Computer-Integrated and Digital Manufacturing

Introduction and Context

Manufacturing has evolved from isolated machines and paper-based planning toward connected, data-driven enterprises. In a modern factory, a product definition created in computer-aided design (CAD) may generate manufacturing instructions through computer-aided manufacturing (CAM), receive process plans from computer-aided process planning (CAPP), release work through a manufacturing execution system (MES), and be financially and commercially coordinated by enterprise resource planning (ERP). Product, process, asset, quality, and production data are governed across the product life cycle through product life-cycle management (PLM).

Computer-integrated manufacturing (CIM) is the coordinated use of computers, communications, automation, and information systems to integrate manufacturing activities. Digital manufacturing extends this idea by using structured digital models, connected equipment, analytics, simulation, and digital threads to represent and control manufacturing processes. Integration is not simply connecting software applications. It requires common definitions, reliable interfaces, appropriate authority, traceable transactions, cybersecurity, and disciplined management of data.

The strategic objective is to create a coherent flow of information from customer demand to product design, process planning, production, inspection, delivery, service, and eventual retirement. This chapter explains that flow using CIM and ISA-95 principles, with attention to vertical integration within the enterprise and horizontal integration across products, processes, suppliers, and customers.

Learning Objectives

After studying this chapter, a student should be able to:

1. Explain CIM, digital manufacturing, the digital thread, and the digital twin.
2. Distinguish the roles of CAD, CAM, CAPP, PLM, MES, and ERP.
3. Describe the ISA-95 enterprise-control system hierarchy and its information exchanges.
4. Design a digital manufacturing workflow with appropriate master data and interfaces.
5. Explain production traceability, genealogy, and key performance measures.
6. Analyze vertical and horizontal integration, including benefits, risks, and governance requirements.
7. Evaluate a practical implementation using technical, organizational, and economic criteria.

1. Foundations of Computer-Integrated Manufacturing

1.1 From functional automation to enterprise integration

Traditional automation often optimized an individual machine or department. A CNC machine could execute a program accurately, while production planning, quality records, maintenance information, and inventory records remained separate. Such functional optimization can create local efficiency but poor enterprise performance. For example, a machine may report high utilization while producing parts that cannot be assembled because the wrong revision was used.

CIM addresses this problem by integrating three dimensions:

• Physical integration: machines, robots, sensors, material-handling equipment, and inspection systems exchange signals and material.
• Information integration: systems share consistent product, process, order, quality, and equipment data.
• Organizational integration: responsibilities, decision rights, procedures, and performance objectives are aligned.

A CIM system is therefore socio-technical. Software and networks are necessary but insufficient. Standard operating procedures, training, data ownership, and change control determine whether integration produces reliable results.

1.2 Digital manufacturing concepts

A digital model is a structured representation of a product, process, resource, or organization. A CAD model represents product geometry and engineering attributes. A process model represents operations, sequences, tools, parameters, and labor requirements.

A digital thread is the governed chain of linked information connecting life-cycle stages. It enables a user to trace a production result back to a material lot, machine, program revision, process plan, operator authorization, and inspection record.

A digital twin is a dynamic digital representation of a physical object, process, or system that is connected to relevant operational data. A static CAD model is not automatically a digital twin. A useful twin requires a defined purpose, a correspondence between physical and digital entities, suitable data, and a method for updating or interpreting the representation.

1.3 CIM as an integration architecture

CIM normally includes design engineering, process planning, production planning, shop-floor control, quality, maintenance, logistics, purchasing, sales, and service. Integration can be represented as a loop rather than a one-way pipeline: operational feedback affects planning; quality findings affect design and process engineering; maintenance data affects capacity; and customer information affects future product configurations.

Figure: A simplified CIM architecture showing downward authorization and upward operational feedback.

