Automation, Sensors and Industrial Control

Introduction and Context

Modern manufacturing systems depend on coordinated sensing, decision-making and physical action. A machining centre, robotic cell, packaging line or automated inspection station must detect conditions, interpret them, control energy and motion, and respond safely when conditions depart from expectations. Automation therefore combines mechanical engineering, electrical engineering, computer technology, instrumentation, communications and human factors.

In a basic automated system, sensors measure process conditions, a controller evaluates those measurements against a program or control rule, and actuators alter the process. Industrial control extends this basic loop by incorporating feedback, sequencing, diagnostics, networking, supervisory monitoring and safety functions. The engineer’s task is not merely to make a machine operate. It is to design a system that operates accurately, reliably, maintainably and safely throughout its service life.

This chapter develops the principles needed to design basic automation solutions using sensors, actuators, programmable logic controllers (PLCs), control logic, machine vision, industrial networks and supervisory systems. It also examines reliability, maintainability and functional safety, because a technically successful system that is difficult to repair or unsafe to operate is not a successful manufacturing system.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the architecture and operation of an industrial automation system.
2. Select appropriate sensors and actuators for common manufacturing tasks.
3. Develop PLC-based combinational, sequential and feedback control logic.
4. Describe the operation and application of machine-vision systems.
5. Compare common industrial communication and supervisory-control arrangements.
6. Evaluate automation designs using reliability, maintainability and functional-safety principles.
7. Propose and justify a basic automated manufacturing-cell solution.

1. Architecture of an Automated Manufacturing System

An automated system can be viewed as a hierarchy of interacting layers:

• Physical process: machines, workpieces, tooling, material flow and energy sources.
• Field layer: sensors, switches, encoders, valves, motors and drives.
• Control layer: PLCs, programmable automation controllers, motion controllers and safety controllers.
• Supervisory layer: human–machine interfaces (HMIs), supervisory control and data acquisition (SCADA), historians and alarm systems.
• Plant and enterprise layer: manufacturing execution systems, production planning, maintenance management and quality databases.

The layers are functionally distinct but connected. A photoelectric sensor may detect a workpiece, a PLC may determine whether the workpiece is correctly positioned, a drive may command a motor to stop, and an HMI may display the resulting status to an operator.

Figure: Functional layers of a typical industrial automation system; information and commands move in both directions.

A useful design distinction is between open-loop and closed-loop control. In open-loop control, the controller commands an action without measuring the resulting output. A timer-operated lubrication pump is an example. In closed-loop control, a sensor measures the output and the controller adjusts the input in response. Temperature control using a thermocouple and heater is a closed-loop system.

Feedback improves disturbance rejection and accuracy, but it introduces additional components and possible failure modes. Sensor errors, measurement noise, delays and poorly tuned controllers can cause instability or poor performance. Feedback must therefore be designed, tested and maintained rather than added without analysis.

2. Sensors and Measurement

A sensor converts a physical variable into a usable electrical, digital or network signal. Common measured variables include position, proximity, displacement, speed, force, pressure, temperature, flow, level, vibration and optical intensity.

Important sensor-selection criteria include:

• Range: the minimum and maximum measurable values.
• Resolution: the smallest distinguishable change.
• Accuracy: closeness to the accepted true value.
• Repeatability: agreement among repeated measurements under the same conditions.
• Response time: time required to respond to a change.
• Environmental suitability: resistance to dust, coolant, vibration, temperature and electromagnetic interference.
• Output and interface: switching, voltage, current, pulse, fieldbus or industrial Ethernet.
• Installation and maintenance: accessibility, alignment, calibration and replacement requirements.

Accuracy and repeatability are not identical. A sensor can be highly repeatable but inaccurate if it has a consistent calibration offset. In production, repeatability may be sufficient for presence detection, whereas dimensional inspection requires calibrated accuracy and an understanding of measurement uncertainty.

2.1 Common Industrial Sensors

Inductive proximity sensors detect metallic objects without contact. They are robust and appropriate for position confirmation, but their sensing range is relatively short and depends on target material and geometry.

Capacitive sensors detect both metallic and many non-metallic materials, including plastics, powders and liquids. They can be affected by humidity, material buildup and changes in dielectric properties.

