6. Industrial Automation, Robotics and Flexible Manufacturing

Introduction and Context

Industrial automation is the coordinated use of control systems, sensing devices, actuators, software, machines, and communication networks to perform manufacturing operations with limited direct human intervention. It is not simply the replacement of workers by machines. Rather, it is a production-system design decision concerning how information and physical work should flow through a factory.

Automation has become central to modern production management because manufacturers must simultaneously pursue productivity, quality, delivery reliability, flexibility, safety, and cost control. These objectives may conflict. A highly specialized transfer line can produce very large volumes at low unit cost, but it may be uneconomical when demand is uncertain or product variety is high. A flexible manufacturing system can accommodate variety, but it requires greater investment, sophisticated control, and capable technical personnel.

The production manager therefore has to select an appropriate degree and type of automation. The key question is not “How much automation is possible?” but “Which activities should be automated, at what level, and with what economic and organizational consequences?” This chapter examines fixed, programmable, and flexible automation; the roles of sensors, actuators, programmable logic controllers, machine vision, robots, collaborative robots, automated material handling, and integrated manufacturing systems; and the methods used to justify and govern automation investments.

Learning Objectives

After studying this chapter, the learner should be able to:

1. Distinguish fixed, programmable, flexible, and integrated automation.

2. Explain the interaction of sensors, controllers, actuators, software, and communication networks.

3. Evaluate PLC-based control, machine vision, industrial robots, and cobots for production applications.

4. Design the basic architecture of an automated manufacturing cell and an FMS.

5. Select suitable automated material-handling and storage technologies.

6. Assess safety, cybersecurity, workforce, maintenance, and ethical implications.

7. Justify automation economically using capacity, quality, labour, downtime, and life-cycle measures.

8. Critically evaluate implementation risks and emerging developments.

Automation as a Production-System Choice

Automation may be classified according to the relationship between equipment design and product variety. In fixed automation, the sequence of operations is built into the equipment through dedicated mechanisms, tooling, transfer devices, or hard-wired controls. It is appropriate for high-volume, stable production. Its strengths are high throughput, consistent cycle time, and low variable cost. Its weakness is limited adaptability: a product change may require substantial redesign.

Programmable automation uses equipment whose operation can be changed through programs, tooling changes, or parameter adjustments. Numerical-control machine tools, industrial robots, and batch-processing equipment are common examples. It is suitable for batch production and moderate product variety. Changeover time, programming effort, and verification become important performance factors.

Flexible automation extends programmability by enabling different products to be manufactured with little or no manual intervention during changeover. A flexible manufacturing system normally combines computer-controlled machine tools, automatic tool and pallet handling, product identification, supervisory control, and production-planning interfaces. Flexibility may refer to product mix, routing, volume, expansion, or recovery from disturbances. It should therefore be measured rather than assumed merely because programmable equipment has been installed.

Integrated automation connects production equipment with planning, quality, maintenance, logistics, and business information systems. This may involve manufacturing execution systems, enterprise resource planning, industrial internet platforms, digital twins, and real-time production data. Integration increases visibility and coordination but also expands cybersecurity, data-governance, and systems-integration responsibilities.

A useful production-management perspective is the volume–variety relationship. High volume and low variety generally support fixed automation; moderate volume and variety support programmable automation; and high variety with recurring families of products may justify flexible automation. However, demand volatility, labour availability, quality risk, safety exposure, floor-space constraints, and strategic capabilities must also be considered.

The Architecture of an Automated System

An automated system can be understood as a closed-loop arrangement consisting of a physical process, sensors, a controller, actuators, human interfaces, and communication infrastructure. Sensors collect information about the process. The controller compares actual conditions with desired conditions and issues commands. Actuators convert those commands into physical action. Feedback allows the system to correct deviations.

Sensors may detect position, proximity, temperature, pressure, force, speed, level, vibration, colour, dimension, or presence. Common devices include inductive and photoelectric proximity sensors, encoders, thermocouples, pressure transducers, load cells, barcode readers, RFID devices, and cameras. Sensor selection depends on range, accuracy, response time, environmental resistance, installation, calibration, and diagnostic capability. A sensor that is accurate in a laboratory may be unreliable in the presence of coolant, dust, vibration, glare, or electromagnetic interference.