2. Roles of Core Manufacturing Systems

2.1 CAD and engineering definition

CAD creates the product definition: geometry, dimensions, tolerances, materials, configurations, annotations, and related documents. Modern model-based definition may place product manufacturing information directly in the three-dimensional model. CAD data should be revision-controlled and associated with approved configurations. Geometry alone is not enough to manufacture a product; functional requirements, tolerances, inspection criteria, and approved materials must also be available.

2.2 CAM and numerical control

CAM converts product and process information into manufacturing operations and machine instructions. Typical functions include tool-path generation, tool selection, cutting-condition definition, collision checking, simulation, post-processing, and setup documentation. CAM output must be validated because a mathematically valid tool path can still be unsafe or unsuitable due to fixture errors, tool wear, machine limits, or incorrect post-processor settings.

2.3 CAPP

CAPP determines how a product will be made. It selects operation sequences, machines, fixtures, tools, inspection steps, labor skills, and expected times. Variant CAPP retrieves and modifies an existing standard plan, whereas generative CAPP derives a plan from product features, rules, resources, and constraints. In practice, many systems combine both methods.

CAPP is the bridge between engineering intent and shop-floor execution. It should identify the required resources and quality controls and should produce a process plan that is consistent with the approved product revision. A process plan is not merely a list of operations; it is a controlled manufacturing instruction set.

2.4 PLM

PLM governs product information and decisions throughout the life cycle. It commonly manages requirements, CAD files, bills of material (BOMs), configurations, revisions, engineering change orders, approvals, supplier documents, and service information. PLM establishes which product definition is valid and who may approve changes.

A key distinction is that PLM is product-centric, while MES is execution-centric. PLM asks, “What is the approved product and process definition?” MES asks, “What is being produced now, by which resource, using which instructions, and with what result?”

2.5 MES

MES coordinates and records production operations between enterprise planning and automation. Functions may include dispatching, detailed scheduling, work instructions, resource status, labor reporting, quality management, nonconformance management, maintenance coordination, electronic batch or device records, and genealogy.

MES is not necessarily a replacement for ERP or a supervisory control system. Its value lies in real-time or near-real-time execution context. It links a production order to operations, materials, personnel, equipment, measurements, deviations, and output.

2.6 ERP

ERP coordinates enterprise resources and transactions. Manufacturing-related ERP functions include sales and operations planning, master production scheduling, material requirements planning, purchasing, inventory valuation, costing, order management, finance, and human-resource records. ERP normally works at a longer planning horizon than MES, although the boundary depends on the implementation.

The distinction between ERP and MES is not absolute. Vendors may offer overlapping capabilities. The correct architecture is determined by business requirements, transaction ownership, timing, data granularity, and operational risk rather than by product labels.

3. ISA-95 and the Manufacturing Hierarchy

ISA-95, also published as IEC 62264 in related international standardization, provides a framework for integrating enterprise and control activities. Its commonly used hierarchy contains four levels:

ISA-95 is a model for organizing responsibilities and information, not a mandatory software product architecture. It helps define boundaries and interfaces. For example, ERP may send a production schedule or order to MES; MES may send a detailed dispatch list and parameters to control systems; control systems return actual production, status, and alarms.

Important ISA-95 information categories include personnel, equipment, material, process segments, production capability, production schedules, production performance, and definitions of products and operations. Consistent use of these concepts reduces ambiguity. A machine should not be known as “Machine 7” in one system and “MC-007” in another without a controlled cross-reference.

Vertical and horizontal integration

Vertical integration connects levels of the hierarchy: business planning to operations management, operations management to supervisory control, and control to physical processes. Its purpose is to maintain alignment between business requirements and actual production.

Horizontal integration connects processes or organizations at the same or adjacent functional level. Examples include CAD-to-CAM integration, quality-to-MES integration, maintenance-to-production integration, supplier-to-manufacturer information exchange, and customer-to-service integration. Horizontal integration supports continuity across the value chain.

Both types are necessary. Vertical integration without horizontal integration may produce a well-connected factory that still has disconnected engineering, quality, or suppliers. Horizontal integration without vertical control may produce many interfaces without coherent business priorities.