Photoelectric sensors use emitted and received light. Through-beam arrangements generally provide long range and reliable detection; retroreflective and diffuse arrangements reduce installation complexity but may be more sensitive to target reflectivity and background conditions.

Ultrasonic sensors measure distance or detect objects using sound waves. They are useful for transparent objects and level measurement, although air temperature, turbulence and surface geometry influence performance.

Encoders measure rotary position or speed. Incremental encoders produce pulses and require a reference or homing procedure after power loss. Absolute encoders provide a coded position value and retain positional information without the same homing requirement.

Temperature sensors include resistance temperature detectors, thermocouples and semiconductor devices. A thermocouple generates a small voltage related to temperature difference and requires appropriate signal conditioning and reference-junction compensation.

Pressure and force sensors commonly use strain-gauge or piezoresistive elements. Their outputs may be analog, digitally communicated or integrated into a smart instrument.

Sensor wiring must also be considered. A 24 V discrete input may represent an on/off condition, while a 4–20 mA analog loop represents a continuous measurement. The 4–20 mA standard permits a non-zero lower value, allowing a controller to distinguish a valid zero measurement from a broken wire in many installations.

For a linear analog signal, engineering value can be obtained by interpolation:

Here, x is the measured engineering value, I is the measured current, and x_min, x_max, I_min and I_max are the corresponding lower and upper scale limits. The relationship assumes a correctly calibrated linear transmitter and a current within the specified range.

Worked example. A pressure transmitter is scaled from 0 to 10 bar over 4–20 mA. If the measured current is 12 mA:

x = 0 + (12 - 4)/(20 - 4) × 10 = 5 bar.

The result is 5 bar, assuming the input signal is valid and the transmitter is linear.

3. Actuators and Final Control Elements

Actuators convert controller commands into physical action. Typical examples are motors, solenoid valves, pneumatic cylinders, hydraulic cylinders, heaters, relays and robotic joints.

A motor may be controlled through a contactor for simple on/off operation, a variable-frequency drive for adjustable speed, or a servo drive for coordinated position and motion control. Pneumatic systems offer rapid and clean linear movement but may have compliance and difficulty achieving precise intermediate positions. Hydraulic systems provide high force density but require attention to leakage, fluid condition and energy isolation.

A final control element should be selected by considering force or torque, speed, duty cycle, accuracy, energy efficiency, environmental conditions, braking requirements and failure response. The design must also specify what happens when power, communication or control signal is lost. For example, a valve may be required to fail closed to prevent hazardous flow, whereas a cooling valve may need to fail open.

Actuator sizing requires more than selecting a nominal rating. The engineer should account for friction, acceleration, external loads, transmission efficiency, duty cycle and safety factors. Excessive oversizing can reduce controllability and increase cost; undersizing causes overheating, missed positions or premature failure.

4. PLCs and Control Logic

A PLC is an industrial computer designed for deterministic control, electrical noise tolerance and maintainable programming. A typical scan consists of reading inputs, executing the user program, updating outputs and performing communications and diagnostics. The exact sequence varies by platform, but the scan-cycle concept is fundamental.

PLC programs commonly use ladder diagrams, function-block diagrams, structured text and sequential-function charts. Ladder logic resembles relay circuits and is accessible to maintenance personnel. Structured text is effective for calculations and data handling. Function blocks support reusable control structures.

4.1 Combinational and Sequential Logic

Combinational logic depends only on present inputs. For example, a motor may be permitted to run when the start command is true, the guard is closed and the overload is healthy:

RunPermit = Start ∧ GuardClosed ∧ OverloadHealthy.

Sequential logic depends on present inputs and previous states. A conveyor may start only after a start command, remain running through a seal-in condition, and stop when a stop command or fault occurs. A robust sequence normally includes explicit states such as Idle, Loading, Clamping, Machining, Unloading and Fault.

A conventional motor-control pattern uses a normally closed stop condition, a start command and a run feedback or seal-in bit. In modern systems, however, the safety function should not rely solely on ordinary PLC logic. Emergency-stop and guard-monitoring functions typically require a safety relay or safety PLC and appropriate energy isolation.

