Industrial Communication and System Integration

Introduction and Context

Modern automated systems are not collections of isolated machines. Sensors, actuators, programmable logic controllers (PLCs), variable-speed drives, robots, supervisory control and data acquisition (SCADA) systems, manufacturing execution systems (MES), and cloud platforms exchange information continuously. Industrial communication provides the physical and logical infrastructure that makes this exchange possible.

Industrial networks must satisfy requirements that differ from ordinary office networks. They may need deterministic timing, high availability, electromagnetic compatibility, simple maintenance, functional safety, cybersecurity, and long service life. A communication failure can stop production, damage equipment, or create a safety hazard. Consequently, network selection is an engineering decision involving performance, environment, cost, lifecycle, and risk.

This chapter develops a practical framework for designing industrial communication systems. It examines fieldbus technologies, Industrial Ethernet, Modbus, PROFIBUS and PROFINET, EtherNet/IP, CAN, IO-Link, OPC UA, wireless systems, gateways, and time-sensitive networking (TSN).

Learning Objectives

After completing this chapter, the student should be able to:

1. Explain the functions and architecture of industrial communication networks.
2. Distinguish fieldbus, Industrial Ethernet, device-level, controller-level, and information-level communication.
3. Compare Modbus, PROFIBUS/PROFINET, EtherNet/IP, CAN, IO-Link, OPC UA, and industrial wireless systems.
4. Select a communication technology using timing, distance, topology, interoperability, environmental, and cybersecurity criteria.
5. Explain the role of gateways, protocol conversion, edge computing, and OPC UA information models.
6. Analyze latency, bandwidth, availability, and determinism in an industrial network.
7. Design a basic integrated network for a manufacturing cell.
8. Apply professional practices for commissioning, diagnostics, documentation, and security.

1. Industrial Communication Architecture

An automated plant is commonly organized into levels, although modern architectures increasingly blur these boundaries.

A communication system has four interacting parts:

• Physical medium: copper cable, fiber optic cable, radio, or a backplane.
• Protocol: rules for addressing, framing, error detection, timing, and access to the medium.
• Data model: meaning and structure of values, such as temperature, status, or batch number.
• Engineering and management tools: configuration, diagnostics, firmware management, and cybersecurity controls.

The OSI reference model is useful for analysis. Industrial engineers commonly focus on the physical layer, data-link layer, network and transport functions, and application layer. A device can be electrically compatible yet still fail to interoperate if its addressing, data representation, configuration, or application semantics differ.

Cyclic and acyclic communication

Cyclic communication transfers data repeatedly at a configured interval. A drive may receive a speed reference every 10 ms and return actual speed and status in the same cycle. Cyclic communication is suitable for control loops.

Acyclic communication occurs when required, for example when a controller reads a device parameter or an operator requests diagnostic information. It uses bandwidth efficiently for infrequent data but is generally less predictable for control.

A simplified network data-rate relationship is:

where R is the payload data rate in bit/s, N is the number of messages per cycle, L is the payload length in bytes per message, and T is the cycle time in seconds. This is only a payload estimate. Protocol headers, acknowledgements, inter-frame gaps, retransmissions, and physical-layer encoding increase the actual required capacity. The estimate is valid when messages are periodic and their payload lengths are known.

For example, 20 devices each transmit 16 bytes every 20 ms:

A network should not be selected with only 128 kbit/s of nominal capacity, because overhead and traffic bursts must also be allowed for.

2. Fieldbus and Industrial Ethernet

Fieldbus principles

A fieldbus is a digital, multi-device communication system designed to replace large quantities of point-to-point wiring. One cable may carry process values, commands, diagnostics, and device parameters. Examples include PROFIBUS, CANopen, DeviceNet, Foundation Fieldbus, and Modbus RTU.

Advantages include reduced wiring, distributed diagnostics, modular installation, and standardized device behavior. Limitations may include lower bandwidth than modern Ethernet, specialized engineering tools, vendor-specific profiles, and difficulty integrating directly with enterprise systems.