Actuators include electric motors, servo motors, pneumatic cylinders, hydraulic systems, solenoid valves, grippers, drives, and heating elements. Selection involves force or torque, speed, positioning accuracy, duty cycle, energy consumption, controllability, maintainability, and safety. Pneumatics are often economical for simple high-speed movements, whereas servo systems are preferred where precise motion and programmable profiles are required.

The programmable logic controller is a central industrial controller designed for reliable sequential and logical control. A PLC scans inputs, executes a control program, and updates outputs. Its typical elements are a processor, input and output modules, power supply, communications interfaces, programming software, and diagnostic functions. Ladder logic remains widely used because it resembles relay-control diagrams, while structured text, function-block diagrams, and sequential-function charts support more complex applications.

A PLC should not be treated as a complete automation strategy. The design must address input filtering, emergency stops, interlocking, fail-safe states, alarm management, network communication, backup, program version control, and maintenance access. A control system that operates correctly under normal conditions but fails dangerously during a sensor fault is poorly designed.

Supervisory control and data acquisition systems, human–machine interfaces, and manufacturing execution systems operate at higher levels. The HMI presents status, alarms, recipes, and trends to operators. SCADA systems supervise geographically distributed or process-oriented operations. MES coordinates production orders, work instructions, traceability, quality records, and performance data. Clear separation of control and supervisory functions improves reliability and cybersecurity.

Machine Vision and Data-Based Inspection

Machine vision uses cameras, optics, lighting, image-processing algorithms, and decision rules to inspect or guide production. A typical vision application includes image acquisition, preprocessing, feature extraction, classification or measurement, and a pass/fail or motion decision. Applications include dimensional verification, surface inspection, assembly confirmation, character recognition, orientation detection, and robot guidance.

Lighting is often more important than the camera itself. Diffuse, back, coaxial, or structured lighting can improve contrast and reduce reflections. The system must be validated against variation in colour, finish, position, and environmental conditions. False rejects increase cost, while false accepts threaten quality and customer satisfaction. Performance should therefore be expressed through measures such as detection probability, false-rejection rate, repeatability, and measurement uncertainty.

Artificial intelligence and machine learning are increasingly used where defect appearance is variable or difficult to describe through fixed rules. Nevertheless, a trained model requires representative data, controlled labelling, monitoring for data drift, and defined escalation procedures. AI does not eliminate the need for process capability, root-cause analysis, or human accountability.

Industrial Robots and Collaborative Robots

An industrial robot is a reprogrammable, automatically controlled manipulator capable of movement in several axes. Major robot configurations include articulated, SCARA, Cartesian, cylindrical, and delta robots. Selection depends on payload, reach, repeatability, speed, workspace, wrist configuration, environmental rating, tooling, and programming requirements.

Robots are used for welding, painting, assembly, machine tending, palletizing, packaging, dispensing, inspection, and material transfer. The robot itself is only one part of a robotic work cell. The complete cell may include an end effector, fixture, part-present sensors, safety devices, vision, conveyors, PLC control, and a means of loading and unloading. Poor fixture design, uncertain part presentation, and excessive changeover effort commonly reduce the value of an otherwise capable robot.

End effectors determine how the robot interacts with the product. Grippers may be mechanical, pneumatic, magnetic, vacuum-based, or adaptive. Tool changing can increase flexibility but adds cost, weight, maintenance, and failure modes. Robot programming may be performed by teach pendant, offline programming, simulation, or increasingly by more intuitive interfaces.

Collaborative robots, or cobots, are designed to operate in applications where people and robots share a workspace under specified conditions. Their force, speed, power, and collision-detection features can support low-volume assembly, inspection, machine tending, and ergonomic assistance. A cobot is not automatically safe merely because it is marketed as collaborative. The application must be risk-assessed, including the robot, tool, workpiece, pinch points, sharp edges, speed, force, foreseeable misuse, and operating environment. Guarding or restricted access may still be necessary.

