Computer-process interface for data acquisition and

control

• In direct computer control individual controllers are replaced

by digital computer.
• The feedback, P, PI, PID control functions will be performed
by an executable program in BASIC, FORTRAN, assembly
language, etc.
• The control programs receive the input data from measured
output and estimate the output result as manipulated
variables to control the process.
• For a digital computer both input data and output results
are in digital form and correspond to discrete time value.
 Samplers
• Flow rates, pressures, liquid levels, temperatures, etc, are
provided continuously by the various sensors and
transducers.
• The computer can handle information on a discrete time
basis.
• Sampler is a switch which closes at specific time intervals
within which the continuous signal is converted into
sampled values. Each sampled value is used by the computer
control algorithm to give output result within the specific
time interval of sampling.
• If during this period (reading value, calculating error and
making a control action) measured value changed, it is not
recognized by the computer.

Cont.
 Hold element
• Most of the final control elements (pneumatic valves
particularly) are actuated by continuous time signals (e.g.,
compressed air).
• The discrete time signal from computer output (computed
by program) is converted into continuous signal by hold
element.
 Analog-to-digital converter
• A/D converters convert an analog voltage signal lying within
a specific range (such as ± 10 V, 0 to 5 V, etc.) to an integer
number.
• The range of the integer I is given by Irange = 2N, where N is
the number of bits.
• For a 12 bit converter, maximum integer is 212 = 4096.
• So the range is 2048 ≤I ≤+ 2047, I=0 exists between +ve and
–ve integers.
• For the positive range (2047), the resolution of the converter
is the voltage range divided by the intervals between
integers, Resolution = (10/2047) V = 0.00489 V.
• This gives an expected error of ± 1/2 of the resolution, or±
0.00244 V, Relative error = 0·00244 V/│measured voltage│
• The throughput speed of A/D converters is routinely 50,000 -
100,000 conversions/s.
Cont.
Ex: As an illustration of A/D conversion, suppose the 12-bit
converter measure voltages and reports the integer 1261
and 21 as a result of the conversion. Determine the actual
measured voltage for I=1261 and 21. Assume expected error
is ± 1/2 of the resolution.
Ans: For I=1261, Vmeasured = (10 ×1261/2047) = 6.16023 V
Relative error = ± 0·00244/6.16023 = ± 0.000397 =± 0.0397%
 Digital-to-analog converter
• D/A converters are the reverse of A/D conversion. An integer
with range 2N is converted to an analog voltage output.
• Ex: For output of 3.5 V determine the input integer for a D/A
converter.
• Iinput =2047× Voutput/10 = (2047) (3.5) /10=716.45 ≈716
• It is rounded to Iinput = 716. The actual output voltage of V
output = 716 X 10 /2047 = 3.49780 V≈3.5 V
 Multiplexer
• Multiplexers multiplex several similar measurements from a
remote location and sequentially transmit these over only a
few lines.
• Consider 10 thermocouple measurements from a process,
each with a voltage signal of – 10 mV. First of all, one may
not wish to run 10 lines and use 10 channels of our A/D
converters just for these signals, so multiplexing is
necessary.
• This is an electronic switch with several ports, which can
serve sequentially several lines carrying analog signals.

A/D converter with

multiplexer
 Digital I/O
• A digital computer is digital I/O. It is used in a control system for
1. Information concerning
• Status of relays turning pumps, valves, lights, and other devices on
or off.
• Status of multiplexers
• Settings of various switches
• Status of communication peripherals and various digital logic
devices
2. Control commands to
• Relays, switches, solenoid, digital logic devices
• Stepping motors
3. Communication between
• Several computers
• A computer and its peripherals, etc.
Cont.
• The digital signals are fully compatible with a computer.
• The transmission (input/output) of digital signals by I/O
interface can either in parallel (two way, in and out,
simultaneous transmission) or in series (one way, in or out
transmission).
• The transmission rates vary from very low to very high, and
are expressed in terms of baud rates
• Baud rate=10 × (number of character transmitted/sec)
Computer control loops (multiple)
 Modes of Computer Control
• Computer control is usually carried out in one of two modes:
supervisory control or direct digital control (DDC).

