PredictBack White Paper

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 38

RMPCT with Predict Back and CV Measurement Compensation White Paper

Contents:
1. Introduction
2. Concepts
3. Implementation Guidelines
4. Model Identification
5. Model Building
6. Examples

INTRODUCTION
This purpose of this document is to provide a detailed description on how to use two of
the more advanced features of RMPCT, Predict-back and CV Measurement
Compensation. The discussion is mainly focused on practical issues relating to model
identification, model building and implementation. General recommendations are
shown in Italics wherever applicable.

CONCEPTS
Predict-back: is an optional RMPCT feature that provides additional feed forward
information and disturbance rejection to Controlled Variables (CV) when the process
value (PV) of a Manipulated Variable (MV) or a Disturbance Variable (DV) deviates
from its setpoint (SP). Predict-back is particular useful when the disturbance rejection
of the MV or the DV is slower than the effects of the unrejected (leak-through)
disturbances on the CVs.
When an MV is the SP of DCS-based PID controllers, it is usually good practice to keep
the PID cascade closed if the control loop performs well, and use their SPs as MVs
within a RMPCT controller design. Keeping higher level cascades closed can help

August 2000, Rev. 2 1


RMPCT with Predict Back and CV Measurement Compensation White Paper

linearize the process, improve disturbance rejection and thus improve RMPCT
controller performance across different operating conditions.
Examples include heater outlet temperature, tower overhead temperature, accumulator
pressure, etc.
The dynamic responses of higher level PID cascades are generally slower than the
basic regulatory controllers (i.e., flow controllers), thus disturbances leak through these
Manipulated Variables and can take longer to reject. In a traditional RMPCT controller
design where only the SP of the MV is considered, process disturbances that result in a
deviation of the MV.PV from its SP are not detected until the CVs are affected. When
the CVs do respond and violate their limits, RMPCT will move the MVs accordingly,
even though the MV.PV may have already returned to its SP.

Example:
Consider the heater outlet temperature illustrated in Figure 1 as a MV and the reactor
outlet temperature as a CV.

Figure 1: Hydrocracker Heater/Reactor

DV

FIC

Feed
TIC
MV

FIC

TI

CV
Fuel Gas

Assume the heater is often subject to load disturbances, such as a sudden reduction in
feed flow rate, an increase in fuel gas BTU content or a change in heater inlet
temperature caused by changes in the preheat exchanger train. When these
disturbances occur, the heater outlet temperature PV starts to deviate from its SP then
returns to the SP due to the PID feedback control action as illustrated in Figure 2.
Some time later, the reactor outlet temperature will respond to these changes and
follow a similar path due to the leak through effect of the disturbances. With the
RMPCT controller off, the CV disturbance is effectively rejected in a single open loop
settling time because the MV PV eventually returns to its SP.

August 2000, Rev. 2 2


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure 2: Hydrocracker Heater/Reactor Open-Loop Response

CV response to MV deviation
SP open loop

CV.PV

0 1 2 3 4 5 6 7 8 9 10 11 12

MV
disturbance MV return
to SP

MV.PV

MV.SP 0 1 2 3 4 5 6 7 8 9 10 11 12

Figure 3 illustrates the RMPCT controller response to the heater disturbance using a
traditional RMPCT design without Predict-back. In closed loop the heater disturbance
isn’t noticed until the reactor outlet temperature starts to respond, because without
Predict-back the controller does not utilize the MV PV to detect disturbances. By the
time the CV responds to the disturbance, the heater outlet temperature may already
have recovered from the disturbance, so any additional control moves to restore the CV
to within its funnel will be detrimental to the control of the reactor outlet temperature.
Notice how the MV cycles due the lack of information in the controller that accounts for
the MV returning to its SP.

August 2000, Rev. 2 3


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure 3: Hydrocracker Heater/Reactor Closed-Loop Response – No Predict-back

CV response to MV deviation
SP with RMPCT control
No Predict back

CV.PV

0 1 2 3 4 5 6 7 8 9 10 11 12

MV.PV
return to SP

0 1 2 3 4 5 6 7 8 9 10 11 12

Figure 4 illustrates the RMPCT controller response to the disturbance when predict-
back is used. Control of the CV can be improved if we utilize the intermediate feedback
information of the MV.PV. When Predict-back is configured for the heater outlet
temperature MV, a prediction of the MV.PV is maintained based on the MV.SP and
possibly other inputs. When the MV.PV deviates from its predicted value due to a
disturbance, the future predictions of the MV are corrected to account for the effect of
the disturbance and then “predicted back” to the MV.SP, accounting for the PID
1
controllers PID nominal response . In addition CV predictions are augmented with the
MVs deviation and future response using a model of the CV from the MV.PV. The
RMPCT controller response is a set of small, offsetting control moves of the MV.SP to
minimize the predicted deviation of the reactor outlet temperature. The end result is a
better performing controller with less MV movement and less violation of the CV limits.
Note; Predict-back does not include state estimation of disturbances that affect the MV
PV, for causality reasons.