Industrial Ethernet

Industrial Ethernet uses Ethernet technologies adapted for industrial conditions and automation requirements. Adaptations may include rugged connectors, shielded cables, ring redundancy, managed switches, real-time scheduling, synchronization, and environmental ratings.

Standard Ethernet is fundamentally a packet network. At higher layers, industrial protocols define how controllers and devices exchange data. Therefore, “Ethernet” alone does not identify interoperability. PROFINET, EtherNet/IP, Modbus TCP, and OPC UA use Ethernet-related infrastructure but provide different application models and real-time mechanisms.

Important topology choices are:

• Line: inexpensive and convenient, but a cable or device failure can disconnect downstream devices unless the devices support bypass or redundancy.
• Star: each device connects to a switch; faults are easier to isolate, but the switch is a potential single point of failure.
• Ring: can provide path redundancy when managed by an appropriate protocol; recovery time must be verified.
• Tree: combines switches and segments and is common in plants, but requires careful traffic and fault analysis.

3. Major Industrial Protocols

Modbus

Modbus is a simple request-response protocol widely used for PLCs, meters, drives, and building systems. Modbus RTU commonly operates over serial links such as RS-485, while Modbus TCP operates over Ethernet and TCP/IP.

A Modbus device exposes logical data areas such as coils, discrete inputs, input registers, and holding registers. The actual interpretation of a register—scaling, signedness, byte order, and engineering unit—must be defined by the device documentation.

Modbus strengths are simplicity, low cost, and broad availability. Its weaknesses include limited standardization of data meaning, polling overhead, and limited intrinsic security. Basic Modbus does not provide authentication or encryption; protection must therefore be implemented through segmentation, secure remote access, firewalls, or secure tunneling.

With RS-485, termination and biasing must follow the transceiver and network design requirements. Excessive stubs, incorrect polarity, poor shielding, and multiple termination resistors are common causes of intermittent faults.

PROFIBUS and PROFINET

PROFIBUS is a fieldbus family. PROFIBUS DP is widely associated with distributed I/O, drives, and deterministic cyclic communication over a serial physical layer. PROFIBUS PA is designed for process instrumentation and hazardous-area applications using suitable systems.

PROFINET is an Industrial Ethernet system. It supports standard TCP/IP traffic as well as real-time communication. PROFINET RT provides prioritized real-time exchange, while higher-performance applications may use isochronous real-time mechanisms. Device descriptions, commonly represented through GSD or GSDML files, support engineering and parameterization.

A PROFINET system generally includes an IO controller, such as a PLC, and IO devices, such as remote I/O or drives. Engineering must assign names, addresses, modules, and communication parameters. A device may be electrically connected yet remain unavailable if its configured device name or module structure is incorrect.

EtherNet/IP

EtherNet/IP uses standard Ethernet and the Common Industrial Protocol (CIP) at the application level. It supports explicit messaging for configuration and information access and implicit or I/O messaging for time-critical cyclic data. Producer-consumer communication can allow multiple devices to receive the same data without separate point-to-point requests.

Its major benefits include a rich object-oriented device model, extensive controller and device support, and integration with CIP Safety, CIP Motion, and other profiles. Engineering must account for connection limits, packet rates, switch behavior, multicast traffic, and controller resources.

CAN and CANopen

Controller Area Network (CAN) uses message identifiers to determine arbitration priority. A dominant bit can overwrite a recessive bit, allowing a node with the highest-priority identifier to gain access without destroying the frame. This is a major reason CAN is effective for embedded control.

CAN provides strong error detection, automatic retransmission, and fault confinement. CANopen adds standardized object dictionaries, network management, process data objects (PDOs), and service data objects (SDOs). CAN is common in mobile machinery, robots, drives, and embedded mechatronic systems.

A key limitation is that CAN bandwidth is shared and messages are prioritized rather than scheduled equally. Increasing bus length generally requires reducing bit rate. Exact limits depend on the physical layer, transceivers, cable, and timing configuration.

IO-Link

IO-Link is a point-to-point communication system for sensors and actuators. An IO-Link master connects individual devices over standardized three-wire connections. It provides process data, device identification, parameters, and diagnostics.

