12. Automation Systems vs.
Control Systems
It is important at this stage to understand some of the differences in
the senses that these two terms are generally interpreted in technical
contexts and specifically in this course. These are given below.
1. Control Systems: the main function of control systems is to
ensure that outputs follow the set points.
2. Automation Systems: may have much more functionality,
such as computing set points for control systems, monitoring
system performance, plant startup or shutdown, job and
equipment scheduling, ….etc. Automation Systems are
essential for most modern industries.
Automation Systems may include Control Systems but the reverse
is not true. Control Systems may be parts of Automation Systems.
13. Industrial Automation vs.
Industrial Information Technology
Industrial Automation makes extensive use of Information Technology. Some of the major IT
areas that are used in the context of Industrial Automation are:
1. Control and Signal Processing
2. Simulation, Design, Analysis, Optimization
3. Communication and Networking
4. Real-time Computing
5. Database
However, Industrial Automation is distinct from IT in the following senses:
Industrial Automation also involves significant amount of hardware technologies, related to
Instrumentation and Sensing, Actuation and Drives, Electronics for Signal Conditioning,
Communication and Display, Embedded as well as Stand-alone Computing Systems etc.
As Industrial Automation systems grow more sophisticated in terms of the knowledge and
algorithms they use, as they encompass larger areas of operation comprising several units or
the whole of a factory, or even several of them, and as they integrate manufacturing with other
areas of business, such as, sales and customer care, finance and the entire supply chain of the
business, the usage of IT increases dramatically. However, the lower level Automation
Systems that only deal with individual or , at best, a group of machines, make less use of IT
and more of hardware, electronics and embedded computing.
14. Features of IT
There are some other distinguishing features of IT for the factory that differentiate it
with its more ubiquitous counterparts that are used in offices and other business.
Industrial information systems are generally
I. Reactive in the sense that they receive stimuli and in turn produce responses.
Naturally, a crucial component of an industrial information system is its interface.
II. Have to be real-time, by that we mean that the computation not only has to be
correct, but also must be produced in time. An accurate result, which is not timely
may be less preferable than a less accurate result produced in time. Therefore
systems have to be designed with clear considerations of meeting computing time
deadlines.
III. Considered mission-critical, in the sense that the malfunctioning can bring about
catastrophic consequences in terms of loss of human life or property. Therefore
extraordinary care must be exercised during their design to make them flawless. In
spite of that, elaborate mechanisms are often deployed to ensure that any unforeseen
circumstances can also be handled in a predictable manner. Fault-tolerance to
emergencies due to hardware and software faults must often be built in.
18. 7/65
Definitions: Variable and
Parameters
Variables - outputs of the process
Parameters - inputs to the process
Continuous variables and parameters - they are uninterrupted as time proceeds (e.g. flow
rate, force , temperature, pressure & velocity) are continuous over time during the process.
Also considered to be analog - can take on any of an infinite number of possible values
within a certain practical range.
They are not restricted to a discrete set of values
Discrete variables and parameters - can take on only certain values within a given range.
The most common types of discrete variable and parameters are:
1) Binary, i.e., ON/OFF, open/closed, and so on, i.e., limit switch open/closed, motor on/off,
work part present/not present.
2) Discrete, are variables that can take on more than two possible values but less than an infinite
number. Examples include daily piece counts in a production operation and the display of a
digital tachometer.
3) Pulse data, which consist of a series of pulses (called a pulse train). As a process variable, it
might be used to indicate piece counts, i.e.; parts passing on a conveyor activate a photocell to
produce a pulse for each part detected. As a process parameter, a pulse train might be used to
drive a stepper motor.
21. Continuous Vs. Discrete
Control
In reality, most operations in the process and discrete manufacturing industries include both
continuous and discrete variables and parameters. Consequently, many industrial controllers are
designed with the capability to receive, operate on, and transmit both types of signals and data.
Hence, in digital computer process control, even continuous variables and parameters possess
characteristics of discrete data, and these characteristics must be considered (why?) in the
design of the computer–process interface and the control algorithms used by the controller.
10/63
29. 17/63
Steady-State Optimization
Class of optimization techniques in which the process exhibits the
following characteristics:
1. Well-defined index of performance (IP) such as product cost,
production rate, or process yield;
2. Known relationship between process variables and IP
3. System parameter values that optimize IP can be determined
mathematically
When these characteristics apply, the control algorithm is designed to
make adjustments in the process parameters to drive the process toward
the optimal state.
Open-loop system
Several mathematical techniques are available for solving steady-state
optimal control problems, including differential calculus, calculus of
variations, and various mathematical programming methods.