The managerial choice between a conventional industrial robot and a cobot depends on cycle time, payload, required separation, product mix, floor space, worker interaction, and risk. Conventional robots often provide higher speed and payload inside guarded cells; cobots may offer easier redeployment and closer human interaction, but not necessarily higher productivity for every task.

Automated Material Handling and Storage

Material handling strongly influences the performance of an automated factory. A fast machine cell can still have poor productivity if parts arrive late, pallets are misplaced, or finished goods accumulate. Technologies include conveyors, automated guided vehicles, autonomous mobile robots, robotic carts, cranes, lifts, pallet systems, automated storage and retrieval systems, and modular shelving systems.

Automated guided vehicles generally follow predefined routes or guidance systems, whereas autonomous mobile robots use mapping, sensors, and dynamic navigation to select routes. Selection should consider payload, travel distance, traffic, floor condition, charging, interoperability, safety, and the stability of material presentation. Automation of transport without standard containers, identification, and location control often produces confusion rather than flow improvement.

An automated storage and retrieval system can improve space utilization, inventory accuracy, retrieval speed, and traceability. Its economic value depends on transaction volume, storage density, service requirements, building constraints, system availability, and integration with inventory records. Buffer capacity must be designed carefully: too little buffer starves machines, while too much conceals problems and increases work-in-process.

Designing a Flexible Manufacturing System

An FMS is not simply a collection of CNC machines. It is an integrated arrangement designed to manufacture a family of parts through alternative or shared routings. Its principal elements normally include computer numerical-control machines, automatic tool systems, fixtures and pallets, automated transport, part-identification devices, a central or distributed control system, and production-planning interfaces.

FMS design begins with product-family analysis. Parts are grouped according to geometry, material, processing requirements, tolerance, tooling, and routing. The designer then studies demand, operation times, machine capacity, setup requirements, tool life, fixture standardization, inspection, and material flow. A process plan should identify which operations are common, which require specialized resources, and where alternative routing can provide resilience.

A simplified design sequence is:

1. Define the strategic problem and performance targets.

2. Analyse demand, product families, routings, and variability.

3. Map the current process and identify quality, flow, and ergonomic losses.

4. Select machines, fixtures, tools, sensors, controllers, and handling systems.

5. Develop control architecture, data interfaces, and operating rules.

6. Simulate capacity, queues, transport, failures, and changeovers.

7. Pilot the system and validate safety, quality, and maintainability.

8. Commission in stages, train personnel, and monitor performance.

Important performance measures include overall equipment effectiveness, throughput, manufacturing lead time, schedule adherence, changeover time, first-pass yield, mean time between failures, mean time to repair, utilization, work-in-process, and energy per unit. OEE is commonly interpreted through availability, performance, and quality, but it must not be used in isolation. A machine can show high utilization while producing the wrong mix or creating excessive inventory.

Discrete-event simulation is valuable before investment because it can test alternative layouts, buffer sizes, dispatching rules, robot assignments, and maintenance policies. Digital twins extend this approach by linking a virtual representation with operational data. The model remains useful only when assumptions, data quality, and model boundaries are transparent.

Economic Justification and Organizational Change

Automation investment requires a life-cycle business case. Initial capital expenditure may include equipment, tooling, software, integration, building modification, utilities, training, validation, and contingency. Operating effects include labour, energy, consumables, maintenance, quality losses, downtime, programming, cybersecurity, and eventual replacement.

Traditional measures such as payback period, return on investment, net present value, and internal rate of return are useful but incomplete. Payback is simple but ignores benefits after the payback date and the time value of money. Net present value is more rigorous because it discounts future cash flows. Sensitivity analysis should examine demand, uptime, labour cost, quality improvement, implementation delay, maintenance cost, and product life.

Benefits may arise from reduced direct labour, higher capacity, fewer defects, lower injury exposure, shorter lead time, improved traceability, reduced material waste, and greater resilience. Some benefits are difficult to monetize, but they should not be treated as automatically zero. A transparent business case should identify assumptions and distinguish committed benefits from uncertain strategic options.