1
When MVs are wound up, the MV.PV is not predicted back to SP and the entire MV prediction error is applied to
the CV predictions.

August 2000, Rev. 2 4


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure 4: Hydrocracker Heater/Reactor Closed-Loop Response – With Predict-back

New CV
CV response to MV deviation from
predictions
SP with RMPCT control action,
with Predict Back

Old CV
predictions

0 1 2 3 4 5 6 7 8 9 10 11 12

MV.PV
disturbance Prediction of MV.PV
MV.PV deviation
return to SP
to be mapped to
CV predictions

Old prediction
of MV.PV

0 1 2 3 4 5 6 7 8 9 10 11 12
Future MV.SP
control moves

August 2000, Rev. 2 5


RMPCT with Predict Back and CV Measurement Compensation White Paper

CV Measurement Compensation: is an optional RMPCT feature that provides additional


feedforward information and stability to Controlled Variables (CV) when the process
value (PV) of a Manipulated Variable (MV) or a Disturbance Variable (DV) deviates
from its setpoint (SP). CV Measurement Compensation subtracts a known effect from
a CV measurement, which is either not controllable or not desirable to control.
A common case is a MV exhibiting oscillatory behavior and the response of CV’s to this
behavior. CV Measurement Compensation will predict the unwanted and often high
frequency portions of the response, and in contrast to Predict-back, exclude this part of
the predictions used for control, thereby eliminating any unnecessary MV movement
from the controller.
The use of CV Measurement Compensation is particular useful in the following cases:
• Type I: A Manipulated Variable or Disturbance Variable PV deviates from its
SP with high frequency oscillations. These oscillations are leaking through to
downstream Control Variables that will exhibit similar oscillations. If the
oscillations exceed the CV funnel width, a controller not using CV
Measurement Compensation will make compensating moves to offset the CV
violations. Due to the nature of the oscillations, it is not desirable to react to
these, but instead exclude the effect of the oscillations from the CV
predictions. Refer to Figure 6 below.
• Type II: Unusual dynamic behavior of a CV response to movement in either
MV or DV. The unusual dynamic can include undesirable components such
as strong inverse response or strong lead type response, both of which can
cause unnecessary MV movement. An example of inherent inverse response
in shown in Figure 7 below.
• Type III: Disturbances with near zero gain effects. In certain situations it may
be desirable to allow the controller to absorb a disturbance by subtracting the
known effect of the offending disturbance. Certain type of event disturbances
falls into this category. An event disturbance is defined as, but not limited to
an abruptly and disrupting occurring event such as a sudden feed change,
semi batch mode switch or sudden closure/opening of a valve. It is
important that these disturbances have close to zero gain, to enable the
controller to ignore the excursion and wait until the CV is brought back within
its funnel.

August 2000, Rev. 2 6


RMPCT with Predict Back and CV Measurement Compensation White Paper

The principle of CV Measurement Compensation is illustrated in Figure-5. Figure 5


illustrates a Type I scenario where a MV exhibiting PV oscillations is used to control
downstream Control variables. The “PID Controller” boxes and the “Process” box
illustrates the MV controller action. A disturbance is added to the MV PV to illustrate
the effect of a load change such as line pressure (flow) or inlet temperature (temp)
affecting the MV PV.
The Process Model G'(s) (MV PV to CV) is derived from step tests using the MV PV as
the independent variable. This output of this model is compared to the actual CV
response and the resulting CV bias is filtered and added to the Control Model
predictions G”(s) that are based on the MV SP. The Control Model is a pseudo model
using the exact same model as the Process Model G’(s).

August 2000, Rev. 2 7


August 2000, Rev. 2

RMPCT with Predict Back and CV Measurement Compensation


Disturbance Disturbance

(CV LIMITS)

+ +
(MV SP) + (RAW CV)
PID Process
RMPCT Process
Controller G(s) +
+
-
Compensated CV prediction
8

(MV PV)

Figure 5

(MV PV)
Main RMPCT Model CV Compensation (CC) Model

Control G'(s) = G"(s) Process


Model Prediction
G"(s) G'(s)

White Paper
+
+ (Model error bias update) -
+
RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure 6 illustrates the closed loop control performance for a type I scenario with a
controller configured with CV Measurement Compensation. A CV target change was
introduced to illustrate the control performance.

