Manufacturing Economics, Sustainability and Decision-Making

Introduction

Manufacturing technology is not selected solely because a machine can produce a part. A technically feasible process may be economically unattractive, environmentally burdensome, difficult to scale, or exposed to unacceptable supply-chain risk. CNC machining and additive manufacturing (AM) therefore have to be evaluated as complete production systems rather than as isolated machines.

CNC processes remove material from a workpiece through cutting, while AM processes build parts layer by layer from digital data. CNC is often associated with high dimensional accuracy, established quality systems and predictable production rates. AM is often associated with geometric freedom, rapid design iteration and reduced tooling requirements. Neither statement is universally true: economics depend on geometry, material, quantity, machine loading, post-processing, quality requirements, energy sources and supply-chain conditions.

This chapter develops a structured basis for comparing CNC and AM. It introduces cost models, lead-time and throughput measures, break-even analysis, sustainability metrics, life-cycle thinking, and make-or-buy decision methods. The emphasis is on engineering decisions supported by transparent assumptions rather than on a single universally superior technology.

Learning Objectives

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

1. Construct and interpret unit-cost models for CNC and AM.
2. Compare lead time, throughput, quality, scalability and flexibility.
3. Calculate a break-even production quantity between competing processes.
4. Assess energy, material waste and life-cycle impacts.
5. Incorporate supply-chain, quality and capacity risks into technology selection.
6. Apply weighted decision methods to make-or-buy and process-selection problems.
7. Explain the limitations and ethical implications of economic and environmental claims.

1. The Manufacturing Decision as a System Problem

A manufacturing decision links product design, process planning, equipment, labor, materials, inspection, logistics and end-of-life treatment. The relevant question is not “What is the machine-hour rate?” but “What is the delivered cost and performance of an acceptable part over its required life?”

A useful decision boundary includes:

• Technical feasibility: Can the process produce the required geometry, material condition and tolerances?
• Economic viability: Is the total cost acceptable at the expected quantity and utilization?
• Operational capability: Can the process meet demand, takt time and delivery reliability?
• Environmental performance: What are the material, energy, emissions and waste consequences?
• Risk exposure: How sensitive is the result to suppliers, machine failure, qualification and demand uncertainty?

A process that fails a mandatory requirement should normally be rejected before weighted scoring. For example, a low-cost process that cannot meet a critical fatigue requirement is not rescued by a favorable sustainability score.

2. Cost Structure of CNC and Additive Manufacturing

2.1 Total-cost logic

For a production lot, total cost can be represented as:

where C_total is total production cost, C_fixed is quantity-independent or lot-level cost, Q is the number of acceptable parts, and C_variable is variable cost per acceptable part. This is a first-order model. It assumes approximately linear variable cost and does not capture capacity jumps, quantity discounts or learning effects.

The unit cost is:

The fixed component becomes less important as quantity increases. In real factories, some costs are semi-variable: an additional machine, operator shift or inspection cell may be required when demand crosses a capacity threshold.

2.2 CNC cost model

A practical CNC estimate may be written as:

Here, setup, programming, running and inspection times are in hours; the corresponding rates are in currency per hour; C_tool is tooling and consumables cost; C_material is material cost; C_external is outsourced or additional cost; and y is the yield, expressed as acceptable parts per attempted parts. If yield is 0.98, the numerator must be divided by 0.98 to obtain cost per good part.

CNC material cost should account for purchased stock, not only the mass of the finished part. If a billet costs p_m currency per kilogram and has mass m_stock, then:

The economic value of chips may be recovered through recycling, but recycling credit should be included only when it is actually obtained. Material-removal time, tool wear, workholding, coolant management and inspection can dominate the result for complex parts.

2.3 AM cost model

An AM estimate may be structured as:

The variables represent build, setup, post-processing and inspection times; machine, post-processing and quality rates; feedstock and support costs; energy cost; other external costs; and yield. AM setup may include orientation, support generation, slicing, file verification and machine preparation. Post-processing may include powder removal, support removal, stress relief, heat treatment, machining, surface finishing and qualification.

A major AM issue is batch economics. Several parts may be nested in one build, so machine time is allocated across the number of acceptable parts. If a build takes 20 h and contains 10 equivalent parts, the build-time allocation is 2 h per part, subject to the effects of packing density, thermal interaction and post-processing.

2.4 Rate construction

A machine-hour rate can be estimated from annualized cost:

where H_productive is productive machine time per year. The denominator must not be confused with calendar hours. If a machine is available for 4,000 h but is productively loaded for only 2,000 h, using 4,000 h will understate the rate.

