAPOR PRESSURE
MELT INDEX
GAS COMPOSITION
MOLECULAR WEIGHT
INFERENTIAL CONTROL
• SECONDARY MEASUREMENTS ARE USED
WITH THE FOLLOWING FOR INFERENTIAL
CONTROL
– PROCESS MODEL EQUATIONS
– THERMODYNAMIC RELATIONSHIPS, I.E. LINKING
TEMPERATURE TO CONCENTRATION
– EMPIRICAL MODELING
– ISOTHERMAL VISCOSITY VERSUS LIQUID
COMPOSITION
INFERENTIAL CONTROL
• MEASURES A VARIABLE USING AN
INDIRECT METHOD
• USED WHEN
– IT IS NOT PRACTICAL TO MEASURE THE TARGET
VARIABLE
– EXCESSIVE COST FOR CONTROL EQUIPMENT
TO DIRECTLY MEASURE THE VARIABLE
– EXCESSIVE DOWNTIME IN A TARGET VARIABLE
SENSOR
– THERE IS AN INFERENTIAL VARIABLE AVAILABLE
INFERENTIAL CONTROL
• CHARACTERISTICS OF THE INFERENTIAL
VARIABLE
• IT MUST BE CLOSELY RELATED TO THE
TARGET VARIABLE
• IT MUST NOT BE AFFECTED BY CHANGES IN
THE PROCESS CONDITIONS
• DYNAMICS ARE ADEQUATE FOR FEEDBACK
CONTROL
INFERENTIAL CONTROL
• CORRECTIONS TO INFERENTIAL CONTROL
VARIABLE
• CAN USE A CASCADE CONTROL SOURCE
• CAN BE MANUALLY ADJUSTED
INFERENTIAL CONTROL
• EXAMPLE USING TEMPERATURE TO
CONTROL COMPOSITION FOR ISOBARIC
FLASH
Vapor
PC
SetPoint
Heating or
Cooling
Media
Process
Feed
Heat
Exchange
Product
TC
Flash Tank
LC
Liquid
Product
INFERENTIAL CONTROL
• EXAMPLE USING TEMPERATURE TO
CONTROL COMPOSITION FOR ISOBARIC
FLASH
• CONTROLS COMPOSITION BASED ON
FLASH TEMPERATURE
• DIRECT CONTROLLED VARIABLE IS FLASH
PRESSURE
• LEVEL IS ALSO DIRECTLY CONTROLLED
INFERENTIAL CONTROL
• EXAMPLE USING TEMPERATURE TO
CONTROL COMPOSITION FOR ISOBARIC
FLASH
• HOW IS THE TEMPERATURE SETTING
CHECKED FOR THIS EXAMPLE?
• MANUAL ANALYSIS CAN BE USED TO
ADJUST
• A FEED FORWARD SIGNAL FROM A
PROCESS ANALYZER CAN ALSO BE USED
(SEE SKETCH NEXT SLIDE)
INFERENTIAL CONTROL
• IT IS ASSUMED
THAT THE LAG TIME
FOR THE ANALYZER
LOOP IS LONGER
THAN THAT FOR
THE TEMPERATURE
LOOP.
• THIS ALSO WILL
TAKE CARE OF ANY
STEADY-STATE
OFFSET FOR THE
TEMPERATURE
CONTROL
SetPoint
Signal
Heating or
Cooling
Media
Process
Feed
Heat
Exchange
PC
Vapor
Product
TC
Flash Tank
LC
Liquid
Product
AC
INFERENTIAL TEMPERATURE CONTROL
FOR DISTILLATION COLUMNS
• REBOILER CONTROL BASED ON TRAY
TEMPERATURE
INFERENTIAL TEMPERATURE CONTROL
FOR DISTILLATION COLUMNS
• CHOOSING A PROPER TRAY TEMPERATURE
LOCATION
• TRAY TEMPERATURE USED FOR INFERENTIAL
CONTROL SHOULD SHOW STRONG SENSITIVITY
INFERENTIAL TEMPERATURE CONTROL
FOR FLOW REACTOR
• SEE EXAMPLE 13.2 IN TEXT
ARTIFICIAL NEURAL NETWORKS
(ANN’s)
• THESE ARE NON-LINEAR
CONTROLLERS THAT ARE USED TO
CONTROL NON-LINEAR PROCESSES
• THE MODEL TAKES INPUT(S) FROM
THE SYSTEM AND USES THESE WITH
WEIGHTED FUNCTIONS, TO PROVIDE
THE OUTPUT FOR THE CONTROLLER
ARTIFICIAL NEURAL NETWORKS
(ANN’s)
• THE WEIGHTING FUNCTIONS ARE
REVISED OVER TIME TO OPTIMIZE
THE OUTPUT
• THE ANN IS TUNED BY THE SYSTEM
AND ONLY APPLIES TO ONE SYSTEM.
ARTIFICIAL NEURAL NETWORKS
(ANN’s)
• SOFT SENSORS BASED ON
NEURAL NETWORKS
• NEURAL NETWORK (NN)
PROVIDES NONLINEAR
CORRELATION.