4.2 Timers, Counters and Interlocks

Timers support delays, debounce periods and maximum-operation limits. Counters support batch quantities and production counts. Interlocks prevent incompatible actions, such as extending a cylinder when a guard is open or commanding two opposing valves simultaneously.

Good logic design makes abnormal states explicit. A timeout should not simply stop a sequence without explanation; it should generate a diagnostic message identifying the expected condition, elapsed time and likely causes. Inputs from limit switches should be checked for contradictory states, such as both extended and retracted sensors being active when that is physically impossible.

4.3 Feedback Control

For a simple proportional controller, the control signal is:

where u(t) is the controller output, e(t) = r(t) - y(t) is the error, r(t) is the reference or setpoint, y(t) is the measured output and K_p is the proportional gain. This equation assumes a proportional-only controller and compatible scaling of input and output. A large gain may reduce error but can produce oscillation, actuator saturation or sensitivity to noise.

Industrial temperature, pressure and speed systems often use proportional–integral–derivative control. In ideal continuous form:

The integral term removes persistent steady-state error, while the derivative term responds to the rate of error change. Practical implementations require sampling, output limits, anti-windup treatment and filtering of derivative action. PID tuning must reflect process dynamics and cannot be selected reliably from gain values alone.

5. Machine Vision

Machine vision uses illumination, optics, an image sensor and software to make decisions from images. A typical inspection sequence is:

1. Present and stabilize the part.
2. Trigger controlled illumination and image capture.
3. Locate the region of interest.
4. Process the image using thresholding, edge detection, geometric measurement, pattern matching or classification.
5. Compare results with acceptance criteria.
6. Communicate pass/fail information to the PLC and record traceability data.

Lighting is often more important than the algorithm. Diffuse lighting can reduce glare, backlighting can produce a clear silhouette, and structured lighting can reveal surface geometry. Camera resolution, lens selection, field of view and working distance determine whether the required feature can be measured.

Vision systems are affected by vibration, contamination, changing ambient light, part orientation and surface finish. A pass/fail classifier should be validated using representative good and defective samples. False acceptance can create safety or quality risks; false rejection causes waste and may indicate unstable presentation or over-restrictive thresholds. Artificial intelligence and deep-learning vision can classify complex patterns, but they require representative training data, version control, performance monitoring and human review of uncertain cases. They do not eliminate the need for controlled optics and sound process design.

6. Industrial Networks and Supervisory Systems

Industrial networks connect field devices, controllers, drives, HMIs and higher-level systems. The choice of network depends on determinism, bandwidth, topology, distance, interoperability, diagnostics, installed base and cybersecurity requirements.

Common categories include sensor-level networks, controller networks and industrial Ethernet. Protocols such as Modbus, PROFINET, EtherNet/IP, EtherCAT and OPC UA serve different purposes and have different implementation characteristics. A protocol name alone does not guarantee performance; network loading, update time, switch configuration and device behavior must be evaluated.

An HMI presents operating status, commands, trends, alarms and maintenance information. It should support rapid comprehension rather than reproduce every internal variable. Alarms should be prioritized, actionable and limited to conditions requiring attention. Alarm floods during a fault can obscure the initiating event.

SCADA systems supervise geographically distributed or multi-cell processes. They may collect historical data for production analysis, energy monitoring, quality correlation and predictive maintenance. Supervisory software must not be treated as the primary safety layer. Loss of an HMI or network should not create an unsafe machine state.

Industrial connectivity increases the attack surface. Good practice includes network segmentation, least-privilege accounts, controlled remote access, secure backups, patch management, asset inventories, change control and incident-response procedures. Cybersecurity and functional safety are related but distinct: cybersecurity protects systems from unauthorized or malicious actions, while functional safety reduces risk from hazardous failures and control behavior.

7. Reliability, Maintainability and Functional Safety

Reliability is the probability that a system performs its required function for a specified time under stated conditions. For a constant failure rate, a simple exponential model is:

where R(t) is reliability at time t, and λ is the constant failure rate in failures per unit time. The model assumes statistically independent failures and a constant rate; it is not universally valid, especially during early-life or wear-out periods.

For independent components connected in series, system reliability is:

A series system fails when any essential component fails. Redundancy can improve reliability, but only if redundant channels are sufficiently independent and common-cause failures are controlled.