Automation changes jobs rather than simply eliminating work. Operators may become cell technicians, robot programmers, quality analysts, or maintenance specialists. Successful implementation involves early consultation, skills assessment, reskilling, revised work design, and attention to workload and decision authority. Resistance often reflects legitimate concerns about job security, reliability, or loss of craftsmanship. Technical deployment without organizational preparation usually reduces adoption.

Safety, Cybersecurity and Governance

Safety must be designed into the system rather than added after commissioning. Risk assessment should identify hazards associated with motion, stored energy, crushing, cutting, heat, electricity, pressure, unexpected restart, tooling, maintenance, and human error. Safeguarding may involve fixed guards, interlocked gates, light curtains, laser scanners, two-hand controls, safety-rated monitoring, lockout/tagout, safe speed, safe torque off, and clearly defined operating zones.

Safety functions should be validated under normal, abnormal, maintenance, cleaning, setup, and recovery conditions. Emergency stops are important but do not replace risk reduction, guarding, and safe procedures. Documentation, training, inspection, and management of change are essential because a safe design can become unsafe after a new tool, program, product, or bypass procedure is introduced.

Connected automation creates cybersecurity exposure. Threats include unauthorized access, ransomware, manipulated parameters, insecure remote maintenance, unpatched devices, removable media, and compromised suppliers. A practical governance approach uses asset inventories, network segmentation, least-privilege access, strong authentication, secure backups, patch management, monitoring, incident response, and controlled remote access. Availability and safety are especially important in operational technology; a security measure that disrupts a critical control function must be engineered carefully.

Data governance also matters. Production data may reveal employee performance, proprietary processes, supplier information, or customer requirements. Organizations should define ownership, retention, access, accuracy, and permitted uses. Automated decisions affecting workers require transparency, review, and protection against unfair surveillance or misleading performance comparisons.

Illustrative Case: Flexible Automation in a Pump-Component Plant

Background

The following is an illustrative case. A medium-sized plant manufactures several families of pump components. Demand is stable overall but varies by customer and season. The plant uses conventional machining centres, manual fixture changes, paper-based dispatching, and forklift transport.

Situation and Problem

The plant experiences long changeovers, inconsistent production records, frequent machine starvation, and late deliveries. Management proposes a fully automated FMS with robotic loading and an automated storage system. The initial proposal assumes that labour savings alone will justify the investment.

Analysis

A cross-functional team maps product families and discovers that only 62 percent of components share compatible fixturing and tooling. Several low-volume parts require special inspection and cannot be efficiently processed in the proposed cell. Historical data also show that machine downtime is caused partly by tool-management errors and delayed maintenance. Consequently, buying additional machines would not by itself solve the flow problem.

The team evaluates three alternatives: process improvement with limited automation, a partial FMS for the highest-volume family, and full automation. Simulation indicates that the partial FMS provides most of the achievable lead-time reduction with substantially lower integration risk. It includes standardized pallets, RFID identification, a pallet buffer, tool-life monitoring, a robot for machine tending, and an MES interface. Special parts remain in a controlled conventional area.

Approach and Solution

The plant first standardizes fixtures and cutting tools, introduces preventive maintenance, and cleans production data. It then pilots two machines and a robot cell. Operators are trained as cell technicians, while maintenance staff learn servo, PLC, and network diagnostics. Safety validation is completed before production release. A staged rollout is used, with manual bypass capability during the stabilization period.

Outcome

The pilot improves schedule visibility, reduces unplanned waiting, and shortens changeover for the targeted family. More importantly, the plant avoids over-automating unsuitable products. Performance is reviewed using throughput, first-pass yield, changeover time, downtime, schedule adherence, and total cost rather than robot utilization alone.

Learning Points

The case demonstrates that automation should follow product-family and process analysis. Standardization, data quality, maintenance, skills, and safety are prerequisites. A partial solution may create greater economic value than a technically impressive but poorly matched total system. Automation should be staged so that learning is incorporated before expansion.

