Control System Architectures
Uploaded by
ScribdTranslationsControl System Architectures
Uploaded by
ScribdTranslationsIndustrial Communication Systems
Unit 2. Integration, Control, Communication and Computing
Control System Architectures
Introduction
A control system can be defined as a set of devices responsible for managing, ordering, directing or
regulating the behavior of another system, in order to reduce the probability of failure and obtain
the desired results.
Almost all processes that are intended to be controlled are multivariable in nature or also called
multiple input and multiple output systems, MIMO (multiple input, multiple output), which act on
two or more input variables of the process, and are designed taking into account the multivariable
nature of the process to be controlled.
In the design of multivariable control systems, an architectural style is applied, which defines the
system in terms of structural, control, communication patterns, etc. An architectural style defines
(at a high level) the architecture of a system, that is, its conceptual structure and logical
organization.
Control System Architectures
The IEEE defines system architecture as the fundamental organization of a system, including its
components, their relationships with each other and the environment, and the principles
governing their design and evolution.
A control system architecture is nothing more than the combination of sensors, actuators,
processors, software, etc. within the same structure, which allows control to be carried out in a
plant or process.
It is a detailed graphical representation of the control, communications and supervision equipment
of a plant, subprocess or machine, including communication networks and buses.
The evolution of control systems has generated architectural models where the components are
specific to the tasks to be performed and little by little these components tend to be independent
elements from each other, thereby achieving the isolation of the technology: objects or
components.
Different alternatives for Control System Architectures
An architectural style describes:
• A set of components with their responsibilities.
Control System Architectures Prof. Julima Anato 1/5
Industrial Communication Systems
Unit 2. Integration, Control, Communication and Computing
• A set of connectors or interfaces between components (communication, coordination,
cooperation, etc.).
• Constraints that define how components are integrated to form the system.
• Models that allow us to understand the properties of a general system based on the
known properties of the parts that comprise it.
There are different alternatives or architectures for the design of control systems under an
architectural style. Some control system architectures include:
• Classic Control Architecture.
• Centralized Control Systems Architecture
• Distributed Control Systems Architecture
• Multilevel Control Systems Architecture
Classical Control Systems Architecture
This architecture is based on the concept of separate control loops. This consists of the
decomposition of the system into independent loops, each associated with a single state variable
and a single control variable. Figure 1 shows a basic diagram of this type of architecture. And Figure
2 shows the architecture of continuous and discrete control systems.
Figure 1. Architecture of classical control systems
The distributed nature of the system contributes to increased reliability, so some local failures can
be safely handled by other controllers. Assigning a handler to each thread makes system
maintenance easier. However, the concept of separate control loops is effectively applied only
when this separation can also be done physically, which is considered a disadvantage as it is
impossible to do an integrated optimization of the system.
Control System Architectures Prof. Julima Anato 2/5
Industrial Communication Systems
Unit 2. Integration, Control, Communication and Computing
(to) (b)
Figure 2. Architecture of classical control systems in: (a) continuous system, (b)
discrete system
Centralized Control Systems Architecture
The centralized control system architecture (Figure 3) consists of a series of elements connected to
a central processor that monitors all process state variables and adjusts all control variables. This
type of architecture uses a shared communication medium to transport information between the
main control device and the controlled process.
Figure 3. Centralized control system architecture
Optimization, which results from process integration, and the flexibility of human-machine
Control System Architectures Prof. Julima Anato 3/5
Industrial Communication Systems
Unit 2. Integration, Control, Communication and Computing
communication, represents a great advantage in this architecture. However, the centralization
itself makes them less reliable and, in addition, the complexity of the software is not of much help.
Distributed Control Systems Architecture and its structure.
It is based on an architecture that uses high-speed industrial local networks for plant control. In
this case, each of the process controllers is connected to a bus (see figure 4) that allows connection
Local
Industrial
Network
Figure 4. Distributed control systems architecture
with the rest of the system components.
This architecture is characterized by its flexibility, any type of logical structure can be implemented
without changing the physical structure, including fault-tolerant hierarchical systems; secure high-
speed communications, integration with other control systems, access to shared resources and
greater autonomy.
Multilevel Control Systems Architecture and its structure
It is based on an architecture that divides control tasks into levels according to their hierarchy
within the process (supervision, control, etc.), as shown in Figure 5.
Control System Architectures Prof. Julima Anato 4/5
Industrial Communication Systems
Unit 2. Integration, Control, Communication and Computing
Figure 5. Multi-level control system architecture
This architecture combines centralized and distributed systems, which is considered an advantage,
as well as its independence and optimization derived from globality. However, it has the
disadvantage of showing a rigid HW/logical structures relationship and the difficulty in
communication between lower levels.