Maintainability concerns how quickly and effectively a failed system can be restored. It is improved by modular construction, accessible components, clear diagnostics, standard connectors, documentation, spare-parts planning and safe isolation procedures. Mean time to repair (MTTR) includes fault identification, access, replacement, testing and return to service—not merely the time spent changing a part.

Availability is often approximated by:

where MTBF is mean time between failures and MTTR is mean time to repair, both expressed in the same time units. This steady-state approximation assumes repeated operating and repair cycles and does not account for scheduled maintenance, logistics delays or complex failure distributions.

Functional safety begins with hazard identification and risk reduction. A typical process includes defining machine limits, identifying hazards, estimating risk, specifying safety functions, designing protective measures, validating performance and managing changes. Examples of safety functions include preventing motion when a guard is open, stopping hazardous movement after emergency actuation, limiting speed in an access mode and preventing unexpected restart.

Safety functions should be separated from ordinary production control where necessary. A risk assessment determines the required performance level or safety integrity target under applicable regulations and standards. The design must consider architecture, diagnostic coverage, common-cause failure, component reliability and validation. A reset command should not automatically restart hazardous motion unless the risk assessment and design explicitly permit it; reset should generally require a deliberate and safe restart sequence.

8. Design Method for a Basic Automation Solution

A disciplined design procedure is more valuable than selecting components prematurely.

1. Define the process and success criteria. Identify product flow, cycle time, quality requirements, utilities and operator interaction.
2. Map inputs and outputs. List sensors, actuators, safety devices, analog values, communications and diagnostic signals.
3. Identify hazards. Consider crushing, cutting, entanglement, ejection, heat, stored energy, unexpected startup and electrical hazards.
4. Select the control architecture. Decide on PLC type, remote I/O, drives, safety controller, network and HMI.
5. Develop the sequence. Use states, transitions, permissives, interlocks, fault handling and recovery procedures.
6. Specify devices. Check electrical compatibility, environmental ratings, accuracy, response time, installation and maintainability.
7. Design software and data structures. Use meaningful tags, modular routines, diagnostic codes and controlled versions.
8. Verify and validate. Test normal operation, boundary conditions, sensor failures, communication loss, power restoration and safety functions.
9. Document and train. Provide schematics, I/O lists, network diagrams, programs, risk assessment, maintenance instructions and operator training.
10. Monitor performance after commissioning. Use quality, downtime, alarm and maintenance data to improve the system.

Applied Case: Illustrative Automated Inspection and Sorting Cell

Background

A small manufacturing company produces machined aluminium housings. Each housing must be present, correctly oriented, within an acceptable height range and free from a major surface defect before packaging. Manual inspection is slow and inconsistent.

Problem

The proposed cell must detect parts on a conveyor, stop each part at an inspection station, verify orientation and height, use vision to identify a visible defect, and divert rejected parts. Operators must be protected from the conveyor and pneumatic ejector.

Analysis

A photoelectric sensor detects arrival. A diffuse or structured light source and camera inspect the top surface. A laser displacement sensor measures height if the tolerance is too small for a simple switch. A pneumatic cylinder diverts rejected parts. The PLC coordinates the sequence, while a safety controller monitors the access guard and emergency-stop circuit.

The principal failure modes include missed arrival detection, camera contamination, incorrect part orientation, cylinder failure to retract, loss of network communication and unexpected restart after power restoration. These require diagnostics and safe responses rather than only normal-cycle programming.

Approach and Solution

The cell uses a conveyor motor drive, an arrival sensor, a stop cylinder with extended and retracted feedback, a vision camera, a height sensor, a reject cylinder with position feedback, a PLC, a safety controller and an HMI. The sequence is:

• Idle until the guard is closed, safety circuit is healthy and the operator selects automatic mode.
• Convey a part until the arrival sensor is active.
• Stop and clamp the part; verify the clamp position within a time limit.
• Trigger the camera and height measurement.
• Accept the part only if orientation, defect result and height are within validated limits.
• Release accepted parts or extend the reject cylinder for rejected parts.
• Confirm the reject cylinder returns before permitting the next cycle.
• Record the result and display a diagnostic message if any timeout or contradictory sensor state occurs.

