Thanks to visit codestin.com
Credit goes to www.scribd.com

0% found this document useful (0 votes)
37 views41 pages

Principle of Process Control

The document discusses the principles of process control, focusing on block diagrams, control system classifications, and the behavior of processes in both static and dynamic states. It outlines the components of closed-loop systems, the types of controller actions (P, I, D), and the importance of tuning for optimal performance. Additionally, it includes exercises and examples to illustrate key concepts in process control and optimization.

Uploaded by

ARJUN ABUBAKAR
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
37 views41 pages

Principle of Process Control

The document discusses the principles of process control, focusing on block diagrams, control system classifications, and the behavior of processes in both static and dynamic states. It outlines the components of closed-loop systems, the types of controller actions (P, I, D), and the importance of tuning for optimal performance. Additionally, it includes exercises and examples to illustrate key concepts in process control and optimization.

Uploaded by

ARJUN ABUBAKAR
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 41

Shell Global Solutions

Principle of Process Control

By Somboon Pattanawijit GSEO


Process Control & Optimization Group
28 September 2005 (rev. 1.0)
Table of contents
• Introduction
• Process Behaviour
• Controller actions and their effect on control
• The Closed Control Loop
• Characterisation of a Process
• Optimum Controller settings

Shell Global Solutions 2


Introduction – Block Diagrams
• Block diagram is a short hand, pictorial
representation of the relationship between
input and output of a system
Solid

• The simplest form of a block diagram is the


single block with one input and one output

Input Output

Block
Shell Global Solutions 3
Introduction – Block Diagrams
• The input is the stimulus or excitation from an
external energy source applied to a system in
order to produce a specified response from
Solid

this system
• The output is the actual response obtained
from the system

Control Valve, Temperature,


Fan speed, Flow, Level,
etc. Process etc.

Shell Global Solutions 4


Introduction – Block Diagrams
• The output is the actual response obtained
from the system
Solid
particle
Fi h
Process
Fi

ρA dh = ρFi - ρFo
dt
dh
kA = kFi - h ; Fo=h/k
h Fo dt

H(s) = K
Laplace Transform : = K
Fi(s) kAs+1 τs+1
Shell Global Solutions 5
Introduction – Block Diagrams
• Block diagram of a control valve
Solid
particle

MV [bar]

Flow [m3/s] MV [bar] Flow [m3/s]


Valve

Shell Global Solutions 6


Introduction – Block Diagrams
• Block diagram Operations
z
Solid
particle particle particle
+
x + x+y x + x-y x + x+y+z

+ - +

y y y

Summing point

Take-off point

Shell Global Solutions 7


Introduction – Classification of Control Systems
• Classified into 2 general categories:
• open-loop system
particle
The control action is independent of the process output
• close-loop system
The control action is somehow dependent on the process output

• Example
Is automatic toaster an open/close loop control
system? Why?

Shell Global Solutions 8


Introduction – Block Diagrams of a Feedback
Control system
• The blocks representing the various of a
control system are connected in a way which
characterises their functional relationship
Solid
particle

within the system.


Manipulated
Set Value Deviation Variable
(SV) (θ) (MV)
Controller Actuator Process
+

Process Variable
(PV)
Transmitter

Shell Global Solutions 9


Introduction – Block Diagrams of a Feedback
Control system

Feed flow
Solid
particle

Fuel flow

TT
MV PV
TC

SV

SV MV F T
[mA] θ [mA] [t/d] [oC]
Controller Actuator Process
+
− PV
[mA]
Transmitter

Shell Global Solutions 10


Introduction – Terminology
• Feedback control action
• Feed forward control action
particle

• Open-loop
• Closed-loop
• Direct-acting controller
• Reverse-acting controller
• Failure open or Air to close control valve
• Failure closed or Air to open control valve
• DCS – Distributed Control System

Shell Global Solutions 11


Introduction – Terminology
Flue gas

Controlled
condition Steam
Water
Solid
particle
Furnace
Air
PT

FT
Feedback Correcting Feedback
loop condition loop
D
FRC MV PV
PRC
R
Feed forward SV
loop X
Simple boiler
FT Control scheme

Fuel Shell Global Solutions 12


Introduction – Exercises
• What is the control system type called if the
temperature controller in the furnace control figure is
in the manual mode?
particle

• If the fuel control valve is “failure closed” what should


be the action of the temperature controller (direct or
reverse)?
• If, in a DCS, the controller output changes from 25%
to 75% full scale, what would the equivalent output
signal change be if the controller were:
a: a pneumatic controller?
b: an electronic controller?
Shell Global Solutions 13
Process Behaviour
• There are two different types of behavior can be
distinguished. These are:
•Static behavior
Refers to overall change in the output of the process with
respect to certain input change when the process has reached
its steady state or equilibrium (i.e. has stopped moving).
•Dynamic behavior
Refers to the behavior of the process when it is in the transient
or ‘moving’ state.