Depreciation is an accounting allocation and not necessarily a cash payment in the current period. For investment decisions, cash flow, financing, tax treatment and salvage value may also be relevant. A rigorous investment analysis may use net present value rather than a simple cost rate.

3. Worked Break-Even Example

Consider an illustrative component that can be made either by CNC or by polymer AM. Assume all values are in currency units and that both methods meet the required performance.

For CNC:

• fixed programming and setup cost: 1,800;
• variable cost per good part: 42.

For AM:

• fixed preparation and qualification cost: 600;
• variable cost per good part: 78.

The total costs are:

At break-even, the two costs are equal:

Therefore:

Since production quantity is an integer, AM is less costly through 33 parts, while CNC becomes less costly from 34 parts onward, under the stated assumptions. At 20 parts, CNC costs 2,640 and AM costs 2,160. At 50 parts, CNC costs 3,900 and AM costs 4,500.

This result is not a universal rule. If AM has a shorter lead time, avoids purchased tooling, or permits a lighter redesign, its total economic value may exceed its direct manufacturing cost. Conversely, if AM requires extensive machining and inspection, its variable cost may be higher than initially estimated.

The general break-even expression for two linear alternatives is:

where F₁ and F₂ are fixed costs and v₁ and v₂ are variable costs. The result is meaningful only when the denominator is nonzero and both methods are technically feasible over the quantity range. If one method has both higher fixed and higher variable cost, it is economically dominated unless it provides a non-cost advantage.

4. Lead Time, Throughput and Scalability

4.1 Lead time

Lead time is the elapsed time from an authorized order or released design to delivery of acceptable parts. It includes waiting, queueing, setup, production, inspection, rework and transport. A process time calculated from machine specifications is not necessarily customer lead time.

AM can reduce lead time when conventional tooling would require weeks. CNC can be faster for simple parts when fixtures, programs and stock are already available. Design iteration is often rapid with AM, but certification and post-processing may extend delivery.

A simple production lead-time model is:

Each term must be measured or estimated consistently. Queue time is particularly important in highly utilized facilities.

4.2 Throughput and capacity

Throughput is the rate of acceptable output. If a resource produces N_good acceptable parts in T hours:

where throughput TH has units of parts per hour. If a CNC cycle produces one part every 12 min with 95% yield, ideal throughput is 5 parts/h and effective throughput is:

For AM, throughput depends on build quantity and build duration:

where n_good is the number of acceptable parts per build. AM can be highly productive for geometrically complex parts when many parts fit in one build, but its output may be constrained by build volume, powder handling, post-processing or inspection.

4.3 Scalability

Scalability concerns whether output can increase without disproportionate cost, delay or quality loss. CNC scaling may involve additional machines, fixtures, operators and cutting tools. AM scaling may involve more printers, larger machines, multiple shifts, distributed production or automated post-processing. Adding printers does not automatically multiply output if powder qualification, inspection or finishing is the bottleneck.

The bottleneck principle is central: system throughput is governed by the constrained resource. A high-speed printer feeding a slow heat-treatment furnace does not increase finished-part output beyond furnace capacity.

5. Quality, Capability and Compliance

Cost comparison is valid only after quality requirements are made explicit. Relevant dimensions include:

• dimensional accuracy and repeatability;
• surface roughness and waviness;
• material properties and anisotropy;
• porosity, inclusions and residual stress;
• traceability and process monitoring;
• inspection method and measurement uncertainty;
• qualification requirements and documentation.

CNC commonly offers mature process capability, particularly for prismatic and rotational features. However, thin walls, deep cavities, tool deflection and chatter can reduce capability. AM may produce complex internal geometry, but layer orientation, thermal history, support strategy and powder condition can cause anisotropy or defects.

Process capability is often summarized using:

Here, USL and LSL are upper and lower specification limits, μ is the process mean and σ is the standard deviation. These indices assume a stable process and are meaningful only when the data and distributional assumptions are appropriate. C_p measures potential capability relative to spread; C_pk also considers centering. They should not replace engineering judgment or acceptance testing.

Quality cost includes prevention, appraisal, internal failure and external failure. A cheap process with high rework or warranty exposure may have a higher expected cost than a more expensive controlled process.

6. Sustainability and Life-Cycle Thinking

6.1 Material efficiency

A simple material-utilization ratio is:

where m_product is the mass incorporated in the accepted product and m_input is material purchased or consumed. The ratio must define system boundaries. CNC may have low utilization when a large billet is machined into a small part, although chips can be recycled. AM can have high geometric material efficiency, but supports, failed builds, purge material, sieving losses and unused powder may reduce actual efficiency.