IO-Link is not a multidrop fieldbus. Its value lies in making conventional sensors and actuators identifiable and configurable. A replacement sensor can obtain parameters from the master, reducing commissioning time and configuration errors. The master then connects upward through a fieldbus or Industrial Ethernet protocol.

OPC UA

OPC Unified Architecture (OPC UA) is a platform-independent framework for secure, structured information exchange. Unlike a basic register protocol, OPC UA can represent objects, variables, methods, events, relationships, units, and metadata.

OPC UA supports client-server communication and, in appropriate deployments, publish-subscribe communication. Security mechanisms include application authentication, signing, encryption, certificates, and user authorization. Security is not automatic: certificates must be managed, trust relationships controlled, and insecure endpoints disabled.

OPC UA is especially useful between controllers, SCADA, MES, historians, and analytics systems. It can also be used near the control level, but its suitability for a specific hard real-time loop must be demonstrated rather than assumed.

4. Wireless Industrial Communication

Wireless systems are useful for mobile equipment, rotating machinery, temporary installations, difficult-to-wire areas, and condition monitoring. Technologies include industrial Wi-Fi, WirelessHART, ISA100.11a, private cellular systems, and low-power wide-area networks.

Wireless design must consider radio interference, channel planning, antenna placement, roaming, coexistence, metal structures, multipath propagation, battery life, and security. A link that works in an empty workshop may fail when machinery, materials, or vehicles change the propagation environment.

Wireless is often appropriate for monitoring and supervisory functions. Closed-loop motion or safety functions generally require a specifically engineered wireless technology with verified timing and safety behavior. Ordinary Wi-Fi should not be assumed to provide deterministic control.

5. Gateways and System Integration

A gateway connects networks that use different protocols, addressing schemes, or data models. Examples include Modbus RTU-to-Modbus TCP gateways, CANopen-to-PROFINET gateways, and OPC UA edge gateways.

Protocol conversion is not merely translation of bytes. A gateway must map:

• addresses and identifiers;
• data types and byte order;
• scaling and engineering units;
• update rates and quality states;
• alarms, timestamps, and diagnostics;
• command ownership and permissible control directions.

A gateway can introduce latency, buffering, loss of event information, and a single point of failure. Control-critical data should remain within a network designed for its timing requirements. Gateways are usually better suited to well-defined boundaries, such as exposing production data to an MES, than to concealing an unsuitable control network.

Data modeling and semantic integration

Integration is strongest when systems share meaning, not merely values. A temperature value should include its unit, sensor identity, timestamp, quality, and status. OPC UA information models, standardized device profiles, naming conventions, and asset models support this semantic layer.

A practical naming convention should identify plant, area, line, cell, equipment, signal, and unit. Consistent naming reduces commissioning effort and enables historians and analytics systems to interpret data correctly.

6. Determinism, Timing, and Availability

For a periodic control message, the total communication delay can be considered as:

The terms represent source processing, queueing, transmission, propagation, intermediate processing, and destination processing delays. In a real network, these values vary. The variation is called jitter. Control performance depends not only on average delay but also on the worst credible delay and the consequences of lost or late data.

Transmission delay for a frame can be estimated by:

where L is the frame length in bits and C is the link rate in bit/s. For a 1,200-byte frame on a 100 Mbit/s link, ignoring encoding and inter-frame overhead:

This is only the time to place the frame on the link. It is not the end-to-end latency.

Availability is the probability that a system is operational when required. For a repairable component with mean time between failures (MTBF) and mean time to repair (MTTR), a simple steady-state approximation is:

The approximation assumes stable failure and repair rates and does not fully represent planned downtime, common-cause failures, or redundant architectures. For example, MTBF = 10,000 h and MTTR = 10 h gives:

or approximately 99.90 percent. Redundancy can improve availability, but only if the redundant paths do not share the same power supply, cable route, switch, or configuration failure.

7. Time-Sensitive Networking

Time-Sensitive Networking is a family of IEEE Ethernet standards intended to provide bounded latency, low jitter, synchronization, and reliability over Ethernet. Relevant mechanisms include precise time synchronization, scheduled traffic, traffic shaping, and frame replication or elimination.