4. Designing a Digital Manufacturing Workflow

A robust workflow begins with controlled master data. Typical objects include item numbers, revisions, units of measure, BOMs, routings, process segments, work centers, tools, fixtures, inspection characteristics, personnel qualifications, and equipment identifiers.

A representative workflow is:

1. Customer requirements are captured and translated into product requirements.
2. Engineering creates and approves a CAD model, drawings, specifications, and product structure in PLM.
3. CAPP derives or retrieves an approved routing, operation sequence, resources, and inspection plan.
4. CAM generates, simulates, verifies, and releases machine programs linked to the correct revision and setup.
5. ERP plans demand, materials, capacity, purchasing, and production orders.
6. MES receives the authorized order and dispatches operations according to current resources and priorities.
7. Operators receive controlled instructions and identify materials, tools, equipment, and personnel.
8. Machines and inspection systems produce actual values, status information, and quality results.
9. MES records genealogy, labor, quantities, scrap, deviations, and approvals.
10. ERP receives confirmed quantities, material consumption, inventory movements, and costs; PLM receives relevant lessons and approved changes.

The workflow must define hand-offs. At each hand-off, the organization should specify the sending system, receiving system, data object, trigger, validation rule, error response, and accountable owner. A connection that transfers data but does not define error handling is not a reliable integration.

Interoperability and standards

Integration may use application programming interfaces (APIs), message brokers, databases, file exchange, industrial protocols, and standardized information models. STEP supports product-model exchange in relevant applications; MTConnect and OPC UA are examples of industrial data communication technologies, although their suitability depends on the equipment and required semantics. A protocol transports information, but it does not automatically provide correct meaning. Semantic mapping, units, timestamps, identity, and version control remain essential.

A practical architecture often uses an integration layer or manufacturing service bus. It can transform messages, validate schemas, manage routing, monitor failures, and prevent every application from requiring a separate point-to-point connection. Excessive point-to-point interfaces create high maintenance costs and obscure ownership.

5. Traceability, Genealogy, and Performance

Traceability is the ability to follow the history, application, location, or use of an object through specified records. In manufacturing, traceability may include raw-material heat or lot, supplier, machine, tool, program revision, operator, environmental condition, inspection result, deviation, rework, and shipment.

Genealogy records relationships between inputs and outputs. For example, a finished assembly may be linked to its subassemblies, component serial numbers, material lots, software version, test results, and production order. Genealogy is especially important in aerospace, medical devices, automotive, food, energy, and regulated products.

Traceability must be designed according to risk. Recording every possible data point can increase storage, cybersecurity exposure, and false confidence. The organization should define:

• the object being identified;
• the required granularity, such as lot, batch, unit, or serial number;
• the event and time source;
• the responsible system and data owner;
• retention, access, and audit requirements;
• rules for correction without destroying the original record.

Useful measures include first-pass yield, schedule adherence, scrap rate, cycle-time variance, unplanned downtime, and overall equipment effectiveness (OEE). A common OEE decomposition is:

Availability is the ratio of planned production time during which equipment is operating to planned production time. Performance compares actual output rate with the ideal output rate while operating. Quality is the ratio of good units to total units produced. Each factor is dimensionless and normally expressed as a percentage. The multiplication is meaningful only when the definitions and time boundaries are consistent; organizations should not compare OEE values produced by incompatible rules.

For example, if availability is 0.90, performance is 0.80, and quality is 0.95, then:

This result does not identify the cause of loss by itself. It indicates that improvement should investigate all three components.

6. Digital Twins, Simulation, and AI

Simulation can evaluate layouts, schedules, robot reach, machining time, material flow, and capacity before physical implementation. Discrete-event simulation is useful for queues, routing, failures, and variable processing times. Its results depend on assumptions, data quality, and model validation.

A digital twin may support condition monitoring, what-if analysis, predictive maintenance, energy optimization, or remote service. The twin should have a defined fidelity: a planning twin may need approximate capacities, whereas a control-support twin may require time-synchronized equipment data.

