INSTRUMENTATION &

PROCESS CONTROL
PERRY TIPS AND TRICKS
 Chapter 8 – Process Control (7th Ed. and 8th Ed.)
 Frequently Used Laplace Transforms – Table 8-1 (7E, 8E)
 Advanced Control Methods – Page 8-16 (7E), Page 8-21 (8E)
 Unit Operations Control – Page 8-31 (7E), Page 8-39 (8E)
 Process Measurements – Page 8-43 (7E), Page 8-54 (8E)
 Controllers, FCEs, and Regulators – Page 8-61 (7E), Page 8-71 (8E)
 OUTLINE
I. BASICS OF PROCESS CONTROL
II. TRANSFER FUNCTION
III. DYNAMIC BEHAVIOR OF A SYSTEM
IV. FEEDBACK CONTROL
V. DEGREES OF FREEDOM
BASICS OF PROCESS
CONTROL
PROCESS CONTROL VARIABLES
 CONTROLLED VARIABLES
 Keyword: setpoint
 variables that quantify the quality of the final product
 MANIPULATED VARIABLES
 Keyword: flowrate
 variables that keep the controlled variables within setpoint
 usually involves flowrates
 DISTURBANCE VARIABLES
 variables that cause controlled variables to deviate from the setpoint
PROCESS CONTROL VARIABLES
TYPES OF CONTROL
FEED-FORWARD CONTROL
 takes into account the measurement of ALL disturbance variables in controlling
the manipulated variable
 disturbance variables must be accurately predicted; no unaccounted disturbances
TYPES OF CONTROL
FEEDBACK CONTROL
 measures the output error against the setpoint and adjusts the manipulated
variable based on the deviation
 widely used control system
CONTROL SYSTEM DESIGN
TRADITIONAL APPROACH
 based on experience, basic knowledge, insight

MODEL-BASED APPROACH
• uses model-based controller designs, control laws, and computer automation

Industrial Plant Process Control Considerations

1. Safety
2. Environmental Regulations
3. Product Specifications and Production Rate
4. Economic Plant Operation
5. Stable Plant Operation
CONTROL SYSTEM HARDWARE
SENSORS
• measures the value of the output or controlled variable

TRANSDUCERS
• converts signals from one form to another, usually into an electrical signal
• a sensor with a signal conditioning circuit

TRANSMISSION LINES/TRANSMITTERS
• signal carrier; transducers with 4-20 mA output
• least affected by electrical noise and resistance

CONTROLLER
• receives information from the sensor and decides the necessary action to do
TRANSFER
FUNCTIONS
WHAT IS A TRANSFER FUNCTION?
Transfer Function
• correlation in the s-domain of a dependent variable (output variable or “effect”)
and an independent variable (or input variable or “cause”)
• directly applicable to processes exhibiting linear dynamic behavior – only few
processes

𝑌(𝑠) ℒ{𝑦 𝑡 }
𝐺 𝑠 = =
𝑋(𝑠) ℒ{𝑥 𝑡 }
PROPERTIES OF A TRANSFER FUNCTION
Steady-State Gain
• steady-state change in the output against the steady-state change in the input
• constant for a linear transfer function; dependent on y and x for a non-linear
transfer function

𝑦2 − 𝑦1
𝐾= = lim 𝐺(𝑠)
𝑥2 − 𝑥1 𝑠→0
PROPERTIES OF A TRANSFER FUNCTION
Order of a Transfer Function
• The order of a transfer function is said to be the order of the denominator
polynomial, and is equal to the order of the ODE

𝑦2 − 𝑦1
𝐾= = lim 𝐺(𝑠)
𝑥2 − 𝑥1 𝑠→0
Physical Realizability
𝑚 𝑖
𝑖=0 𝑏𝑖 𝑠
• Given 𝐺 𝑠 = 𝑛 𝑎 𝑠𝑖 , n > m for a system to be physically realizable
𝑖=0 𝑖
PROPERTIES OF A TRANSFER FUNCTION
Additive Property
• the response involving multiple inputs in parallel can be written as:
𝑌 𝑠 = 𝑋1 𝑠 𝐺1 𝑠 + 𝑋2 𝑠 𝐺2 𝑠

Multiplicative Property
• the response involving multiple inputs in series can be written as:
𝑌1 𝑠 = 𝐺1 𝑠 𝑈 𝑠
𝑌2 𝑠 = 𝐺2 𝑠 𝐺1 𝑠 𝑈(𝑠)
DYNAMIC BEHAVIOR
OF A SYSTEM
STANDARD PROCESS INPUTS
Step Input
• the step change is sudden and occurs at a time t = 0
• if M = 1, it is called a “unit step change”, S(t)

Ramp Input
• also denoted as “drifting disturbances”
• rate of change is constant
STANDARD PROCESS INPUTS
Rectangular Pulse
• a brief step change that returns back to its original value
• tw is known as the pulse width

Sinusoidal Input
• subject to periodic or cyclic disturbances
• A is the amplitude and ω is the angular frequency
STANDARD PROCESS INPUTS
Impulse Input
• sudden spikes; short, transient disturbances

Random Input
• fluctuations; characterized using statistical terms
FIRST-ORDER SYSTEM
Standard Form of a First-order Transfer Function
𝐾
𝐺 𝑠 =
𝜏𝑠 + 1
where: K = process gain, τ = time constant