The safety controller removes hazardous actuator energy when a guard opens or emergency stop is activated. The standard PLC handles production sequencing but cannot override the safety function. The HMI displays machine state, active alarms, inspection counts and maintenance prompts.

Outcome

In this illustrative case, the design reduces manual inspection and provides traceable reject data. Its success depends on validating lighting and camera thresholds with samples from the full production range, checking sensor response under contamination, and testing all safety functions under documented conditions. The most important improvement is not the camera alone but the integration of sensing, control, diagnostics and safe recovery.

Learning Points

1. A sensor should be selected from the physical measurement problem, not from its popularity.
2. Normal sequence logic and safety logic have different purposes.
3. Feedback and diagnostic confirmation prevent silent failures.
4. Vision performance depends strongly on presentation and illumination.
5. Maintainability must be designed through access, feedback, documentation and fault messages.

Challenges and Limitations

Automation can increase productivity while also introducing integration complexity. Poorly specified sensors create nuisance stops; excessive interlocks reduce availability; inadequate diagnostics increase repair time; and unvalidated vision systems create hidden quality risks. Networked systems introduce configuration and cybersecurity dependencies. Automation may also alter work roles, requiring retraining and careful communication with employees.

The most sophisticated technology is not always the best solution. A reliable mechanical guide and a simple proximity sensor may outperform a complex vision system when the process can be physically standardized. Conversely, a simple sensor may be unsuitable where product variation or traceability demands image evidence. Engineering judgment requires comparing technical performance, lifecycle cost, safety, maintainability and human usability.

Contemporary Developments and Best Practices

Current automation practice is moving toward modular machines, edge computing, digital twins, condition monitoring and interoperable information models. These technologies can improve commissioning and maintenance, but their value depends on trustworthy data and clear operational objectives. Predictive maintenance models should be evaluated against actual failure and intervention records rather than accepted because an algorithm produces a prediction.

Best practices include:

• Use standardized naming, symbols, wiring conventions and reusable software modules.
• Separate safety functions from ordinary sequence logic and validate them independently.
• Design for diagnostic coverage and safe fault recovery.
• Protect sensors and connectors from the process environment.
• Record calibration, software versions and parameter changes.
• Use alarm priorities based on operator action and consequence.
• Segment industrial networks and control remote access.
• Involve operators and maintenance personnel during design reviews.
• Test power loss, communication loss, sensor failure and restart behavior.
• Treat lifecycle documentation and training as engineering deliverables.

Chapter Summary

Industrial automation integrates sensors, controllers, actuators, networks and supervisory systems to convert manufacturing requirements into coordinated physical action. Sensors must be selected according to range, accuracy, repeatability, response, environment and interface. Actuators must be sized for load, speed, duty cycle, controllability and safe failure behavior.

PLCs execute deterministic control programs that include combinational logic, sequences, timers, counters, interlocks and feedback. Machine vision adds image-based measurement and classification but depends on controlled presentation, lighting, optics, validation and data governance. Industrial networks provide communication between field, control and supervisory layers, while HMI and SCADA systems support operation, alarms, trends and traceability.

A professional automation design addresses reliability, maintainability, cybersecurity and functional safety from the beginning. Risk assessment, safety functions, diagnostics, validation and documentation are essential. Effective automation is therefore not simply the replacement of manual work with machines; it is the disciplined design of a dependable socio-technical system.

Key Terms

Sensor: Device that detects or measures a physical variable.

Actuator: Device that converts a control command into physical action.

PLC: Industrial computer designed for deterministic machine control.

Feedback: Information about a measured output returned to a controller.

Open-loop control: Control without measurement of the resulting output.

Closed-loop control: Control that uses measured output to influence the command.

Interlock: Condition that prevents an unsafe or incompatible action.

HMI: Human–machine interface used for operator interaction and information display.

SCADA: Supervisory control and data acquisition system.

Machine vision: Use of cameras, lighting and software for automated inspection or guidance.

Encoder: Sensor that measures rotary position, displacement or speed.

Resolution: Smallest distinguishable change in a measurement.

Repeatability: Agreement among repeated measurements under the same conditions.

Reliability: Probability of performing a required function for a stated time and condition.