31. Adaptive Control
A type of control systems in which the system parameters are automatically
adjusted to keep the system at an optimum level are called adaptive control
systems. Such type of control systems itself detects changes in the plant
parameters and make essential adjustments in the controller parameters to
maintain optimum level or performance.
Steady-state optimal control operates as an open-loop system. It works
successfully when there are no disturbances that invalidate the known
relationship between process parameters and process performance. Because
steady-state optimization is open-loop, it cannot compensate for disturbances.
When such disturbances are present in the application, a self-correcting form of
optimal control can be used, called adaptive control. Adaptive control is a self-
correcting form of optimal control that includes feedback control.
Measures the relevant process variables during operation (as in feedback
control)
Uses a control algorithm that attempts to optimize some index of
performance (optimal control)
33. Adaptive Control Operates in a
Time-Varying Environment
Adaptive control is distinguished from feedback control and steady-
state optimal control by its unique capability to cope with a time-
varying environment.
The environment changes over time and the changes have a potential
effect on system performance
If the control algorithm is fixed, the system may perform quite
differently in one environment than in another.
An adaptive control system is designed to compensate for its changing
environment by altering some aspect of its control algorithm to achieve
optimal performance.
In a production process, the “time-varying environment” consists of the
variations in processing variables, raw materials, tooling, atmospheric
conditions, and the like, any of which may affect performance.
34. Three Functions in AC
1. Identification function – current value of IP is determined
based on measurements of process variables
2. Decision function – decide what changes should be made to
improve system performance
Change one or more input parameters
Alter some internal function of the controller
3. Modification function – implement the decision function
Concerned with physical changes (hardware rather than
software)
In modification, the system parameters or process inputs are
altered using available actuators to drive the system toward
a more optimal state.
37. On-Line Search Strategies
Special class of adaptive control in which the decision function cannot be
sufficiently defined
Relationship between input parameters and IP is not known, or not known
well enough to implement the previous form of adaptive control
Instead, experiments are performed on the process
Small systematic changes are made in input parameters to observe effects
Based on observed effects, larger changes are made to drive the system
toward optimal performance.
On-line search strategies include a variety of schemes to explore the effects of
changes in process parameters, ranging from trial-and-error techniques to
gradient methods.
All of the schemes attempt to determine which input parameters cause the
greatest positive effect on the index of performance and then move the
process in that direction. There is little evidence that on-line search techniques
are used much in discrete parts manufacturing.
42. Discrete Control Systems:
Logic Control & Sequence Control
The two types of change correspond to two different types of
discrete control are:
1. Logic Control – is used to control the execution of event-driven
changes.
Output at any moment depends on the values of the inputs
Parameters and variables = 0 or 1 (OFF or ON)
2. Sequential Control – is used to manage time-driven changes. It
uses internal timing devices to determine when to initiate changes
in output variables.
Example: in the operation of transfer lines and automated
assembly machines, sequence control is used to coordinate the
various actions of the production system (e.g., transfer of parts,
changing of the tool, feeding of the metal cutting tool, etc.).
43. Discrete Control Systems
There are many industrial actuators which have set of command
inputs. The control inputs to these devices only belong to a specific
discrete set. For example in the control of a conveyor system, analog
motor control is not applied. Simple on-off control is adequate.
Therefore for this application, the motor-starter actuation system may
be considered as discrete having three modes, namely, start, stop and
run. Other examples of such actuators are solenoid valves, discussed in
a subsequent lesson.
Similarly, there are many industrial sensors (such as, Limit Switch /
Pressure Switch/ Photo Switch etc.) which provide discrete outputs
which may be interpreted as the presence/absence of an object in close
proximity, passing of parts on a conveyor, or a given pressure value
being higher or lower than a set value. These sensors thus indicate, not
the value of a process variable, but the particular range of values to
which the process variable belongs.
50. Two Basic Requirements for
Real-Time Process Control
1. Process-initiated interrupts
Controller must respond to incoming signals from the process (event-driven
changes). Depending on relative priority, controller may have to interrupt
current program to respond
2. Timer-initiated actions
Controller must be able to execute certain actions at specified points in time (time-
driven changes). Timer-initiated actions can be generated at regular time intervals,
ranging from very low values (e.g., 100 s to several minutes), or they can be
generated at distinct points in time.
Examples: (1) scanning sensor values from the process at regular sampling
intervals, (2) turning on and off switches, motors, and other binary devices
associated with the process at discrete points in time during the work cycle, (3)
displaying performance data on the operator’s console at regular times during a
production run, (4) re-computing optimal parameter values at specified times.
These two requirements correspond to the two types of changes mentioned
previously in the context of discrete control systems: (1) event-driven changes
and (2) time-driven changes.