Implementation Risks and Common Mistakes

Frequent mistakes include automating an unstable process, selecting technology before defining requirements, underestimating integration costs, ignoring product variation, treating supplier demonstrations as proof of production capability, and failing to budget for spare parts and training. Other errors include excessive customization, weak alarm design, inadequate simulation, poor data interfaces, and measuring only labour reduction.

Automation can also create common-mode failures. When many machines depend on a single network, database, controller, or power system, one fault may stop the entire line. Redundancy, graceful degradation, manual recovery, and clear escalation procedures reduce this exposure. Maintenance must include condition monitoring where appropriate, but predictive analytics should be judged by whether it produces timely and actionable intervention, not by the sophistication of its dashboard.

Current Developments and Best Practices

Current developments include edge computing, industrial wireless networks, interoperable robotics, low-code interfaces, additive-manufacturing integration, AI-assisted inspection, autonomous mobile robots, energy-aware scheduling, and digital twins. These technologies can improve responsiveness, but their value depends on reliable data, process knowledge, cybersecurity, and workforce capability.

Best practice is to begin with a clearly defined production problem and measurable target. Establish a baseline before changing the process. Involve production, maintenance, quality, industrial engineering, information technology, safety, finance, and operators. Use modular architecture and open interfaces where feasible. Standardize data, fixtures, containers, and naming conventions. Simulate and pilot before full deployment. Validate safety and cybersecurity throughout the life cycle. Maintain a skills plan, documented operating procedures, and a management-of-change process. Finally, review the system periodically because product mix, technology, regulation, and business strategy will change.

Chapter Summary

Industrial automation is a production-system choice involving equipment, information, people, economics, and risk. Fixed automation is suited to high-volume, stable production; programmable automation supports batches and changing programs; and flexible automation integrates programmable resources, material handling, identification, and supervisory control for product families. Sensors provide process information, PLCs execute reliable control logic, actuators create physical action, and machine vision supports inspection and guidance. Industrial robots provide speed and repeatability, while cobots enable selected forms of human–robot cooperation subject to application-specific risk assessment.

An FMS must be designed from product families, routings, demand, capacity, tooling, data, and material flow—not from equipment availability alone. Economic justification should include life-cycle costs, uncertainty, quality, capacity, safety, maintenance, and strategic benefits. Successful implementation requires process stability, simulation, staged commissioning, workforce development, safety engineering, cybersecurity, and governance. The central managerial principle is to automate purposefully: technology should improve the total production system rather than merely add machinery.

Key Terms

Automation: Use of control systems and technology to perform work with limited direct human intervention.

Fixed automation: Dedicated automation designed for high-volume, low-variety production.

Programmable automation: Automation whose operating sequence can be changed through programs or settings.

Flexible automation: Integrated programmable automation capable of producing varied products with limited changeover.

PLC: Industrial computer that executes logic and controls machines through input and output signals.

Sensor: Device that detects a physical condition and provides information to a control system.

Actuator: Device that converts a control signal into physical movement, force, heat, or flow.

HMI: Human–machine interface used to display information and accept operator commands.

Machine vision: Camera-based system for inspection, measurement, identification, or guidance.

Industrial robot: Reprogrammable manipulator capable of controlled movement across several axes.

Cobot: Robot designed for specified applications involving physical proximity or interaction with people.

End effector: Tool attached to a robot wrist to perform a task or hold a workpiece.

FMS: Integrated system of programmable machines, handling, fixtures, tools, and control for part families.

AGV: Automated guided vehicle that follows defined guidance or routes.

AMR: Autonomous mobile robot that navigates dynamically using sensors and maps.

MES: Manufacturing execution system that coordinates and records production operations.

OEE: Composite measure based on availability, performance, and quality.

Digital twin: Virtual representation connected to data from a physical asset or process.

Payback period: Time required for cumulative cash benefits to recover an investment.

Risk assessment: Structured identification and evaluation of hazards and required controls.