Compensated CV Prediction

CV Process Predictions

MV.SP

MV.PV

Figure-6

It is noticed that oscillations in the MV PV are propagating on to the CV, which is


exhibiting similar oscillations. However MV moves are complete suppressed until the
CV target is changed. Upon changing the CV target the controller then raised the MV
SP in small discrete steps until the predicted compensated CV was within the width of
the funnel.

August 2000, Rev. 2 9


RMPCT with Predict Back and CV Measurement Compensation White Paper

Hydrogen
Quench

FIC

TIC

TIC MV
DV

TI
FIC
CV
Fuel Gas

An example of a type II problem is described below based on a reactor multivariable


control function. In reactor control design it is usually desirable to control individual
reactor bed differential temperatures within a given maximum limit for strongly
exothermic processes, based on safe operating guidelines and the catalyst deactivation
rate. The differential temperature is calculated directly from the bed inlet TIC PV and
the reactor outlet TI PV as the difference between the two. Embedded in this
supporting calculation is a combination of MV and CV dynamics. The nature of the
calculation and the process dynamics causes a strong inverse response when the bed
inlet temperature SP is increased. The length and magnitude of the inverse response
portion is a function of the dynamics:

• MV response to SP changes

• Deadtime across reactor bed

• The dynamic lag of the bed outlet temperature

CV Measurement Compensation can effectively be used to model and remove the


inverse response.

August 2000, Rev. 2 10


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure-7

Dynamic response of reactor bed differential temperature.

TIC PV

TIC SP (Reactor bed inlet temperature)

TI PV (Reactor bed outlet temperature)

TI PV - TIC PV (Reactor bed differential temperature)


Supporting calculation

August 2000, Rev. 2 11


RMPCT with Predict Back and CV Measurement Compensation White Paper

IMPLEMENTATION GUIDELINES
Predict-back: The decision to use Predict-back is typically made at the design stage.
Predict-back is normally recommended when:
• The MVs or DVs experience significant disturbances from measurable or non-
measurable sources and the disturbance rejection is slower than the effects of the
unrejected (leak-through) disturbances on the RMPCT controlled variables.
• The MV PID settling time is large relative (cascade controllers) to the RMPCT
execution period.
• The CV settling times (including deadtime) are large (coupled with 1 and/or 2)
• If tight control of CVs is required (reactor temperatures).
When the dynamics of the MV’s PID controller are fast relative to the controller
execution period, then disturbances to the MV are usually attenuated fast enough to
have little impact on the RMPCT CVs. Flow controllers are an example of “fast” MVs.
For example, it is usually not recommended to break a flow controller cascade and
adjust the valve position as an RMPCT MV.
The following is a partial list of MVs where using Predict-back can or may improve
controller performance:
1. Feed heater coil outlet temperature
2. Distillation Column top temperature
3. Reboiler return temperature
4. FCC feed preheat temperature
5. FCC riser outlet temperature
6. FCC regenerator air flow
7. FCC Gas Con strip-off flow
8. Hydrocracker reactor bed inlet temperature
9. Hydrocracker reactor WABT (for conversion control)

August 2000, Rev. 2 12


RMPCT with Predict Back and CV Measurement Compensation White Paper

CV Measurement Compensation: The decision to use CV Measurement Compensation


is typically made at the design stage. CV Measurement Compensation is normally
recommended when:
• Known MVs or DVs experience significant disturbances from measurable or non-
measurable sources and the disturbance rejection is slower than the effects of the
unrejected (leak-through) disturbances on the RMPCT controlled variables. In
addition the PVs of these control loops often don’t return to SP, but instead exhibit
cyclic behavior and/or hysteresis. The CV excursions due to effects described above
are well beyond the noise range and causes frequent violations of CV limits,
resulting in unwarranted MV moves, which adversely affects control performance
(non-minimum MV movement).
• The CV response exhibit up-front dominant inverse or lead type response, and it is
acceptable to allow a transient violation, after which the CV is brought back within
the funnel width (refer to use of calculated CVs).
• When tight control of CVs is not required (surge drum level). CV Measurement
Compensation allows the controller to ignore the offending disturbances and allow
these to propagate though to the CV without any MV movement.
The following is a partial list of MVs and CVs where using CV Measurement
Compensation can improve controller performance:
1. FCC riser outlet temperature (affected CVs are related fractionator loading
such as wet gas compressor flow or amps)
2. Valve position CVs in general
3. Hydrotreater/cracker calculated differential temperatures
4. Calculated or measured CVs with dominant inverse or lead type response
characteristics

August 2000, Rev. 2 13


RMPCT with Predict Back and CV Measurement Compensation White Paper

MODEL IDENTIFICATION
Models used for Predict-back are broken down into three categories:

• Predict-back Models
• Estimated Disturbance Models
• Main RMPCT Models

For simplicity, first order transfer functions have been used in the following illustrations.
Higher order models including lead terms are possible provided that consistency among
the model types is preserved.