6.2 Energy and carbon

Process energy can be estimated as:

where P_i is the power of equipment i in kilowatts and t_i is operating time in hours, giving energy in kilowatt-hours. Ancillary equipment such as chillers, dust extraction, compressors, ovens and ventilation must be included when the boundary requires it.

If electricity has an emissions factor EF in kilograms of CO₂-equivalent per kilowatt-hour:

This calculation is valid only when the emissions factor corresponds to the location, time period and accounting method. Scope definitions, renewable electricity claims and embodied emissions in materials must be stated.

A life-cycle assessment (LCA) normally considers raw-material production, manufacturing, transport, use, maintenance and end of life. Manufacturing-stage energy alone cannot determine total environmental superiority. A lighter AM part may require more energy to produce but save fuel or electricity during years of use. Conversely, a complex AM process may create burdens that outweigh material savings.

6.3 Waste hierarchy and circularity

Sustainability decisions should distinguish reduction, reuse, recycling and disposal. Metal chips and polymer scrap may have different recovery values and contamination risks. Repeated powder recycling may alter particle-size distribution or chemistry and therefore require qualification. Recycling is not automatically closed-loop: collection, sorting, remelting and transport consume energy.

A responsible comparison reports functional performance, such as emissions per acceptable part or per service unit, rather than only emissions per kilogram of material. The functional unit might be “one component meeting a specified load, life and tolerance for 10 years.”

7. Supply-Chain Risk and Resilience

Technology selection is also a risk decision. Important risks include:

• dependence on a single machine, software vendor or material supplier;
• long lead times for specialty powders, billets or cutting tools;
• cyberattack or unauthorized modification of digital build files;
• loss of process qualification when production is transferred;
• operator skill shortages;
• uncertain demand and obsolete inventory;
• energy-price volatility and utility interruption;
• transport disruption and geopolitical restrictions.

Risk can be represented using expected loss:

where p_j is the probability of risk event j and I_j is its financial or operational impact. This simple model assumes discrete events and reliable probability estimates. For rare, severe events, scenario analysis and stress testing are more informative than a single expected value.

Resilience measures include qualified alternative suppliers, safety stock for critical feedstock, duplicate machines, standardized digital workflows, preventive maintenance, local production capability and documented transfer procedures. AM can reduce physical inventory by producing near the point of use, but it can increase dependence on digital files, specialized materials and qualified machines.

8. Make-or-Buy and Technology-Selection Framework

A disciplined decision can proceed through the following stages.

Stage 1: Define the requirement

Specify quantity, demand pattern, delivery date, geometry, material, tolerances, surface requirements, mechanical performance, regulatory obligations and expected service life. Separate mandatory constraints from preferences.

Stage 2: Screen technical feasibility

Reject alternatives that cannot meet the specification or cannot be qualified within the project schedule. Include post-processing and inspection, not just the primary operation.

Stage 3: Estimate total cost

Use consistent boundaries and include setup, labor, machine time, material, tooling, post-processing, inspection, logistics, scrap, maintenance and risk allowances. Present ranges rather than false precision when data are uncertain.

Stage 4: Evaluate capacity and lead time

Model queueing, batch size, available hours, bottlenecks and ramp-up. Determine whether the proposed process can meet both average demand and peak demand.

Stage 5: Evaluate sustainability

Use a functional unit and state the life-cycle boundary. Report material, energy, emissions and waste assumptions. Consider use-phase effects and end-of-life recovery.

Stage 6: Evaluate risk and strategic fit

Assess supplier concentration, intellectual-property protection, workforce competence, cybersecurity, qualification burden and reversibility of the decision.

Stage 7: Select and monitor

Use sensitivity analysis and define performance indicators such as cost per good part, on-time delivery, first-pass yield, energy per part, material utilization and changeover time.

A weighted scoring model may be used after mandatory screening:

where w_k is the weight assigned to criterion k, x_k is the normalized score, and ∑w_k = 1. Scores must be defined consistently; a high score should always mean better performance. Weight selection should be documented and tested through sensitivity analysis.

9. Illustrative Case: Low-Volume Aerospace Bracket

Background

This illustrative case concerns a small aerospace bracket required in a quantity of 120 units per year. The bracket contains a curved load path, several mounting holes and a pocket intended to reduce mass. The design authority requires traceability, dimensional inspection and a defined fatigue performance.