TSN is important because it can allow conventional Ethernet infrastructure to carry information technology traffic and highly time-sensitive industrial traffic under controlled conditions. However, TSN is not one protocol and does not automatically make every Ethernet network deterministic. A successful design requires compatible devices, switches, profiles, time synchronization, traffic schedules, and validated configuration.

TSN can support convergence between motion control, machine vision, safety-related communication, and plant information systems. Engineers must still separate functions when risk, certification, or maintenance requirements make convergence inappropriate.

8. Network Design Method

A disciplined design process is more valuable than choosing a fashionable protocol.

Step 1: Define requirements

Record device count, message types, cycle times, payload sizes, permitted latency and jitter, distance, environmental conditions, safety classification, availability target, maintenance practices, and future expansion.

Step 2: Classify traffic

Separate hard or soft real-time control, alarms, diagnostics, configuration, video, and enterprise data. Do not place large noncritical transfers on the same path as time-critical control without traffic analysis.

Step 3: Select the physical medium and topology

Consider distance, electromagnetic noise, grounding, hazardous areas, vibration, temperature, connector protection, and service access. Fiber may be preferable across buildings or areas with severe electromagnetic interference.

Step 4: Select protocol and device profiles

Choose a protocol supported by the required controllers and devices. Confirm actual interoperability, not merely a product label. Check device description files, firmware versions, connection limits, update rates, and diagnostic functions.

Step 5: Engineer addressing and data models

Create an address plan, device naming convention, register map, scaling table, time-stamping policy, and quality-state representation. Document byte order explicitly.

Step 6: Engineer security and resilience

Use zones and conduits, least-privilege access, managed switches, secure remote access, backups, patch procedures, port controls, logging, and network monitoring. Separate safety functions from ordinary control and information traffic where required.

Step 7: Validate and commission

Perform cable tests, device identification, configuration verification, traffic measurements, fault injection, redundancy testing, recovery-time tests, and acceptance tests. Record baseline diagnostics for future maintenance.

9. Illustrative Case: Packaging Cell Integration

Background

A packaging cell contains a PLC, six variable-speed drives, two robot controllers, 24 distributed I/O stations, 18 IO-Link sensors, an HMI, a vision camera, and a plant MES. The existing design used separate serial links and point-to-point wiring, making diagnostics slow and production data incomplete.

Problem

The plant requires 10 ms control updates for drive and I/O data, recipe transfer from the MES, device-level diagnostics, and secure remote maintenance. A network outage should not stop the entire line unnecessarily.

Analysis

The drive and I/O traffic is cyclic and time-sensitive. Sensor parameterization is best handled by IO-Link. MES communication is information-oriented and should not burden the control cycle. The camera produces relatively large data bursts. The cell also contains motors and inverters that create electromagnetic interference.

Approach and solution

The design uses Industrial Ethernet for the PLC, drives, distributed I/O, and robot interfaces. IO-Link masters connect the sensors and expose process data and diagnostics through the controller network. A managed switch provides star and line connections, while a redundant ring is used on the cell backbone. The vision camera is placed in a separately prioritized traffic class.

An industrial firewall separates the cell zone from the plant network. An OPC UA server publishes selected production status, counts, alarms, quality values, and equipment metadata to the MES. The MES cannot directly write motion commands. A gateway is used only for a legacy energy meter using Modbus RTU; its values are mapped with units, timestamps, and quality flags.

Commissioning includes measuring the actual 10 ms update behavior, checking switch utilization, testing ring recovery, disconnecting cables, verifying drive fault diagnostics, and confirming that remote access requires authentication and approval.

Outcome

The integrated design reduces point-to-point wiring, improves device replacement through stored parameters, and provides structured production data. Control traffic remains within the real-time network, while enterprise integration occurs through a controlled information boundary. The principal improvement is not simply higher speed; it is clearer separation of control, diagnostics, and business information.