Artificial intelligence can assist visual inspection, anomaly detection, demand forecasting, tool-wear estimation, scheduling, and knowledge retrieval. However, AI outputs should be evaluated for false positives, false negatives, drift, explainability, cybersecurity, and responsibility for decisions. In safety-critical production, an AI recommendation should not bypass validated controls or authorized human decision-making. Training data may also encode historical bias, such as systematically accepting defects from one product family and rejecting similar defects from another.

7. Implementation Strategy and Governance

A successful integration program normally follows a staged approach:

1. Define business outcomes: quality, delivery, flexibility, compliance, cost, or safety.
2. Map current processes: identify delays, duplicate entry, uncontrolled spreadsheets, and data ownership.
3. Create a capability and information model: define entities, identifiers, states, revisions, and transactions.
4. Prioritize a pilot: select a product family or value stream with measurable problems and manageable scope.
5. Cleanse master data: resolve duplicate part numbers, inconsistent units, obsolete routings, and ambiguous equipment names.
6. Build and test interfaces: include normal, delayed, duplicate, incomplete, and rejected messages.
7. Validate in realistic operation: test revision changes, machine downtime, rework, scrap, and quality holds.
8. Train users and establish support: include operators, engineers, planners, quality staff, IT, and maintenance.
9. Scale through standards: reuse information models, interface patterns, and governance rules.

Governance should assign ownership. Engineering should generally own product definitions; manufacturing engineering should own process definitions; operations should own execution records; quality should own inspection and nonconformance rules; IT or enterprise architecture should govern platforms and integration; cybersecurity should govern access, segmentation, monitoring, and response.

Cybersecurity and operational resilience

Connecting shop-floor systems increases the attack surface. Essential practices include asset inventory, network segmentation, least-privilege access, multifactor authentication where feasible, secure remote access, patch and vulnerability management, backups, logging, incident response, and recovery testing. Safety and availability require special care: a security update that interrupts a validated control system can create operational risk, while an unpatched system may create cyber risk.

Ethical, legal, and regulatory considerations

Digital records affect worker privacy, intellectual property, product liability, and regulatory compliance. Monitoring operator actions should have a legitimate purpose, transparent policy, proportionality, and access control. Records used for regulated products must be attributable, legible, contemporaneous, original or appropriately controlled, and accurate. Automated decisions should preserve auditability and provide a method for correction or appeal where people are materially affected.

8. Applied Case: Illustrative Aerospace Component Factory

Background

An illustrative aerospace component factory machines serialized aluminum housings on several CNC cells. Engineering stores CAD and drawings in PLM, ERP manages orders and materials, and operators previously selected programs from local network folders. Inspection results were entered manually into spreadsheets.

Problem

The factory experienced three related incidents: an obsolete CNC program was used, a material certificate could not be linked quickly to finished serial numbers, and production status in ERP lagged behind the physical shop floor. Investigation showed inconsistent part identifiers, weak revision control, and no electronic confirmation that the correct program had been selected.

Analysis

The problem was not merely a machine-control failure. It involved horizontal gaps between PLM, CAM, quality, and production, together with vertical gaps between ERP, MES, and the CNC cells. The organization also lacked a formal data owner for program release and a defined response to interface failures.

Approach and solution

The factory introduced a controlled workflow. PLM became the authority for approved product revisions. CAM programs were linked to the product and process-plan revision and were released only after simulation and manufacturing approval. MES received the production order from ERP and dispatched operations only when the required material lot, tool status, and program revision were valid.

At the machine, the operator scanned the work order and material identifier. MES verified the authorized program and recorded machine, operator, fixture, tool data, start time, completion time, quantity, scrap reason, and inspection results. A quality hold prevented shipment when required measurements were absent or outside approved limits. ERP received confirmed production and material consumption, while genealogy linked each housing to its material certificate and inspection record.

Outcome

The revised system reduced manual transcription, made revision status visible, and shortened investigation of a suspected material issue from several days to minutes. The factory also discovered that some delays were caused by tool-availability rules rather than software. This illustrates an important principle: integration exposes process weaknesses; it does not automatically eliminate them.