Predict-back Models
These models are used to predict the MV PVs or DV PVs. These are simply the
models of the MV.PV response to its’ own SP change, as shown in equation 1. Notice
that the process gain is exactly 1 by nature, because the PV always returns to its SP,
unless the controller is wound up.
−td 1
s
dMV.PV 1e
=
dMV.SP T1s + 1 [1]

The Predict-back models may also include other inputs that affect the MV.PV. Normally
these additional disturbance models have a zero gain in steady state and are
sometimes referred to as “hump” models or just zero-gain models. Equation 2 is an
example of a “hump” model used in this context. The purpose of these models is to
predict the effect of external disturbances to the MV PV so that the controller can
anticipate the effect and take action even before the disturbances hit the MV.
−t d 2

dMV.PV K p 2 s ⋅ e
s

=
dX T2 s +1 [2]

Where X = other MV or DV disturbance input

August 2000, Rev. 2 14


RMPCT with Predict Back and CV Measurement Compensation White Paper

Estimated Disturbance Models


The estimated disturbance models are simply the models between the CV and the
MV.PV, and are shown in equation 3. These models are used to map the prediction
error of models [1] and [2] onto the CVs (i.e., the prediction error of the MVs is used to
estimate the disturbance on the CVs).
−t d 3
s
dCV .PV K p 3e
=
dMV .PV T3 s + 1 [3]

Main RMPCT Models:


These are the models that we are all familiar with. They represent the effect on the
CV.PV by the MV SPs and DV SPs. These models are generated by the convolution of
equations 1 and 3 as illustrated in equation 4.
−t d 1 −t d 3 −(t d 1 + t d 3 )
s s
dCV.PV dMV.PV dCV.PV 1e K p3e K p3e s
= = =
dMV.SP dMV.SP dMV.PV T1s + 1 T3s + 1 (T1s + 1)(T3s + 1)
[4]
If the Predict-back models include other inputs described by equation 2, then these
additional zero-gain “hump” models must also be included in the Regular RMPCT
Models to maintain consistency. These models are generated by the convolution of
equations 2 and 3 as illustrated in equation 5.
− td 2 − td 3 − ( td 2 +td 3 )

dCV.PV dMV.PV dCV.PV K s ⋅ e K e K K s⋅e s s s

= ⋅ = ⋅ = p2 p3 p2 p3

dX dX dMV.PV T s +1 T s +1 (T s + 1)(T s + 1)
2 3 2 3

[5]
Where X = other MV or DV disturbance input

August 2000, Rev. 2 15


RMPCT with Predict Back and CV Measurement Compensation White Paper

Typically, Predict-back models only include those described by equations 1, 3 and 4.


In some cases additional models are needed (equation 5) to describe to interactions in
a series of similar process segments (f.ex. multiple reactor beds).
During the “Build Controller” procedure in Profit Design Studio, model consistency is
checked. Particular attention should be paid to the Message file to ensure that no
modeling errors or inconsistencies exist. If inconsistencies are present, the model
builder will usually try to make adjustments in either gains or deadtimes, however the
model builder does currently not handle inconsistencies in the dynamic part. If any
warnings related to Predict-back are issued, it is recommended to fix the problem
before building the final model. Ignoring the warnings can in some cases cause a
model mismatch, and thus later on adversely affect controller performance. Minor
differences in the gains due to round off errors can usually be ignored even though a
warning message is usually created.
Although defining all three of these model types is an over-specification of the system
rd
(i.e., only 2 of the 3 models are required to completely specify the system, the 3 model
can be calculated), Profit Design Studio requires the specification of all three models.

August 2000, Rev. 2 16


RMPCT with Predict Back and CV Measurement Compensation White Paper

Zero-gain RMPCT Models:

Zero-gain or “hump” models can be represented in two different ways:

Ks + 1
G(S ) = k
τs + 1 [6]

Ks
G ( s) = k
τs + 1 [7]

The main difference between the two representations is that only [7] is a true zero-gain
model while [6] has a non zero steady state gain, where the value of the offset depends
on the values of the parameters. Consider the following transfer function [8]:
100 s + 1
G ( S ) = 0 .1
(10 s + 1)(5 s + 1) [8]

The limit value of [8] using the final value theorem is 0.1.