Learning points

1. Protocol selection must follow traffic and risk requirements.
2. Semantic data mapping is essential for useful MES integration.
3. Redundancy must be tested under realistic faults.
4. A gateway should have a documented responsibility and failure behavior.
5. Cybersecurity and maintainability must be designed from the beginning.

10. Cybersecurity, Safety, and Governance

Industrial communication connects cyber systems to physical processes. Threats include unauthorized commands, ransomware, credential theft, protocol abuse, firmware tampering, denial of service, and manipulation of process data. Controls should include asset inventories, network segmentation, secure configuration, identity management, backups, vulnerability management, monitoring, and incident-response plans.

Security and safety are related but different. A firewall may reduce unauthorized access but does not by itself provide a safety function. Safety communication requires appropriate certified mechanisms, hazard analysis, validation, and compliance with applicable standards such as IEC 61508, IEC 62061, or ISO 13849 where relevant.

Engineers should follow applicable standards and organizational policies, including industrial cybersecurity guidance such as the IEC 62443 family. Data collection should be proportionate to operational needs. Access logs, production records, and worker-related information require appropriate governance and privacy protection.

11. Contemporary Developments and Best Practices

Current industrial systems increasingly combine edge computing, digital twins, machine learning, private wireless networks, and cloud analytics. These technologies can improve predictive maintenance and quality analysis, but they do not remove the need for reliable source data and deterministic control.

AI-generated configuration or diagnostic recommendations must be reviewed by competent engineers. An AI system should not autonomously alter safety parameters or control logic without authorization, validation, and traceability. Training data may contain biased or incomplete operating conditions, and model failures can be difficult to detect.

Recommended practices are:

• design from performance and risk requirements rather than brand preference;
• use documented device profiles and version control;
• separate cyclic control from high-volume information traffic;
• provide diagnostics that identify cable, device, configuration, and application faults;
• use meaningful timestamps and quality states;
• maintain current topology, address, firmware, and backup records;
• test failure recovery and cybersecurity controls;
• plan spare parts and lifecycle support;
• use standards-based interfaces at system boundaries;
• train maintenance personnel in both networking and automation.

Chapter Summary

Industrial communication enables sensors, controllers, actuators, supervisory systems, and enterprise applications to operate as an integrated system. Fieldbus technologies remain valuable for economical device-level communication, while Industrial Ethernet provides greater bandwidth, integration, and flexibility. Modbus is simple and widespread but requires careful data mapping and external security controls. PROFIBUS, PROFINET, EtherNet/IP, CAN, and IO-Link serve different levels and application needs. OPC UA adds secure, structured, semantic communication across automation and information systems.

Reliable design requires attention to topology, physical media, cyclic and acyclic traffic, latency, jitter, availability, diagnostics, redundancy, and cybersecurity. Gateways must translate meaning as well as protocol syntax. Wireless systems require radio and security engineering, and TSN provides a path toward bounded Ethernet communication when its complete configuration is engineered and validated. The central professional principle is to select and integrate technologies according to timing, safety, environmental, lifecycle, and information requirements.

Key Terms

Fieldbus: A digital industrial network connecting multiple field devices over shared communication media.

Industrial Ethernet: Ethernet adapted for industrial environments and automation requirements.

Cyclic communication: Periodic exchange of data at a configured interval.

Acyclic communication: Event-driven or on-demand exchange of data.

Modbus RTU: A serial Modbus implementation commonly used over RS-485.

Modbus TCP: Modbus carried over TCP/IP and Ethernet.

PROFIBUS: A family of industrial fieldbus technologies.

PROFINET: An Industrial Ethernet system for automation communication.

EtherNet/IP: Ethernet communication using the Common Industrial Protocol.

CAN: Controller Area Network using identifier-based arbitration.

CANopen: A higher-layer CAN protocol with standardized device and network services.

IO-Link: Point-to-point communication for intelligent sensors and actuators.

OPC UA: A secure, platform-independent framework for structured industrial information exchange.

Gateway: A device that connects networks or protocols and maps their data.

Determinism: The ability to provide predictable communication timing.

Jitter: Variation in message delay or periodic timing.