Response of a First Order System on a Step Input

SECOND-ORDER SYSTEM
Standard Form of a Second-order Transfer Function
𝐾
𝐺 𝑠 = 2 2
𝜏 𝑠 + 2ξ𝜏𝑠 + 1

where: K = process gain, τ = time constant

ξ = damping coefficient

Response of a First Order System on a Step Input

• Response has a higher degree of oscillation as the damping coefficient decreases
• Large damping coefficients yield slow responses
• Fastest response without overshoot is critically damped
SECOND-ORDER SYSTEM
Critically damped and Overdamped Underdamped
CHARACTERISTICS OF AN
UNDERDAMPED PROCESS
Rise Time, tr – time to first reach the new
steady-state value
Time to First Peak, tp – time to reach first
maximum value
Settling time, ts – time required to remain
within a band (±5% of the total change in y for
95% response time)
Overshoot, OS = a/b
Decay Ratio, DR = c/a
Period of Oscillation, P – time between two
successive peaks or valleys
POLES AND ZEROS
𝑏𝑚 (𝑠 − 𝑧1 )(𝑠 − 𝑧2 )(𝑠 − 𝑧3 )
𝐺 𝑠 =
𝑎𝑛 (𝑠 − 𝑝1 )(𝑠 − 𝑝2 )(𝑠 − 𝑝3 )

POLES ZEROS

complex pole unstable pole overshoot inverse response

integrating process
TIME DELAYS
• Time delay, represented by θ, occur in processes due to various phenomena:
1. fluid flow in a pipe
2. transfer of solid materials via conveyor belts
3. sampling line delay
4. on-line analyses
• Time delays are bad for control since they involve a delay in the information

Representation of a Time Delay

𝐺 𝑠 = 𝑒 −𝜃𝑠
FEEDBACK CONTROL
BASIC CONTROL MODES
• GOAL: reduce the signal error to zero

PROPORTIONAL CONTROL (P) 𝑃′ 𝑠

• controller output is proportional to the error signal 𝐺 𝑠 = = 𝐾𝑐
• bias can be reset – manual reset 𝐸(𝑠)
• controller gain can be adjusted to make controller output changes as sensitive as
desired
• sign of the controller gain can be chosen to proportionate controller output and error
signal
• steady-state error occurs after a set-point change or a sustained disturbance – needs
offset
BASIC CONTROL MODES
INTEGRAL CONTROL (I)
• controller output depends on the integral of the error signal over time
• changes as long as e(t*) ≠0
• τI is also known as ‘reset time’ or ‘integral time’
• eliminates offset
• disadvantageous due to reset windup or integral windup

PROPORTIONAL-INTEGRAL CONTROL (PI)

𝑃′ 𝑠 1
𝐺 𝑠 = = 𝐾𝑐 (1 + )
𝐸(𝑠) 𝜏𝐼 𝑠
BASIC CONTROL MODES
DERIVATIVE CONTROL (D)
• anticipate the future behavior of the error signal by considering its rate of change
• τD is also known as ‘derivative time’
• tends to stabilize the controlled process

PROPORTIONAL-INTEGRAL-DERIVATIVE CONTROL (PID)

PARALLEL FORM SERIES FORM

𝑃′ 𝑠 1 𝑃′ 𝑠 𝜏𝐼 𝑠 + 1 𝜏𝐷 𝑠 + 1
𝐺 𝑠 = = 𝐾𝑐 (1 + + 𝜏𝐷 𝑠) 𝐺 𝑠 = = 𝐾𝑐
𝐸(𝑠) 𝜏𝐼 𝑠 𝐸(𝑠) 𝜏𝐼 𝑠 𝛼𝜏𝐷 𝑠 + 1
COMPARING CONTROLLERS
EFFECTS ON CONTROLLER

P PID PI PI
DEGREES OF
FREEDOM
DEGREES OF FREEDOM
• refers to the number of variables that DEGREES OF FREEDOM ANALYSIS
must be specified to define the complete
process CASE REMARK NO. OF
SOLUTIONS
Exactly
𝑵𝑭 = 𝑵𝑽 − 𝑵𝑬 NF = 0
Specified
1

NF > 0 Underspecified Infinite

where: NF < 0 Overspecified No solution
NF = degrees of freedom
NV = no. of variables
NE = no. of equations
 ADDITIONAL
INFORMATION
OVERVIEW OF PROCESS CONTROL IN MANUFACTURING

Scheduling

Production and Target

Monitoring

SCADA (HMI)  DBMS

I/O modules, PLCs,

RTUs

Flow sensors,
Temperature sensors,
Control Valves
SCADA vs DCS
Supervisory Control and Data Acquisition Distributed Control System (DCS)
(SCADA) • monitoring and control system used in
• monitoring and control system used in industrial applications
industrial applications • collects data on a factory-wide scale
• collects data on a factory-wide scale • data collection is distributed throughout
• centralized real-time data collection the factory
from one or multiple sources • single incident will not bring down the
• used in control rooms system
• single incident may bring down the • more costly
system • for more geographically-dispersed
• less costly systems
• usually for small-scale industrial systems
PLC vs DCS
Programmable Logic Computer (PLC) Distributed Control System (DCS)
• needs an Human Machine Interface • distributed throughout a machine
(HMI) most of the time • provides instructions to different parts
• handles binary input and output of the • one DCS corresponds to one section or
logic statement part of a machine to control and manage
• can be referred to as a small DCS its operation
DISTURBANCE VS NOISE
DISTURBANCE NOISE
• unwanted input that affects the control • can come from external interferences or
system’s output electrical fields, instrument wear and
• can increase system errors in control tear, or any physical obstruction in the
path of the instrument
• could result in erroneous readings
HYSTERESIS AND DEAD BAND
HYSTERESIS DEAD BAND
• creates ability to regulate a variable to • a defined range of set point values
its set point value that is within a • no outputs to the system are made
specified dead band