53. 1. Polling (Data Sampling)
Periodic sampling of data to indicate status of process. In some systems, the polling
procedure simply requests whether any changes have occurred in the data since the last
polling cycle and then collects only the new data from the process. This tends to shorten the
cycle time required for polling. Issues related to polling include issues:
1. Polling frequency or rate – reciprocal of time interval between data samples
2. Polling order – sequence in which data collection points are sampled
3. Polling format – which refers to the manner in which the sampling procedure is
designed. The alternatives in polling format include:
All sensors polled every cycle
Update only data that has changed this cycle
Using High-level and Low-level scanning,
1. High-level scanning: in which only certain key data are collected each polling
cycle (high-level scanning),
2. Low-level scanning: but if the data indicates some irregularity in the process, a
low-level scan is undertaken to collect more complete data to ascertain the
source of the irregularity.
64. Direct Digital Control (DDC)
DDC represents a transitory phase in the evolution of computer process control technology.
Form of computer process control in which certain components in a conventional analog
control system are replaced by the digital computer (The difference between direct
digital control and analog control can be seen by comparing Figures 5.8 and 5.9).
Components remaining in DDC: sensor, transducer, amplifier and actuator.
Components replaced in DDC: analog controller, recording and display
instruments, set-point dials, and comparator.
New components in the loop include the digital computer, analog-to-digital and
digital-to-analog converters (ADCs and DACs), and multiplexers
It has also motivated the use of distributed control systems, in which a network of
microcomputers is utilized to control a complex process consisting of multiple unit
operations and/or machines.
Applications: process industries
The regulation of the process is accomplished on a time-shared, sampled-data basis
rather than by the many individual analog components working in a dedicated continuous
manner.
With DDC, the computer calculates the desired values of the input parameters and set
points, and these values are applied through a direct link to the process.
69. Programmable Logic Controller (PLC)
The Programmable Logic Controller (PLC) is a microprocessor-based
controller that executes a program of instructions to implement logic,
sequencing, counting, and arithmetic functions to control industrial machines and
processes.
PLC used extensively for sequence control today in transfer lines, robotics,
process control, and many other automated systems.
Introduced around 1970 to replace electromechanical relay controllers in
discrete product manufacturing
Today’s PLCs perform both discrete and continuous control in both process
industries and discrete product industries
In essence, a PLC is a special purpose industrial microprocessor based real-time
computing system, which performs the following functions in the context of
industrial operations
1. Monitor Input/Sensors
2. Execute logic, sequencing, timing, counting functions for Control/Diagnostics
3. Drives Actuators/Indicators
70. Programmable Logic Controller (PLC)
Within a PLC technology, the terms programmable automation
controller (PAC) and remote terminal unit (RTU) have been
coined to distinguish among the types of control devices.
1) A PAC can be thought of as a digital controller that combines the
capabilities of a personal computer with those of a conventional
PLC; specifically, the input/output capabilities of a PLC are
combined with the data processing, network connectivity, and
enterprise data integration features of a PC.
2) A RTU is a microprocessor-based device that is connected to the
process, receiving electrical signals from sensors and converting
them into digital data for use by a central control computer; in
some cases it also performs a control function for local sections of
the process.
RTUs often use wireless communications to transmit data,
whereas PLCs use hardwired connections.
72. Supervisory Control
Supervisory control systems perform, typically the following functions:
Set point computation: Set points for important process variables are computed depending
on factors such as nature of the product, production volume, mode of processing. This
function has a lot of impact on production volume, energy and quality and efficiency.
Performance Monitoring / Diagnostics: Process variables are monitored to check for
possible system component failure, control loop detuning(readjusting) , actuator
saturation, process parameter change etc. The results are displayed and possibly archived
for subsequent analysis.
Start up / Shut down / Emergency Operations : Special discrete and continuous control
modes are initiated to carry out the intended operation, either in response to operator
commands or in response to diagnostic events such as detected failure modes.
Control Reconfiguration / Tuning: Structural or Parametric redesign of control loops are
carried out, either in response to operator commands or in response to diagnostic events
such as detected failure modes. Control reconfigurations may also be necessary to
accommodate variation of feedback or energy input e.g. gas fired to oil fired.
Operator Interface: Graphical interfaces for supervisory operators are provided, for
manual supervision and intervention.
73. Supervisory Control & Data
Acquisition (SCADA)
The term SCADA collect data from the process, which often includes multiple sites
distributed over large distances. SCADA system consists of :
(1) A central supervisory computer system capable of collecting data from the
process and transmitting command signals to the process,
(2) A human-machine interface (HMI) that presents the collected data to the
system operator(s) and enables them to send command signals,
(3) Distributed PLCs and RTUs that are connected directly to the process for data
acquisition and control, and
(4) A communications network that connects the central computer to the remote
PLCs and RTUs.