It involves resetting the

set-point of the local
controller according to
some computer
algorithm.

Cont.
It requires that all the controller action be carried out by the
digital computer. Measurements are sent to the computer
and compared with the set point; then the computed
control action is transmitted to the actuator.
 Programmable logic controller (PLC)
• Modern industrial control systems are microprocessor-based
programmable systems containing hardware and software.
• These are direct digital control, distributed control,
programmable control, and PID action.
• The processor in a PLC system has software that is easily
programmable and flexible, making the initial program,
updates, modifies.
• PLC are configured to receive a certain numbers of
inputs(both analog and digital) and to control a certain
number of outputs like actuators, displays or other types of
devices.
• PLCs are categorized into low-end, midrange, and high- end.;
Low-end is from 64 expandable up to 256 I/Os, midrange is
expandable up to 2,048 I/Os, and high-end is expandable up
to 8,192 I/Os.
Cont.
• PLCs have the ability to communicate with each other on a
local area network (LAN) or a wide area network (WAN),-
these send operational data to, and be controlled from, a
central computer terminal.
• The individual control loops are not independent in a process
but are interrelated.
• Measured variables may be monitored and manipulated
variables controlled simultaneously.
• Several processors also may be connected to a mainframe
computer for complex control functions.
Block diagram of programmable controller

Block diagram processing unit

• The central processing unit can be divided into the processor,
memory, and input and output units or modules.
• The units are interconnected by a two-way 16-bit data bus, a
one-way address bus, and a one-way enable bus.
• The enable and address buses are controlled by the processor,
which uses software instructions for its direction.
• When addressing an input module, the module is selected
with its enable bus code.
• The address bus then can be used to select which external
input data is to be put onto the two-way data bus.
• This data is then transferred to memory to wait for the next
step from the software instructions.
• The output modules are addressed and selected in the same
way as the input modules. The unit is selected by the enable
code, and the address bus directs the data placed on the data
bus from the memory by the processor to its output.
 Operation cycle in the PLC
• The operation cycle of a PLC is made up of two separate modes; these are
the I/O scan mode, followed by the execution mode.
• I/O scan mode is the period when the processor updates the output
control signals, based on the information received from the previous I/O
scan cycle after its evaluation of the signals. The processor then scans the
inputs in a serial mode and updates its internal memory as to the status of
the inputs.
• Execution mode follows the I/O scan mode. In this mode, the processor
evaluates the input data stored in memory against the data programmed
into the CPU. The processor then can determine the actions to be taken by
the output modules, and puts the data into memory for transfer to the
output modules during the next I/O scan mode.
• Scan time is the time required for the PLC to complete one I/O scan plus
the execution cycle. This time depends on the number of input and output
channels, the length of the ladder instruction sets, and the speed of the
processor. A typical scan time is between 5 and 20 ms.
Computer based digitally controlled process
 Distributed control system (DCS)
• DCS was originally considered as a replacement of huge panels, located in
central control rooms and comprising 100 to 1000 process instruments.
• It has expanded information-processing role, adding advanced control
such as, model reference control and expert systems; information-analysis
tools, such as statistical process control and intelligent alarming; decision
support applications such as predictive maintenance and document
management; and business system integration capabilities.
• The DCS has three essential qualities
1. first is to distribute its functions into relatively small sets of semiautonomous
subsystems, which are interconnected via a high-speed communications
network. These functions include data collection, process control, process
analysis and supervision, storage and retrieval of archived information, and
the presentation and reporting of information. The advantages are:
• lower exposure to component or subsystem failure and better isolation to
facilitate maintenance and upgrades.
• better partitioning of application requirements
• improved modularity for application development
• geographical distribution reduces installation costs (reduced wiring) and
provides more localized operational supervision
Cont.
2. 2nd to automate the manufacturing process by integrating advanced regulatory
control, logic and sequential control, it also includes information such as:
• activity-based cost accounting
• production scheduling and dispatching
• preventative or predictive maintenance scheduling
• validation of employee job certification and readiness
• information exchange with business, logistics, and transportation applications
3. 3rd a DCS organizes the command structure and information flow among its
constituent parts so as to have it act as a single automation system unifying the
various subsystems, including:
• process signal input and conditioning
• process actuator signal output
• regulatory, combinatorial, and sequence logic and procedural and supervisory
control
• human readable process displays of current values, alarms, trends, and
calculations
• human actions including setpoint changes, manual overrides and alarm
handling
• application subsystems such as process optimization and manufacturing
support
• information-storage subsystems
• communications subsystems
PROCESS/INDUSTRIAL INSTRUMENTS AND CONTROLS HANDBOOK by Gregory
K. McMillan and Douglas M. Considine.
 Figure: DCS presented by Siemens Energy & Automation, Inc.
Both DCS and PLC systems share the following components: Field devices; Input/output
modules; Controllers ; Human machine interface (HMI); Engineering; Supervisory control;
Business integration
 Comparison between PLC and DCS by Siemens Energy &
Automation, Inc.
PLC is applied DCS is applied
Factory automation - manufacturing or Process automation applications Involves
assembly of specific items “thing” the combination and/or transformation of
raw materials “stuff”
Product is visible as it moves through the Often impossible to visually see the
process product as it moves through the process
High-speed logic control (such as motors) Regulatory/Analog (loop) control
Simple Batch control Complex Batch Control
Value of the individual component being The value of a “batch” can be very high
manufactured is relatively low (either in raw material cost or market
value)
Downtime mainly results in lost production Downtime can result in process equipment
damage (product hardens, etc.)