Learning points

• A single authoritative source is needed for each critical data object.
• Traceability must be designed before interfaces are built.
• Revision control must extend from CAD through CAM and MES to the machine.
• Exception handling is as important as the normal workflow.
• Benefits should be measured through operational outcomes, not the number of connected systems.

9. Challenges and Limitations

Integration projects may fail because of poor data quality, unclear ownership, obsolete equipment, incompatible vendor models, excessive customization, inadequate network capacity, or resistance to changed work practices. A large platform does not guarantee interoperability. Nor does real-time data guarantee correct decisions; inaccurate sensors and incorrect context can produce faster errors.

Legacy equipment may require gateways, manual verification, or phased replacement. Cloud services can improve scalability and collaboration but require careful treatment of latency, availability, data sovereignty, and intellectual property. Digital twins may become expensive demonstrations if their purpose and decisions are not defined. Finally, automation can displace tasks while increasing demand for systems thinking, troubleshooting, data literacy, and cybersecurity skills.

Best-Practice Checklist

• Establish controlled identifiers, units, revisions, and ownership.
• Use ISA-95 concepts to clarify responsibilities and interfaces.
• Separate product authority, process authority, execution authority, and financial authority.
• Validate both data semantics and communication reliability.
• Design genealogy and retention rules according to product and regulatory risk.
• Test exceptions, rework, downtime, duplicate messages, and revision changes.
• Protect operational technology using layered cybersecurity.
• Measure quality, delivery, flexibility, and total cost rather than connectivity alone.
• Involve operators and maintenance personnel in design and validation.
• Scale proven patterns instead of copying uncontrolled point-to-point links.

Chapter Summary

Computer-integrated and digital manufacturing connect product definition, process planning, production execution, enterprise planning, and physical automation through governed information flows. CAD defines the product, CAM develops machine instructions, CAPP defines manufacturing methods, PLM controls product knowledge and change, MES coordinates and records execution, and ERP plans enterprise resources and transactions. ISA-95 provides a useful hierarchy for distinguishing business planning, manufacturing operations, supervisory control, sensing, and the physical process.

Vertical integration aligns business plans with shop-floor activity; horizontal integration links engineering, production, quality, maintenance, suppliers, and customers. Digital threads and genealogy provide traceability, while simulation, digital twins, and AI can improve decision-making when their assumptions, data quality, and risks are controlled. Successful implementation depends on master-data governance, interoperable interfaces, cybersecurity, validated workflows, user participation, and measurable business outcomes. Integration is therefore both a technical architecture and an organizational transformation.

Key Terms

CIM: Coordinated use of computers, communications, automation, and information systems to integrate manufacturing activities.

Digital manufacturing: Manufacturing supported by connected digital models, data, simulation, analytics, and automated information exchange.

CAD: Computer-aided design used to create and manage product geometry and engineering definition.

CAM: Computer-aided manufacturing used to generate and validate manufacturing operations and machine instructions.

CAPP: Computer-aided process planning used to define operation sequences, resources, and methods.

PLM: Product life-cycle management governing product information, configurations, revisions, and changes.

MES: Manufacturing execution system coordinating and recording production operations.

ERP: Enterprise resource planning system coordinating business resources, materials, orders, and financial transactions.

ISA-95: A framework for organizing enterprise-control integration and manufacturing information exchanges.

Vertical integration: Connection between hierarchical levels, from enterprise planning to physical control.

Horizontal integration: Connection among functions, processes, or organizations across a value stream.

Digital thread: Governed chain of linked information across the product life cycle.

Digital twin: A connected digital representation of a physical object, process, or system.

Genealogy: Recorded relationship between production inputs, operations, and outputs.

Traceability: Ability to follow the history, application, location, or use of an object.

Master data: Controlled, relatively stable data such as items, resources, revisions, and units.

Revision control: Governance ensuring that authorized versions are identified and used.