Maintainability: Ease and speed with which a system can be restored after failure.

Availability: Proportion of time a system is operational and able to perform its function.

Functional safety: Reduction of risk through safety-related control functions.

Fail-safe: Design behavior that moves toward a safer condition after a defined fault.

Industrial network: Communication system connecting automation devices.

Alarm management: Design and control of messages requiring operator attention.

Review Questions

Short Answer Questions

1. Distinguish between open-loop and closed-loop control, giving one manufacturing example of each.
2. What is the difference between sensor accuracy and repeatability?
3. State two reasons why lighting is important in machine vision.
4. What is the purpose of an interlock in PLC control?
5. Why should a safety function not normally depend only on an ordinary PLC program?

Long Answer Questions

1. Explain the functional layers of an industrial automation system and describe the information exchanged between them.
2. Compare inductive, capacitive, photoelectric and ultrasonic sensors for manufacturing applications.
3. Explain how a PLC executes a sequential process, including states, transitions, timers, interlocks and fault handling.
4. Discuss the relationship among reliability, maintainability and availability in automated equipment.
5. Describe the stages involved in developing and validating a functional-safety solution for a machine.

Analytical / Application Questions

1. Design a basic control architecture for a conveyor that detects, inspects and rejects parts. Identify sensors, actuators, PLC functions, safety functions, network connections and HMI information.
2. A 4–20 mA transmitter is scaled from 0 to 100 °C and produces 16 mA. Calculate the temperature and explain the assumptions behind the calculation.
3. An automated cell has an MTBF of 800 hours and an MTTR of 8 hours. Calculate its approximate steady-state availability and identify two practical actions that could improve it.

Multiple Choice Questions with Answers

Question 1

Which device is most suitable for non-contact detection of a steel part at a short range?

A. Inductive proximity sensor

B. Thermocouple

C. Pressure transmitter

D. Load cell

Correct Answer: A. Inductive proximity sensor

Question 2

A 4–20 mA transmitter scaled from 0 to 10 bar outputs 12 mA. What pressure does this represent?

A. 2.5 bar

B. 5 bar

C. 7.5 bar

D. 8 bar

Correct Answer: B. 5 bar

Question 3

Which statement best describes closed-loop control?

A. The controller acts only according to a fixed timer

B. The output is never measured

C. Measured output is used to influence the control command

D. The actuator is disconnected from the process

Correct Answer: C. Measured output is used to influence the control command

Question 4

What is the primary purpose of an interlock?

A. To increase network bandwidth

B. To prevent incompatible or unsafe actions

C. To calibrate an analog sensor

D. To store historical data

Correct Answer: B. To prevent incompatible or unsafe actions

Question 5

Which machine-vision factor most directly controls glare and feature contrast?

A. Lighting arrangement

B. PLC scan time only

C. Motor overload rating

D. Network password length

Correct Answer: A. Lighting arrangement

Question 6

For independent components in a series system, system reliability is calculated by:

A. Adding all component reliabilities

B. Subtracting the lowest reliability from one

C. Multiplying the component reliabilities

D. Dividing reliability by repair time

Correct Answer: C. Multiplying the component reliabilities

Question 7

If MTBF is 900 hours and MTTR is 9 hours, the approximate availability is:

A. 90.0%

B. 98.0%

C. 99.0%

D. 99.9%

Correct Answer: C. 99.0%

Question 8

Which function is normally associated with a safety controller rather than ordinary production sequencing?

A. Counting accepted parts

B. Displaying production totals

C. Preventing hazardous motion when a guard is open

D. Calculating a batch average

Correct Answer: C. Preventing hazardous motion when a guard is open

Question 9

What is a principal advantage of an absolute encoder over an incremental encoder?

A. It requires no electrical connection

B. It can retain position information without the same homing requirement

C. It measures temperature directly

D. It cannot be affected by vibration

Correct Answer: B. It can retain position information without the same homing requirement

Question 10

Which practice best supports industrial-network cybersecurity?

A. Connecting every device to an unrestricted public network

B. Disabling all diagnostic records

C. Using network segmentation and controlled remote access

D. Allowing shared administrator passwords

Correct Answer: C. Using network segmentation and controlled remote access