Now consider the following transfer function:

100 s
G ( S ) = 0 .1
(10 s + 1)(5 s + 1) [9]

The limit value of [9] using the final value theorem is exactly 0.
This example is illustrated below, with transient response curves to a step input. In this
case both models are adequate representations of the response of a manipulated
variable process value to a step input in a DV. The general recommendation is to use a
true zero gain model to represent the transient response in the MV PV to avoid conflicts
with the steady state optimizer. The model parameters can manually be adjusted via
the model summary page to force a zero steady state offset. Note, that if the results
from the FIR model identification indicate a steady state gain that is significantly
different than zero, forcing the gain to zero may cause a major dynamic mismatch
between the original FIR model and the corrected model. In this case the user should
revisit the results from the model identification and if necessary go back collect
additional process data for further analysis. Note that if a MV is in a wind-up state then
the predictions of MV PV will not return to its SP. Any “hump” model between a DV and
a CV will not reflect this and a model mismatch is introduced. In addition, the anti-
windup feature for the MVs is permanently disabled when using the Predict-back
feature for the MV.

August 2000, Rev. 2 17


RMPCT with Predict Back and CV Measurement Compensation White Paper

Mathcad Worksheet Example:

RMPC with Predict-Back


Illustration of zero gain models for disturbance inputs

Non zero gain model GNZ(s) with 0.1 steady state offset:
100 s 1
0.1 invlaplace , s 1.8. exp( .1. t ) 1.9. exp( .2.t ) .1
( 10. s 1 ) ( 5 s 1 ) s
GNZ( t ) 1.8.exp( .1. t ) 1.9. exp( .2. t ) .1
Zero gain model GZ(s) with no steady state offset:
100 s
0.1 invlaplace , s 2.. exp( .2.t ) 2.. exp( .1.t )
( 10. s 1 ) ( 5 s 1 ) s
GZ( t ) 2.. exp( .2. t ) 2..exp( .1. t )
1

GNZ( t )
0.5
GZ( t )

0
0 20 40 60 80 100
t
The two curves represents the transient response to a step input.
The solid represents the response of the non zero gain model.
The dotted curve represent the response of the zero gain model.

When “hump” models for the relevant MVs have been identified this way, the
corresponding “hump” models for the CVs can be calculated using a convoluted model.
If we select [9] as the model between the DV and the MV then assume the following
estimated disturbance model between the MV PV and the CV:

1
G(S ) = 2 [10]
5s + 1
The convoluted disturbance model for the CV is:
1 100 s
G(S ) = 2 * 0 .1 [11]
5s + 1 (10 s + 1)(5 s + 1)
It is assumed that the disturbance model [11] has been identified prior to this. It is
recommended to validate that the convoluted model predicts the CV response well.

August 2000, Rev. 2 18


RMPCT with Predict Back and CV Measurement Compensation White Paper

Models used for CV Measurement Compensation are broken down into two categories:

• CV Compensation (CC) Models

• Main RMPCT Models

CV Measurement Compensation Models


The “CV Compensation Models” are used to predict the CV response based on the MV
PVs or DV PVs. Figure 5 illustrates the use of this model as G’(s) or simply referred to
as the “CV Compensation (CC) model”.
The “Main RMPCT Models” are use to predict the CV response based on the MV SPs
or the DV SPs. These are the models that we are all familiar with. They represent the
effect on the CV.PV by the MV SPs and DV SPs.
The two models are exactly identical but used for different purposes in the controller.

August 2000, Rev. 2 19


RMPCT with Predict Back and CV Measurement Compensation White Paper

MODEL BUILDING
Predict-back: APCDE is used to configure a controller with Predict-back. The Regular
RMPCT model file (i.e., *.mdl) is still the main file used to develop the *.cfg, *.xm, *.xs,
etc. files to be used in the on-line controller. Figure 8 illustrates the “Build Controller”
Options that need to be specified for Predict-back.

Figure 8: APCDE Configuration of Predict-back

No special consideration is required to implement the on-line controller when using


Predict-back as long as the off-line “Build Controller” step included the Predict-back
option. To verify that the XM file used on-line is configured with the Predict-back option,
view the LCN or Profit Viewer “MV Control” display and verify that “PB” is displayed in
the right column for all MVs using Predict-back. The on-line displays do not indicate
which CVs are using Predict-back, if the disturbance model for the CV was left blank in
the estimated disturbance model (mdl) file, then essentially predict back has been
disabled for that particular CV (no warnings are given in the model builder, if the user
chooses not to or forgets to add the model). Figure 9 shows an example of the LCN
based RMPCT “MV Control” display. Notice the “PB” for MVs number 3 and 4.