TSN: Time-Sensitive Networking standards for bounded and synchronized Ethernet traffic.

Availability: The likelihood that a system is operational when required.

Network segmentation: Dividing a network into controlled zones to limit traffic and risk.

Information model: A structured representation of entities, properties, relationships, and meanings.

Review Questions

Short Answer Questions

1. Distinguish cyclic communication from acyclic communication.
2. Why is Ethernet alone insufficient to guarantee industrial interoperability?
3. State two advantages and two limitations of Modbus.
4. What is the purpose of an industrial gateway?
5. Define jitter and explain why it matters in control systems.

Long Answer Questions

1. Compare PROFIBUS, PROFINET, EtherNet/IP, and Modbus in terms of physical medium, communication model, applications, and limitations.
2. Explain how IO-Link and OPC UA can contribute to different levels of an integrated automation architecture.
3. Describe a systematic method for designing a reliable Industrial Ethernet network.
4. Discuss the role of topology, redundancy, and managed switches in industrial network availability.
5. Explain why cybersecurity, functional safety, and network engineering must be treated as related but distinct disciplines.

Analytical / Application Questions

1. A cell contains 30 devices transmitting 12-byte payloads every 10 ms. Estimate the payload data rate and explain why the selected link capacity must exceed this value.
2. Design a communication architecture for a robotic assembly cell containing distributed I/O, drives, IO-Link sensors, an HMI, a vision system, and MES connectivity. Justify the protocol, topology, and security boundaries.
3. A legacy Modbus RTU energy meter must be integrated into a PROFINET cell and its data published to an MES. Identify the gateway mappings, timing limitations, diagnostics, and cybersecurity controls required.

Multiple Choice Questions with Answers

Question 1

Which communication type is most appropriate for a repeated drive speed reference exchanged every 10 ms?

A. Cyclic communication

B. Acyclic communication

C. Manual file transfer

D. Email messaging

Correct Answer: A. Cyclic communication

Question 2

What is a principal function of an industrial gateway?

A. Increase motor torque

B. Translate data between networks or protocols

C. Replace all network security controls

D. Eliminate the need for device addressing

Correct Answer: B. Translate data between networks or protocols

Question 3

Which technology is a point-to-point system for intelligent sensors and actuators?

A. IO-Link

B. OPC UA

C. PROFIBUS PA only

D. TSN

Correct Answer: A. IO-Link

Question 4

What does CAN arbitration use to determine message priority?

A. Packet length only

B. Identifier priority

C. Device physical size

D. Cable color

Correct Answer: B. Identifier priority

Question 5

Which statement best describes OPC UA?

A. A motor-control power converter

B. A secure framework for structured industrial information exchange

C. A copper cable termination standard

D. A replacement for all safety systems

Correct Answer: B. A secure framework for structured industrial information exchange

Question 6

Ignoring overhead, what is the transmission time for a 1,200-byte frame on a 100 Mbit/s link?

A. 9.6 μs

B. 96 μs

C. 960 μs

D. 9.6 ms

Correct Answer: B. 96 μs

Question 7

Which factor is especially important when designing an industrial wireless network?

A. Radio interference and antenna placement

B. Printer ink capacity

C. Mechanical gear ratio only

D. PLC enclosure color

Correct Answer: A. Radio interference and antenna placement

Question 8

What is a major purpose of TSN?

A. Provide bounded and synchronized Ethernet communication

B. Convert AC power to DC power

C. Replace all application data models

D. Remove the need for network testing

Correct Answer: A. Provide bounded and synchronized Ethernet communication

Question 9

Which Modbus issue commonly requires documentation by the integrator?

A. Register scaling and byte order

B. Wheel diameter of a vehicle

C. Lubricant viscosity only

D. Room lighting color

Correct Answer: A. Register scaling and byte order

Question 10

For a repairable component with MTBF = 10,000 h and MTTR = 10 h, what is the approximate availability using A = MTBF ÷ (MTBF + MTTR)?

A. 90.0%

B. 99.0%

C. 99.90%

D. 99.999%

Correct Answer: C. 99.90%