Shell Global Solutions 14


Process Behaviour - Static
• A control loop in principle comprises the following
elements:
•Process

•Transmitter

•Controller

•Control valve

• The static gain, G, of an element is defined as:

∆y ∆x
Change in the Output ∆x
G= Change in the Input
G = ∆y

Shell Global Solutions 15


Process Behaviour - Static
• Suppose and electronic temperature transmitter
(output range 4-20 mA) has a range of 200-300 oC.
It generates a signal of 4 mA at 200 oC and 20mA at
300 oC,. It is said to have a span of 100 oC. The
transmitter has a linear characteristic.

• What is the static gain?

Shell Global Solutions 16


Process Behaviour - Static
• The product of all the gains in the control loop is
called the static open loop gain GL.
∆PV = ∆θ.Gcc.Gvv.Gpp.Gtt
GLL = ∆PV = Gcc.Gvv.Gpp.Gtt
∆θ
• Thus, for the closed loop gain.
∆θ = ∆SV−∆PV
Note
Note that
that the
the control
control loop
loop will
will
∆PV = GLL be
be unstable
unstable ifif the
the closed
closed loop
loop
∆SV 1+GLL gain
gain >1.0
>1.0

Shell Global Solutions 17


Process Behaviour - Static
• With reference to the combustion process, the product
of the gains Gv.Gp.Gt can be combined as below:
Controller
SV MV F T
[mA] θ [mA] [t/d] [oC]
Gc=0.5 Gv=0.1 Gp=10
+

PV
[mA]
Gt=2

Controller
SV MV T
[mA] θ [mA] [oC]
Gc=0.5 Gp=2
+

PV
[mA]

Shell Global Solutions 18


Process Behaviour - Static
• In process control calculation it is often assumed that
the gain of each element in a control loop is constant
through the elements’s operating range. In practice
this is rarely the case.
• Example
• Flow transmitter follows a squared relationship for DP cell
• Control valve can be linear, equal-percentage
• Unless precautions are taken the loop gain of a
control system may change if the operating point
shifts. The controller gain may then need retuning to
provide satisfactory control.

Shell Global Solutions 19


Process Behaviour - Dynamic
• The dynamic behavior determines how fast and in
what manner the process responses to an input
change from the original steady-state to the final one.
• There are many types of dynamic behavior and
processes are generally described by a combination
of a number of basic dynamic elements.

Shell Global Solutions 20


Process Behaviour - Dynamic
• Integrator

x(t)
Reference
Reference level
level

u(t)
u(t).dt = A.dx
1
11 tan α =
t 00
α A
00
tt00 1 tt00

A0∫ 1.dt

Shell Global Solutions 21


Process Behaviour - Dynamic
• Deadtime Steam
injection
TC
TC

FF
T1 2 m. T2

Steam flow

time
T1

time
T2 D

time
t0 t1
Shell Global Solutions 22
Process Behaviour - Dynamic
• First-order lag
∆OP(t) First Order ∆PV(t)
Process

Input to the process: ∆OP(t)

Response of the Process: ∆PV(t)

  t   (t)
∆ PV(t) = 1 − e p  * ∆ OP (t)
 τ 
 

Shell Global Solutions 23


Process Behaviour - Dynamic
445 86

Step (∆OP) 440


84

Response (∆PV)
435
95% 82

430 86%
80

425 63.2% OP
PV
78
420

76
415

74
410

405 72
1 11 21 31 41 51 61 71 81 91 101 111 121 131 141
Time
τd τp τp τp

Shell Global Solutions 24


Process Behaviour - Dynamic
• Combination of deadtimes
u1 u2 u3 x
D1 D2 D3

u1(input)

u2
D1

u3
D2
x (output)
D3

D=D1+D2+D3

Shell Global Solutions 25


Process Behaviour - Dynamic
• Combination of first-order lag
u1 u2 u3 x
τ1 τ2 τ3

u1 Unit step

1st-order
u2

u3 2nd-order

x 3rd-order

Shell Global Solutions 26


Process Behaviour - Dynamic
• Open-loop step response
Output PV
Input step MV
TT
Gp, τ, D

Input step MV
1stst order
mathematic model
Output PV
∆MV
∆PV Kp
G p ( s) = e − sD
τs + 1
D τ
time
t1 t2 t3

Shell Global Solutions 27


Process Behaviour - Dynamic
• Exercise
• What is the static gain of an electronic pressure
transmitter with an input scale from 10 to 20 bars?
• If a process is described by a first-order lag and takes
15 minutes to reach 95% of its final value, what is the
first-order time constant τ ?