August 2000, Rev. 2 20


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure 9: RMPCT “MV Control” Display with Predict-back Option for MVs 3 and 4

The three model files discussed in the previous section are mandatory in order to
configure Predict -back. Specific dimensions are required for the different model files
based on the number of MVs, DVs and CVs. Table 1 illustrates the dimension and
input/output requirements for each model file.

Table 1: Predict-back Model Dimensions and Variable Source

Model File Model Dimensions2 Variable Type


(typical)
Independent Dependent Independent Dependent

Predict-back (nMV+nDV) (nMV+nDV) SP PV

Estimated (nMV+nDV) nCV PV PV


Disturbance

Main (nMV+nDV) nCV SP PV

2
nMV = # of MVs; nDV = # of DVs; nCV = # of CVs;

August 2000, Rev. 2 21


RMPCT with Predict Back and CV Measurement Compensation White Paper

Identification of Models:
When Predict-back is used, it can be applied to any subset of MVs and CVs within the
control problem. For instance, in the 3 MV by 5 CV and 1 DV controller illustrated
below, you can choose to only apply Predict-back to MV #2 and CVs # 1,3 and 4, even
though there are 5 CV models for that MV in the Main model file.

Main Models3

MV1.SP MV2.SP MV3.SP DV1.PV


CV1.PV √ √ √ √
CV2.PV √ √ √
CV3.PV √ √ √
CV4.PV √ √ √ √
CV5.PV √ √
Predict-back Models
MV1.SP MV2.SP MV3.SP DV1.PV
MV1.PV
MV2.PV √ √
MV3.PV

Estimated Disturbance Models

MV1.PV MV2.PV MV3.PV DV1.PV


CV1.PV √
CV2.PV
CV3.PV √
CV4.PV √
CV5.PV

3
“√” indicates that a model exists for this CV/MV pair

August 2000, Rev. 2 22


RMPCT with Predict Back and CV Measurement Compensation White Paper

A systematic method to develop the controller models is to:


1. Organize your MVs, DVs and CVs into your Main controller file like normal (i.e.,
main.mdl)
2. Copy main.mdl for use as the Predict-back model file (i.e., pb.mdl)
3. Copy main.mdl for use as the Estimated Disturbance model file (i.e., ed.mdl)
4. In pb.mdl, delete the CVs and paste in the MV.PVs. The MV.PVs should be defined
as type CV.
5. Identify the pb.mdl models for the appropriate MVs and any zero-gain model for
disturbance MVs or DVs, making sure to select final trials and to save the file4.
6. Identify the ed.mdl models for the appropriate CV/MV combinations, making sure to
select final trials and to save the file.
7. Identify the main.mdl models for all CV/MV combinations. Over-ride the appropriate
CV/MV models using the convolution models described in the previous section.
Compare the convoluted models to the FIR models as a reality check.
8. Over-ride the appropriate (if any) CV/DV model using the convolution models
described in the previous section. Compare the convoluted models to the FIR
models as a realty check.
9. Select final models in main.mdl and save the file.
10. Build the controller using the “Options” to specify the Predict-back and Estimated
Disturbance models.
11. Check the bottom of the Message file after building the controller for any warnings.
The build step verifies that the models are consistent and by default will internally
adjust the estimated disturbance model to make the models consistent. The source
models should be investigated to ensure that the proper model file is fixed.

4
The final models need to be saved in each case as the “Build Controller” step reads the file versions from disk.

August 2000, Rev. 2 23


RMPCT with Predict Back and CV Measurement Compensation White Paper

Predict-back Tuning:
Predict-back feedforward corrections can be turned off, by setting the Predict-back ratio
parameter for the MV to zero. Normally a Predict-back ratio = 1 is recommended for full
Predict-back feedforward corrections. A fractional value less than 1 is recommended to
de-tune the feedforward corrections to the CV. Note; a value different than 0 or 1 will
introduce mismatch in the CV predictions.
Essentially the disturbance model gain is multiplied with the Predict-back ratio to
attenuate the feedforward prediction corrections to the CV when the MV PV deviates
from its SP.
Note; the Predict-back ratio is not a gain-scheduling tool, instead the online gain delay
factors should be used to make corrections to the gains, or alternatively recreation of
the XM and XS files.
It is recommended to lower the Predict-back ratio if MV control moves are deemed to
vigorous in the presence of disturbances, and larger transient errors in the CVs can be
tolerated. However the tuning of this factor is mostly ad hoc based on the observed MV
and CV response and normally a value of 1 is recommended.