Interoperability: Ability of systems to exchange and correctly interpret information.

OEE: Product of availability, performance, and quality measures for equipment effectiveness.

Operational technology: Hardware and software that monitors or controls physical processes.

Review Questions

Short Answer Questions

1. What is the difference between PLM and MES?
2. Define vertical integration and give one manufacturing example.
3. What information does CAPP provide to production?
4. Why is genealogy important in regulated manufacturing?
5. State the three factors in the common OEE decomposition.

Long Answer Questions

1. Explain how CAD, CAM, CAPP, PLM, MES, and ERP can form one controlled digital workflow.
2. Describe the ISA-95 hierarchy and discuss the purpose of boundaries between its levels.
3. Compare product-centric PLM, execution-centric MES, and resource-centric ERP.
4. Discuss the technical and organizational requirements for reliable manufacturing traceability.
5. Evaluate the benefits and limitations of using digital twins and AI in production.

Analytical / Application Questions

1. A factory has accurate ERP inventory records but frequent shop-floor shortages. Analyze likely vertical and horizontal integration failures and propose a staged remedy.
2. Design a genealogy scheme for a serialized medical-device assembly. Specify identifiers, events, records, retention considerations, and quality holds.
3. A production cell reports availability of 0.92, performance of 0.85, and quality of 0.97. Calculate OEE and recommend which additional data are needed before selecting an improvement project.

Multiple Choice Questions with Answers

Question 1

Which system is primarily responsible for controlling approved product revisions and engineering changes?

A. PLM

B. MES

C. SCADA

D. Payroll system

Correct Answer: A. PLM

Question 2

In the ISA-95 hierarchy, manufacturing operations management is commonly associated with which level?

A. Level 0

B. Level 1

C. Level 3

D. Level 4

Correct Answer: C. Level 3

Question 3

Which statement best describes horizontal integration?

A. Connecting enterprise planning to physical sensors

B. Connecting functions such as engineering, quality, and maintenance across a value stream

C. Replacing all manual work with robots

D. Increasing the speed of one isolated machine

Correct Answer: B. Connecting functions such as engineering, quality, and maintenance across a value stream

Question 4

If availability is 0.90, performance is 0.80, and quality is 0.95, what is OEE?

A. 0.684 or 68.4%

B. 0.855 or 85.5%

C. 2.65 or 265%

D. 0.150 or 15.0%

Correct Answer: A. 0.684 or 68.4%

Question 5

What is the principal purpose of CAPP?

A. To process payroll

B. To define manufacturing operations, sequences, and resources

C. To supervise electrical signals only

D. To store customer invoices only

Correct Answer: B. To define manufacturing operations, sequences, and resources

Question 6

Which record most directly supports tracing a finished serial number to its material lot and inspection results?

A. Genealogy record

B. Marketing forecast

C. Office seating plan

D. Uncontrolled personal notebook

Correct Answer: A. Genealogy record

Question 7

Why should an integration project test duplicate and incomplete messages?

A. To make interfaces visually attractive

B. To verify behavior under realistic exception conditions

C. To eliminate the need for data ownership

D. To avoid using master data

Correct Answer: B. To verify behavior under realistic exception conditions

Question 8

Which system generally coordinates detailed production execution and records operator, equipment, and quality results?

A. MES

B. CAD

C. ERP financial ledger only

D. Email client

Correct Answer: A. MES

Question 9

A digital twin differs from a static CAD model primarily because it is:

A. Always a physical prototype

B. Connected to relevant physical or operational data for a defined purpose

C. Limited to financial information

D. Unable to represent processes

Correct Answer: B. Connected to relevant physical or operational data for a defined purpose

Question 10

Which practice is most appropriate for protecting connected operational technology?

A. Give every user administrator privileges

B. Remove all audit logs

C. Use segmentation, least privilege, backups, and incident-response planning

D. Permit unrestricted remote access

Correct Answer: C. Use segmentation, least privilege, backups, and incident-response planning