Shell Global Solutions 28


Controller actions and their effect on control
• We have seen that a closed-loop control system
consists essentially of a process and a controller. The
controller automatically adjusts the input to the process
in response to a change in signal from the process
output or to a change in its set value.
• There are three kinds of controller action in a PID
controller (PID algorithm):

P: Proportional action
I: Integral or Reset action
D: Derivative or Rate action

Shell Global Solutions 29


Controller actions and their effect on control
• P: Proportional action

u = K ce Proportional Band
PB = 100/Kcc
e = SV − PV
SV
SV

time
time
uu

time
time

Shell Global Solutions 30


Controller actions and their effect on control
• I: Integral action only
t
Kc
u=
τi ∫ e( s )ds
0
SV
SV

time
time
uu

Integral
Integral action
action only
only

time
time

Shell Global Solutions 31


Controller actions and their effect on control
• PI: Proportional + Integral action

 1
t

u = K c e + ∫ e( s )ds 
 τi 0 
SV
SV

time
time
uu Proportional
Proportional ++ Integral
Integral action
action
Change
Change due
due to
to integral
integral action
action

ττii time
time

Shell Global Solutions 32


Controller actions and their effect on control
• D: Derivative action only

de
u = K cτ d
dt
Deviation
Deviation

time
time
uu

Derivative
Derivative action
action only
only

time
time

Shell Global Solutions 33


Controller actions and their effect on control
• PD: Proportional + Derivative action

 de 
u = K c e + τ d 
 dt 
Deviation
Deviation

time
time
uu Proportional
Proportional ++ Derivative
Derivative action
action

Proportional
Proportional action
action only
only
Change
Change due
due to
to the
the Proportional
Proportional action
action

ττdd time
time

Shell Global Solutions 34


PID Algorithm

Integral Derivative
Proportional

 1
t
de 
u = K c × e + ∫ e( s )ds + τ d 
 τi 0 dt 

Shell Global Solutions 35


Controller actions and their effect on control
• Desired response
Desired behavior or response must be decided upon before loop can be tuned.
For example: a protective anti surge controller needs different (faster) tuning
then cut point temperature controller
Response below could be in general considered as “good” (fast and well
damped).

One Overshoot

Set Value

One Undershoot

Process Variable

Time

Response to a Step Change

Shell Global Solutions 36


Open loop tuning procedure
For First Order Processes – not for Integrators:
1. Go to stable and representative operating point.
2. Put controller on manual and make step wise change to output.
3. Discuss first magnitude of step change with panel operator. If change too
small; no response. If change too big; plant upset.
4. Do two steps in opposite directions.
5. Record (“paper and pen” or via ODS system) responses. No data
compression and check sampling rate ODS system (might be too slow).
6. Take average response. If too much difference between steps, do another
step. Etc.
7. Estimate Process Gain, Process dead time and Process time constant.

Shell Global Solutions 37


Tuning Rules
• There are many published tuning rules/guidelines
commercially/freely available.
• Most of tuning rules have a suggested range of
applicability based on the ratio of d/τ.
• Tuning rules do not work for integrating processes
(e.g. level control).
• Remember these tuning rules are based on
approximation models and thus should be a starting
point in the tuning process.

Shell Global Solutions 38


Ziegler-Nichols Step Response Method

The point where the slop of the


Gain step response has its maximum is
(Kp )
first determined, and the tangent
at this point is drawn a shown in
the figure.

Dead Time Lag Time


(τ) Recommended
Recommended Range
Range of
of Application
Application 1.0<
1.0< (d/τ)
(d/τ) <0.1
<0.1
(d)
Mode
Mode KKcc ττii ττdd
1 τ 
PP  
Kp d 
0.9  τ 
PI   3.33d
PI Kp d 
1.2  τ 
  2.0d 0.5d
PID
PID Kp d 

Shell Global Solutions 39


Typical Tuning values
- Typical tuning values good enough for most loops
- Probably not good enough for the more difficult loops which operations will ask
you to look at. Easy loops do not give problems !

Gain (Kc) Integral time (τi) Derivative time (τd)


- Flow 0.3 - 0.7 0.2 – 0.5 0
- Pressure – liquid 0.3 – 0.7 0.2 – 0.5 0
- Pressure – gas 5 5 0
- Level – tight 1.5 – 3.5 10 0
- Level – buffer 1 20 0
- Temperature 1 5 - 20 0–2

Shell Global Solutions 40


Tuning Exercise
• If an input change of 20oC is required to give an
output change of 50% in a DCS and the temperature
transmitter range is 100oC:
particle

A. what is the controller gain? And


B. what is the Proportional band of the controller

Shell Global Solutions 41

You might also like