August 2000, Rev. 2 24


RMPCT with Predict Back and CV Measurement Compensation White Paper

CV Measurement Compensation: APCDE is used to configure a controller with CV


Measurement Compensation. The Regular RMPCT model file (i.e., *.mdl) is still the
main file used to develop the *.cfg, *.xm, *.xs, etc. files to be used in the on-line
controller. Figure 10 illustrates the “Build Controller” Options that need to be specified
for CV Measurement Compensation.
Figure 10: APCDE Configuration of CV Measurement Compensation

No special consideration is required to implement the on-line controller when using CV


Measurement Compensation as long as the off-line “Build Controller” step included the
CV Measurement Compensation option. The on-line displays do not indicate which
CVs are using CV Measurement Compensation, if the CV Compensation model for the
CV was left blank in the CC model (mdl) file, then essentially CV Measurement
Compensation has been disabled for that particular CV (no warnings are given in the
model builder, if the user forgets to add the model).

August 2000, Rev. 2 25


RMPCT with Predict Back and CV Measurement Compensation White Paper

Table 2: CV Measurement Compensation Model Dimensions and Variable Source

Model File Model Dimensions5 Variable Type


(typical)
Independent Dependent Independent Dependent

CV Compensation (nMV+nDV) nCV PV PV


(CC)

Main (nMV+nDV) nCV SP PV

A systematic method to develop the controller models is to:


1. Organize your MVs, DVs and CVs into your Main controller file like normal (i.e.,
main.mdl)
2. Copy main.mdl for use as the CV compensation model file (i.e., cc.mdl)
3. Identify the cc.mdl models for the appropriate CV/MV combinations, making sure to
select final trials and to save the file.
4. Identify the main.mdl models for those CV/MV combinations not using CV
Measurement Compensation. Type in the models from cc.mdl for those CV/MV
combinations using CV Measurement Compensation.
5. Select final models in main.mdl and save the file.
6. Build the controller using the “Options” to specify the Predict-back and CV
Measurement Compensation Models.
7. Check the bottom of the Message file after building the controller for any warnings.
The build step verifies that the models in the two files are pairwise identical and by
default will internally adjust the CC model (gain and/or deadtime) to make the
models consistent.

CV Measurement Compensation Tuning:


CV Measurement Compensation predictions can be turned off, by setting “PV VALUE
COMP RATIO” ratio parameter for the CV to zero. Normally a ratio = 1 is
recommended for full predictions.
Essentially the CV Measurement Compensation model gain is multiplied with the ratio
to attenuate the prediction to the CV. A ratio less than one will introduce a model
mismatch in G’(s) (refer to Figure 5). The model error bias will be incorporated into the
Main model predictions, and in the absence of MV moves the open loop predictions will
reflect the effect of movement in MV PVs or in DV PVs.

5
nMV = # of MVs; nDV = # of DVs; nCV = # of CVs;

August 2000, Rev. 2 26


RMPCT with Predict Back and CV Measurement Compensation White Paper

EXAMPLES
Predict-back:
Example 1: Simple 1 x 1 controller
Figures E1-1 and E1-2 are APCDE screen dumps that illustrate the main (main.mdl),
predict-back (pb.mdl) and estimated disturbance (ed.mdl) models and variable
configurations.

Figure E1-1: Models

August 2000, Rev. 2 27


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure E1-2: Variable Configurations

August 2000, Rev. 2 28


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 2: Hydrocracker Reactor Outlet Temperature Control


The Hydrocracking process can exhibit highly exothermic reactions. A typical
hydrocracker configuration is presented in Figure E2-1. Feed is heated under heater
outlet temperature control in the feed heater and enters the first bed of the reactor. The
temperature rises due to the exothermic cracking reaction in bed 1, so bed 1 outlet
temperature is higher than bed 1 inlet temperature. The hydrocarbon exiting bed 1 is
cooled using hydrogen quench via the bed 2 inlet temperature controller. Often there
are 4 or 5 beds contained in a single reactor, with each bed inlet temperature control
cascaded to hydrogen quench flow controllers.

Figure E2-1: Typical Hydrocracking Reactor Process

Hydrogen
Quench

FIC

TIC

TIC MV
DV

TI
FIC
CV
Fuel Gas

The objective of the Profit Controller application presented in this example is to provide
control of bed 2 outlet temperature. As in the previous example, any deviation of a
bed’s inlet temperature PV from its SP will propagate through to its respective outlet
temperature. Since the beds are in series, changes in bed 1 outlet temperature PV
result in a load disturbance on bed 2 inlet temperature and thus also affect bed 2 outlet
temperature. Figure E2-2 illustrates the dynamic response of the bed 1 inlet and 2
inlet/outlet temperatures when bed 1 inlet temperature setpoint is changed:

August 2000, Rev. 2 29


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure E2-2: Hydrocracker Process Response

The important things to note here are:


• The hump shaped response of bed 2 inlet temperature (Tin2PV) when the heater
outlet temperature (ThtrSP) is stepped.
• How bed 2 outlet temperature (Tout2PV) follows Tin2PV.
There are two issues that should be addressed in the design of this controller:
1. Account for unexpected deviations of Tin2PV from its setpoint
2. Account for the expected deviation of Tin2PV from its setpoint when ThtrPV
changes
Issue 1. is handled by the traditional Predict-back design outlined in Example 1.
Issue 2. is a bit trickier, and is handled via the model transfer functions illustrated in
equations [2] and [5]. We can see from Figure E2-2 that when ThtrSP is changed,
ThtrPV changes accordingly, which then propagates through the first bed and applies a
load to Tin2PV. Tin2PV deviates from its SP due to the load disturbance then returns

August 2000, Rev. 2 30


RMPCT with Predict Back and CV Measurement Compensation White Paper

back to its SP on PID control. Tout2PV then follows Tin2PV subject to the process
dead-time and lag.
As stated earlier, hydrocracker reactions can be highly exothermic. If the inlet
o
temperature control cannot quickly reject disturbances, a 1 F disturbance in the bed
inlet temperature can propagate through the reactor beds resulting in a +10oF
disturbance in the reactor outlet temperature. Coupling this disturbance with
feed/effluent heat exchange (most designs have this) returns a disturbance back to the
bed 1 inlet temperature which can lead to unstable reactor control in the best case or a
run away reaction and catastrophic metallurgical failure in the worst case. Anything that
can be done to attenuate temperature disturbances within a reactor system improves
the control stability.
From Figure E2-2 we can see that there is feed forward potential in knowing that
ThtrPV has changed and that it will affect Tin2PV and subsequently Tout2PV. With this
feed forward information, Tin2SP could be changed to counteract the disturbance and
minimize the effect on Tout2PV. This feed forward effect is configured into RMPCT by:
• Adding the effect of ThtrPV on Tout2PV to the main model file (hcr_main.mdl). This
is a “hump” model. This provides feed forward information to the CV predictions.
Provided that the performance ratio of Tout2PV is sufficiently aggressive, control
moves will be made to minimize this disturbance.
• Adding the effect of ThtrPV on Tin2PV to the Predict-back model file (pb.mdl). This
is a “hump” model. Since the direct effect of the ThtrPV disturbance is already
mapped to Tout2PV as described in item 1, this disturbance needs to be mapped
onto the Tin2PV predictions to eliminate any prediction error in the MV (i.e., to avoid
any double dipping). Should Tin2PV not respond as predicted, its prediction errors
will be mapped onto the Tout2PV predictions using the typical Predict-back
mechanism.
• Notice, the predict back model between the heater SP (ThtrSP) and its PV (ThtrPV)
is not present. The reason for this is the fact that only the heater PV is used for
feed-forward predictions, which makes the need to use predicts back for the heater
DV obsolete since the heater SP is not used for predictions. In most cases of this
nature, the heater temperature would be a MV and predict back would have to be
included for this variable.
The following figures illustrate the Predict-back model configurations used for this
example:

August 2000, Rev. 2 31


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 2: Figure E2-3 Models

August 2000, Rev. 2 32


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 2: Figure E2-4 Variable Configurations

August 2000, Rev. 2 33


RMPCT with Predict Back and CV Measurement Compensation White Paper

CV Measurement Compensation:
Example 3: Simple 2 x 1 controller
Figures E3-1 and E3-2 are APCDE screen dumps that illustrate the main (main.mdl),
and CV Measurement Compensation (cc.mdl) models and variable configurations.

Figure E3-1: Models

August 2000, Rev. 2 34


RMPCT with Predict Back and CV Measurement Compensation White Paper

Figure E3-2: Variable Configurations

August 2000, Rev. 2 35


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 4: Hydrocracker Reactor Differential Temperature Modeling


Please refer to Example 2 and Figure E2-1. As discussed above it is recommended to
use CV Measurement Compensation for differential temperatures that exhibit dominant
inverse response. The bed inlet temperature controller TC4321 PV affects both bed
outlet temperature and bed differential temperature. The dynamic responses are
shown below:
Figure E4-1: Hydrocracker Differential Temperature Process Response

August 2000, Rev. 2 36


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 4: Figure E4-2 Models

August 2000, Rev. 2 37


RMPCT with Predict Back and CV Measurement Compensation White Paper

Example 4: Figure E4-3 Model Predictions

August 2000, Rev. 2 38

You might also like