Review Questions

Short Answer Questions

1. How does flexible automation differ from programmable automation?

2. What are the three basic functions performed by sensors, controllers, and actuators?

3. Why is lighting important in a machine-vision application?

4. State two factors that influence the selection of an industrial robot.

5. Why should an FMS be designed around product families?

Long Answer Questions

1. Explain the architecture of an automated manufacturing system, including sensors, PLCs, actuators, HMIs, and communication systems.

2. Compare fixed, programmable, and flexible automation in terms of volume, variety, flexibility, investment, and risk.

3. Discuss the design and implementation stages of a flexible manufacturing system.

4. Explain the safety and cybersecurity responsibilities associated with connected robotic systems.

5. Describe how a production manager should construct and evaluate an automation business case.

Analytical / Application Questions

1. A plant produces three product families with frequent changeovers and uncertain demand. Recommend an automation strategy and justify it using operational and economic criteria.

2. A robotic cell has high robot utilization but frequent downstream delays and quality rejects. Diagnose possible system-level causes and propose corrective measures.

3. An organization plans to introduce cobots beside manual assembly workers. Develop a practical evaluation framework covering productivity, safety, skills, data governance, and change management.

22. Multiple Choice Questions with Answers

Question 1

A manufacturer produces a very high volume of one stable product with minimal design variation. Which automation type is generally most appropriate?

A. Fixed automation

B. Programmable automation

C. Flexible manufacturing system

D. Manual batch production

Correct Answer: A. Fixed automation

Question 2

In a closed-loop automated system, what is the primary role of a sensor?

A. Convert electrical energy into motion

B. Detect process conditions and provide information

C. Store production schedules

D. Replace the safety assessment

Correct Answer: B. Detect process conditions and provide information

Question 3

Which PLC design practice most directly helps a system respond safely to sensor failure?

A. Removing diagnostic alarms

B. Using fail-safe states and validated interlocks

C. Increasing conveyor speed

D. Disabling program version control

Correct Answer: B. Using fail-safe states and validated interlocks

Question 4

A vision system produces many false rejects because reflective parts appear different under changing illumination. What should be investigated first?

A. Robot payload

B. Lighting and image-acquisition conditions

C. Forklift routing

D. Payback-period calculation

Correct Answer: B. Lighting and image-acquisition conditions

Question 5

Which factor most clearly distinguishes a complete robotic work cell from a robot alone?

A. The colour of the robot

B. Integration of tooling, fixtures, sensing, control, and safety devices

C. The use of manual paperwork

D. The absence of an end effector

Correct Answer: B. Integration of tooling, fixtures, sensing, control, and safety devices

Question 6

Why is a cobot not automatically safe in every application?

A. Cobots cannot be programmed

B. Their safety depends on the specific tool, workpiece, speed, force, and interaction risks

C. Cobots operate only outdoors

D. Cobots do not require risk assessment

Correct Answer: B. Their safety depends on the specific tool, workpiece, speed, force, and interaction risks

Question 7

Which situation most strongly supports an FMS rather than a dedicated transfer line?

A. One product, stable demand, and no expected changes

B. Several related part families requiring alternative routings and recurring changeovers

C. A one-time prototype with no repeat production

D. A process with no programmable equipment

Correct Answer: B. Several related part families requiring alternative routings and recurring changeovers

Question 8

A proposed automation project claims labour savings but ignores training, integration, maintenance, and cybersecurity costs. What is the main weakness?

A. It uses too many sensors

B. It is not a complete life-cycle economic analysis

C. It measures product quality

D. It includes staged commissioning

Correct Answer: B. It is not a complete life-cycle economic analysis

Question 9

Which implementation approach best reduces the risk of over-automating unsuitable products?

A. Automate the entire plant before analysing product families

B. Pilot a targeted product family and expand after validation

C. Eliminate manual bypass capability immediately

D. Use robot utilization as the only performance measure

Correct Answer: B. Pilot a targeted product family and expand after validation

Question 10

What is a key cybersecurity control for connected operational technology?

A. Unrestricted remote access

B. Network segmentation and least-privilege access

C. Shared passwords for all technicians

D. Disabling all backups

Correct Answer: B. Network segmentation and least-privilege access