Instrumentation
and
Process Control
Lecture – 4
Overview of control, SISO, MIMO, Process Variables
Dr. Hemanth Kumar Tanneru
Department of Chemical Engineering
IIPE Visakhapatnam
Summary of previous lectures
• Why is process control needed?
• Main components needed to implement control?
Lecture 4 : Instrumentation and Process control 2
Summary (cont…)
• Safety (Stability of the system)
• Economics (Optimize the performance of the plant)
• Regulatory rules (Environmental impact)
Lecture 4 : Instrumentation and Process control 3
Summary (cont…)
• Control problems
• Servo
• Regulatory
• Control strategies
• Feed back
• Feed forward
• Process variables
• Controlled variables
• Manipulated variables
Lecture 4 : Instrumentation and Process control 4
In this lecture
Overview on how individual controllers get their set points
MIMO control systems
Why are SISOs popular?
Justification on how MIMO systems can be viewed as SISO in terms of control
Process variables
• State variables
• Input and output variables
Review of control terminologies
Lecture 4 : Instrumentation and Process control 5
MIMO Control
Multiple Inputs
Multiple Outputs
Lecture 4 : Instrumentation and Process control
6
Lecture 4 : Instrumentation and Process control 7
Lecture 4 : Instrumentation and Process control 8
A top-down view of process control
Maximization of profit subject to all
Overall goal in process plants safety and regulatory standards
Consider a chemical process
Overall profit function = Income from selling all the products produced – cost of (raw material + utilities)
Usually,
Profit = 𝑝 𝑥1, 𝑥2,….., 𝑥𝑛 where 𝑥1, 𝑥2,….., 𝑥𝑛 are process variables
Profit function 𝑝 is notionally optimized through two approaches:
Heuristic knowledge or Mathematical approach
(Process insights, studying the operations of (First-principles knowledge of the cause-effect behaviours of the
various units in the process and a precise statement of the
similar systems )
objective function )
Lecture 4 : Instrumentation and Process control 9
Lecture 4 : Instrumentation and Process control 10
Selection of control variables
𝑝∗ is the optimized profit achieved at optimum values for 𝑥1, 𝑥2, ……,𝑥𝑛 at 𝑥1∗ , 𝑥2∗ , … . … 𝑥𝑛∗
Optimize 𝒙∗𝟏 , 𝒙∗𝟐 , … . … 𝒙∗𝒏 c1 (c1*), … cm(cm*)
Process Control variable
with profit
x1’ x2 ,…. , xn selection
objective
Selection of manipulated variables
▪ Should have strong impact on the controlled variables
▪ There should be enough of them
▪ Economics
Lecture 4 : Instrumentation and Process control 11
Controllers in process industry
The controller chooses the values that the manipulated variables need to attain given the setpoints (obtained from optimization of the control
objectives) that the controlled variables should follow
c1 m1
Controlled Variables cr mr Manipulated Variables
Controller
cn mn
• If the controller is designed to manipulate all the variables (𝑚1 , 𝑚2 , … , 𝑚𝑟 , … 𝑚𝑚 ) at the same time to keep all the controlled variables at
their respective setpoints, then such a controller is a multivariable controller.
• Multivariable controllers are often broken down to simple single variable controllers which manipulate
only one variable at a time to control one output variable.
• Single-input Single-output controllers (SISO) are the most prevalent in the industry:
➢ The analysis and design of multivariable controllers is complex
➢ High reliability of SISO controllers
➢ Simple advanced concepts (Cascade, Ratio, Feed-forward) can be easily incorporated
into the SISO controller framework
Lecture 4 : Instrumentation and Process control 12
Multivariable control problem to a group of SISO control problems – How ?
Perform “Pairing” of manipulated and controlled variables through a series of heuristics (such as closeness of
C.V to M.V to minimize delay in response, sensitivity of C.V to changes in M.V and so on), prior knowledge and
sound mathematical principles
Consider a simple 2 x 2 case to illuminate how the different blocks in a SISO control loop flow out of a multivariable control problem
Disturbance d1
m1 Process
m2
SISO Decomposition
c1
Controller
c2
Lecture 4 : Instrumentation and Process control 13
Loop 1 is isolated where changes in 𝑐1 due to changes
in 𝑚2 are also viewed as a disturbance.
m2 d1
Loop1 Loop1
Disturbance (m2) Disturbance (d1)
d
Loop1 Set point for c1 Disturbance
process model
Loop1 Set point for c1
Controller process
Controller
The interactions between the loops are taken into
account by modeling them as disturbances
Lecture 4 : Instrumentation and Process control 14
Chemical process Example Simple Liquid Level System
Fi (m3/s)
Raw materials Products
Processing units Fi = Inlet flow rate
FO = Outlet flow rate
h (m) A = Area of cross-section of tank
Fo (m3/s)
(Through chemical, physical, h = Liquid level in tank
mechanical or thermal
changes) A (m2)
Process variable : Physical or chemical quantities which indicate the current process conditions of a typical processing unit
State variable : Minimum set of variables required to completely describe the system and are inherent to the system
Input variables: Exogenous to the system and capable of inducing changes in the state of the system through some dynamical process
Measurements or Output variables : Probes used to understand the internal conditions of a system and could represent state variables that are
measured or a combination of state variables measured
Lecture 4 : Instrumentation and Process control 15
Control system terminologies
Control systems are used to maintain controlled variables at their desired values by manipulating certain process variables, called
manipulated variables
Controlled variable – Variables that need to be maintained at desired values
Setpoint – The desired value of the controlled variable
Manipulated variable – Input variables that can be manipulated
Disturbance variable – Input or exogenous variables over which no control can be exercised
Error signal – Difference between the the setpoint and controlled variable value
1. Effect of disturbances 2. Change in set-point
Causes output variables to deviate
from their setpoints
Regulatory
Servo Control
Control