Return to steady state production after an Return to steady state production after an
outage is short and relatively unplanned outage can be long, expensive,
straightforward and difficult
Typically, the heart of the system is the Typically, the heart of the system is the
controller HMI
PLC is applied DCS is applied
The operator's primary role is to handle The operator’s interaction is typically
exceptions required to keep the process in its target
performance range
Fast logic scan (approx. 10ms) is required to Control loops require deterministic scan
perform motor or motion control execution at a speed of 100 to 500 ms

Redundancy may not be cost justified System redundancy is often required

System can be taken offline to make Online configuration often required

configuration changes changes

Analog Control: Simple PID only Analog Control: Simple to advanced PID
control up to Advanced Process Control

Diagnostics to tell you when something is Asset Management alerts you to what
broken might break before it does

High level programming languages are Custom logic created from existing function
available for creating custom logic blocks

Solution is generic in nature, to be applied Use of pre-defined, pre-tested functions

on a wide variety of applications saves time
 Data processing for process plants
• In a process plant thousands of variables like flow rates, temperatures, pressures,
levels, compositions, etc. are routinely measured and automatically recorded for
process control, online optimization or process economic evaluation.
• In recent days, computers and data acquisition systems collect and process large
volume of data sampled with low to moderate frequency in the order of minutes
or seconds.
• Modern days computers not only access large volume of data at higher frequency
but also eliminate errors in data to increase its accuracy and validity.
• During process measurement and data transmission the data involve errors.
Measurement errors are of two types, i) Random error, ii) Gross error.
• Outcome of random errors are uncertain with respect to its sign (+ve/-ve) and
magnitude. If measurement is repeated with same process condition and
instrument, dissimilar values are obtained from the measurement. The random
errors are characterized by probability distribution function like Gaussian
distribution. The error is caused by power supply fluctuations, network
transmission and signal conversion noise, analog input filtering, changes in
ambient conditions, etc. The errors are beyond the control of process engineer
and always present in a measured data and it is characterized by high frequency
and low magnitude unlike some occasional spike.