The general mode of operation in SCADA is for the remote devices to directly
control the various control loops in the system, but these devices can be overridden
by the operator at the HMI if that becomes necessary for some reason.
Example: the operator might change the value of a set point in one of the
control loops
74. Supervisory Control & Data
Acquisition (SCADA)
In some applications, Supervisory control is not much more than regulatory
control or feedforward control.
In other applications, the supervisory control system is designed to
implement optimal or adaptive control. It seeks to optimize some well-
defined objective function, which is usually based on economic criteria such
as yield, production rate, cost, quality, or other objectives that pertain to
process performance.
In the context of discrete manufacturing, SCADA is the control system that
directs and coordinates the activities of several interacting pieces of equipment in
a manufacturing cell or system, such as a group of machines interconnected by a
material handling system.
Again, the objectives of supervisory control are motivated by economic
considerations. The control objectives might include minimizing part or product
costs by determining optimum operating conditions, maximizing machine
utilization through efficient scheduling, or minimizing tooling costs by tracking
tool lives and scheduling tool changes
76. Distributed Control Systems (DCS)
Multiple microcomputers connected together to share and distribute the process
control workload. A DCS consists of the following components and features:
Multiple process control stations to control individual loops and devices of
the process. PCs, PACs, PLCs, and RTUs are used at these stations.
Central control room equipped with operator stations, where supervisory
control of the plant occurs.
Local operator stations distributed throughout the plant. This provides the
DCS with redundancy. If a control failure occurs in the central control
room, the local operator stations take over the central control functions. If
a local operator station fails, the other local operator stations assume the
functions of the failed station.
Communications network (data highway)
The distinction between DCS and SCADA is not always clear.
The term distributed system emphasizes an interconnected collection of computers,
whereas supervisory control emphasizes the use of a central computer to manage an
interconnected collection of remote controller and data acquisition devices
81. Enterprise-Wide
Integration of Factory Data
Managers have direct access to factory operations
Process Scheduling: depending on the sequence of operations to be carried on the
existing batches of products, processing resource availability for optimal resource
utilization. Planners have most current data on production times and rates for scheduling
purposes
Inventory Management: Decision processes related to monitoring of inventory status of
raw material, finished goods etc. and deployment of operations related to their
management
Sales personnel can provide realistic delivery dates to customers, based on current shop
loading
Order trackers can provide current status information to inquiring customers
Quality Management : assessment, documentation and management of quality. QC can
access quality issues from previous orders
Accounting has most recent production cost data
Production personnel can access product design data to clarify ambiguities
Maintenance Management: decision processes related to detection and deployment of
maintenance operations
The control hierarchy illustrated above starts from the top and works its way down.
For example, the plant-level aggregate production plan sets the overall boundaries of which products will be produced and when.
This provides a constraint on the shop floor level, which must then allocate the required production to machining cells and/or other production processes in the most effective manner.
Once a specific machining cell or set of production processes is allocated its production schedule for a specific day, it is the responsibility of the work cell/production line level to coordinate the manufacture of the product through the related machines and processes it requires.
Finally, when the machine is assigned its role in partially fabricating the product, it is the responsibility of the machine-level controller to execute the correct steps of the fabrication process.
Adaptive control (AC) machining originated out of research in the early 1960s and was sponsored by the US Air Force at the Bendix Research Laboratory for machining operation. The term ‘adaptive control’ denotes a control system that measures certain output process variables and uses this information to control speed and/or feed. Some of the process variables that have been used in adaptive control machining systems include spindle deflection, force, torque, cutting torque, vibration amplitude, and horse power consumed. The motivation for developing an adaptive machining system lies in trying to operate a process more efficiently.
Example 5.1 Single-Level versus Multilevel Interrupt Systems
Three interrupts representing tasks of three different priority levels arrive for service in the reverse order of their respective priorities. Task 1 with the lowest priority arrives first. Soon after, higher priority Task 2 arrives. And soon after that, highest priority Task 3 arrives. How would the computer-control system respond under (a) a single-level interrupt system and (b) a multilevel interrupt system?
Solution: The response of the system for the two interrupt systems is shown in Figure 5.6 . It response of the computer-control system in Example 5.1 to three priority interrupts for (a) a single-level interrupt system and (b) a multilevel interrupt system.
Task 3 is the highest level priority. Task 1 is the lowest level. Tasks arrive for servicing in the order 1, then 2, then 3. In (a), Task 3 must wait until Tasks 1 and 2 have been completed. In (b), Task 3 interrupts execution of Task 2, whose priority level allowed it to interrupt Task 1.