• WEIGHTS ARE ADJUSTED
UNTIL NN AGREES WITH
PLANT DATA
• NN-BASED SOFT SENSORS
ARE USED TO INFER NOX
LEVELS IN THE FLUE GAS
FROM POWER PLANTS.
SCHEDULING CONTROLLER
TUNING
• THIS IS A METHOD TO COMPENSATE
FOR PROCESS NON-LINEARITY THAT
CAN AFFECT CONTROL RESPONSE
• THE BASIC TECHNIQUE IS TO TUNE
THE CONTROLLER BASED ON
EMPIRICAL DATA
– OPTIMUM TUNING DATA IS OBTAINED
OVER A RANGE OF PROCESS SETTINGS.
SCHEDULING CONTROLLER
TUNING
• THE TUNING DATA IS THEN
CONVERTED INTO PROPORTIONAL,
INTEGRAL AND DERIVATIVE RESET
FUNCTIONS OF THE MANIPULATED
VARIABLE.
• THIS METHOD IS SIMILAR TO ANN
EXCEPT IT ONLY LOOKS AT ONE
INPUT VARIABLE AND RESULTS IN
CLEARLY DEFINED FUNCTIONS
SCHEDULING CONTROLLER
TUNING
• ADJUST
TUNING OF
HEAT
EXCHANGER
CONTROL
FOR VARIOUS
FEED RATES
• LINK TUNING
PARAMETERS
TO THE FLOW
RATES
SCHEDULING CONTROLLER
TUNING
• TYPICAL OPEN LOOP RESPONSE
SCHEDULING CONTROLLER
TUNING
• CLOSE LOOP RESPONSE WITH SCHEDULING
SCHEDULING CONTROLLER
TUNING
• CLOSE LOOP RESPONSE WITH SCHEDULING
SCHEDULING CONTROLLER
TUNING
• IMPLEMENTATION CAN TAKE THE FORM OF
ADJUSTMENT OF PI GAIN AND INTEGRAL
TIME USING THE TUNING FACTORS
• FOR EXAMPLE USING ZEIGLER-NICHOLS
(EQUATION 9.11.2):
OVERRIDE/SELECT CONTROL
• THIS METHOD EMPLOYS A SELECTION
AMONG MULTIPLE INPUTS
– IT CAN BE APPLIED TO ROUTINE CONTROL
– IT CAN BE USED TO IMPLEMENT EMERGENCY
CONTROL
• UNDER NORMAL OPERATION A LOW SELECT
OR A HIGH SELECT METHOD IS USED BY
THE CONTROLLER TO ADJUST THE
MANIPULATED VARIABLE
OVERRIDE/SELECT CONTROL
• INPUT COMES FROM TWO OR MORE
CONTROLLERS TO A SECOND IN A CASCADE
CONFIGURATION
• THE COMPARISON CONTROLLER CHOOSES
THE LOWEST OR HIGHEST TO SEND TO THE
ACTUATOR
• CONSIDER A REACTOR WITH COOLING FOR
TEMPERATURE CONTROL
OVERRIDE/SELECT CONTROL
• THE LOW SELECTOR TAKES THE LOWER
VALUE FROM THE COMPOSITION ANALYZER
AND THE REACTOR TEMPERATURE SENSOR
• THE LOWER VALUE IS SELECTED BECAUSE
THIS ASSURES THE HIGHEST COOLING
FLOW TO THE UNIT.
LS
AC
ISOTHERMAL
CSTR
TC
OVERRIDE/SELECT CONTROL
• TEXT PROVIDES SEVERAL OTHER
EXAMPLES BASED ON HIGH, LOW AND
COMBINED SELECTION
• NOTE THAT IT IS IMPORTANT FOR THE
OPERATOR TO KNOW WHICH SIGNAL IS
BEING USED BY THE CONTROLLER.
• MAY BE USED FOR LOW AND HIGH LEVEL
ALARM ACTIONS
– ALERTS OPERATOR TO OUT-OF-RANGE AND
INITIATES CORRECTION WITHIN THE LOOP
– NOT INTENDED TO REPLACE SEPARATE HI-HI
AND LO-LO ALARMS
COMPUTED MANIPULATED
VARIABLE CONTROL
• THESE ARE APPLIED MASS BALANCES,
ENERGY BALANCES OR REACTION MODELS
THAT ARE USED TO SPECIFY OPERATING
SET POINTS.
• CAN BE USED FOR COMPLICATED SYSTEMS
THAT CAN BE CONVENIENTLY MODELED
• TYPICALLY USED AS A SECONDARY SET
POINT GENERATOR
• MAY BE LINKED TO SIMULATORS
COMPUTED MANIPULATED
VARIABLE CONTROL
• COMPUTED REBOILER DUTY CONTROL
COMPUTED MANIPULATED
VARIABLE CONTROL
• INTERNAL REFLUX CONTROL

similar documents