Cont.
• The gross error corresponds to non random events and is caused by instrument
malfunctioning (due to improper installation of measuring devices), mal
calibration, wear or corrosion of sensors, solid deposits, etc. If measurement is
repeated with same process condition and instrument, similar values of gross
errors are obtained from the measurement. Good installation and maintenance
procedures prevent gross error to involve. Wear and fouling of sensors cause a
gross errors which increases slowly over long period of time and can be eliminated
by proper cleaning. The gross errors maintain a trend and are of larger
magnitude than random error.
• Small error in measured signal leads to deterioration of control performance,
uneconomic process and unsafe operation regimes.
• Different data processing technique have been used to eliminate errors.
• Signal conditioning by analog and digital filters. These are used to reduce the
effect of high frequency noise in the measurement. Today smart sensors performs
diagnostic check about the hardware problem and acceptability of the data.
• Statistical quality control test (SQC) detects significant error in process data. It is
used separately for each process variables and does not use knowledge of process
model; hence does not ensure consistency of the data with respect to
interrelationship among process variables. Obvious gross error is eliminated and
first step for eliminating random error.
• Data reconciliation is a technique to improve accuracy of measurement by
reducing random error. It explicitly makes use of process model constrains to
estimate process variables by adjusting process measurements in such a way to
satisfy the process constrains.
Cont.
• The technique regulates process measurements with random errors by making
them satisfy material and energy balance constraints to improve measurement
from a process via DCS or data collection devices.
• It involves steady state data reconciliation (Kuehn & Davidson, 1961) and is of two
types- linear and nonlinear data reconciliation that utilizes steady state matrix
and Jacobian matrix respectively from system governing mass, species and energy
balances.
• A reconciliation problem can be expressed as (Özyurt & Pike, 2004)
( ,, , )
• ∑ such that = 0 and ≤ ≤
,
• = [ , , , , , ,…, , ,…., , ] is the set of system variables are to be
measured. These variables are estimated using governing equation with
constraints.
• = [ , , , , , ,…, , ,…., , ] is the set of measurement vector.
• Where, i=1,2,…,n is the number of variables and j=1,2,….,m is the number of sets
of measurement, and , is the variance of errors over the measurement sets.
( ,, , )
• Standard error is estimated using different probability distribution
,
functions of errors.
Fig: Typical arrangement between the DCS and the Data Reconciliation,
Simulation, and Optimization procedures (from Simulation Sciences, Inc.,
1989) presented by Romagnoli & Sanchez, DATA PROCESSING AND
RECONCILIATION FOR CH EM ICAL PROCESS OPERATIONS
 The following variables are
monitored through the DCS:
• Temperatures at trays 12, 11, 9,
7, 5, 3, 1
• Temperatures of feed, distillate,
bottoms, and water in and out of
the condenser.
• Flowrates of steam to the
reboiler, water to the condenser,
feed, distillate, bottoms, and
reflux
• Pressure at the bottom of the
column
• Liquid levels in the condenser
and the bottom of the column

• The 23-cm-diameter distillation column under study is used to separate ethanol and water. It
contains 12 sieve trays with a 30-cm spacing (Fig.) as well as three possible feed locations, an
external reboiler, and two condensers, which are used at the bottom and the top of the column.
The second condenser is also used as a reflux drum; a pump sends the reflux back to the
column (tray 1) and the product to the product tank. Presented by Romagnoli & Sanchez, DATA
PROCESSING AND RECONCILIATION FOR CH EM ICAL PROCESS OPERATIONS.
The setup of the column, DCS, and interface with VAX-station and supervisory control
Presented by Romagnoli & Sanchez, DATA PROCESSING AND RECONCILIATION FOR CH EM
ICAL PROCESS OPERATIONS.
Presented by Romagnoli & Sanchez, DATA PROCESSING AND RECONCILIATION FOR CH
EM ICAL PROCESS OPERATIONS.
Presented by Romagnoli & Sanchez, DATA PROCESSING AND RECONCILIATION
FOR CHEMICAL PROCESS OPERATIONS.
Reference
• Stephanopoulos, G., "Chemical Process Control and Introduction to theory
and practice", Prentice Hall.
• William C. Dunn, Introduction to Instrumentation, Sensors, and Process
Control, ARTECH HOUSE, Norwood, MA, 2006.
Thank You