Problem

The existing bracket is machined from aluminum billet. It has a long procurement lead time because the billet is a special size, and material utilization is low. An AM redesign could reduce mass and eliminate some inaccessible machining operations, but it would require qualification and heat treatment.

Analysis

The engineering team compares three options: continue CNC production, outsource metal AM, or purchase an AM machine and produce internally. The team first identifies non-negotiable requirements: fatigue performance, hole accuracy, corrosion protection, inspection records and annual demand.

The CNC option has mature qualification and a predictable inspection route. Its disadvantages are high billet waste, multiple setups and a six-week material procurement lead time. The outsourced AM option has a shorter physical production time but introduces supplier capacity and qualification risk. Internal AM offers control but requires capital investment, operator training and a new quality system.

A cost model shows that at 120 units per year, CNC has the lowest direct unit cost because the fixed qualification cost is already absorbed and the machining cell is available. AM becomes attractive only if the redesign reduces assembly count, lowers mass sufficiently to produce use-phase savings, or supports several additional parts on the same qualified platform.

The team therefore chooses CNC for the immediate production requirement and launches an AM demonstrator project. The AM project is not judged only by prototype cost; it must demonstrate fatigue capability, stable powder reuse, repeatable heat treatment, inspection time and an economically credible batch size.

Outcome and learning points

The decision illustrates a staged strategy: use the qualified CNC route for supply continuity while developing AM capability under controlled conditions. The best present decision and the best long-term technology need not be identical. The case also shows why a lighter part, lower material waste or shorter build time cannot independently justify AM without considering qualification, post-processing, life-cycle effects and demand.

10. Contemporary Developments and the Role of Digital Tools

Digital manufacturing systems increasingly connect CAD, process planning, machine monitoring, inspection and production records. AI can assist with tool-wear prediction, build-defect detection, nesting, scheduling and demand forecasting. These tools are decision aids, not substitutes for engineering validation.

AI-based recommendations require representative data, controlled versioning, cybersecurity and clear accountability. A model trained on one alloy, machine or parameter set may fail when transferred. False negatives in defect detection can create safety risk, while false positives can increase scrap and cost. Human review, validation datasets and audit trails are essential.

Digital inventories also create governance concerns. A production file may encode intellectual property and safety-critical parameters. Access control, encryption, authentication, revision management and secure transfer are therefore part of manufacturing economics and risk management.

11. Limitations and Best Practices

Economic models are simplifications. Common errors include omitting setup and inspection, assuming full machine utilization, ignoring rejected parts, comparing different functional performance, treating accounting depreciation as cash cost, and reporting energy without ancillary equipment.

Best practice includes:

1. Define the functional unit and system boundary before collecting data.
2. Separate mandatory technical constraints from weighted preferences.
3. Use good-part cost, not nominal process cost.
4. Include queueing, post-processing and inspection in lead-time estimates.
5. Use sensitivity analysis for quantity, labor rate, energy price, yield and machine utilization.
6. Validate assumptions with shop-floor measurements and supplier quotations.
7. Report uncertainty ranges and the date and location of emissions factors.
8. Reassess decisions when geometry, demand, material or qualification requirements change.
9. Protect digital manufacturing data and maintain revision traceability.
10. Avoid environmental claims that cannot be supported by a defined life-cycle method.

Chapter Summary

CNC and AM should be compared as complete manufacturing systems. Total cost includes fixed and variable elements, material, labor, machine time, tooling, post-processing, inspection, yield and risk. Break-even analysis identifies quantity regions in which one process has a direct cost advantage, but it does not replace technical, quality or strategic analysis.

Lead time includes queueing and downstream operations, while throughput is governed by effective good-part output and the system bottleneck. CNC often benefits from mature capability and predictable production; AM often benefits from design freedom, low tooling dependence and efficient production of complex or customized parts. Both technologies have limitations.

Sustainability requires life-cycle thinking. Material utilization, process energy, emissions, waste, use-phase performance and end-of-life treatment must be related to a functional unit. Supply-chain resilience includes supplier diversity, qualification transfer, cybersecurity, digital traceability and capacity continuity.

A robust make-or-buy decision defines requirements, screens feasibility, estimates total cost, evaluates capacity and quality, assesses life-cycle effects, analyzes risk and tests sensitivity. Weighted scoring is useful only after mandatory constraints are satisfied and the scoring assumptions are transparent.

Key Terms

Total cost: The complete cost of producing and delivering acceptable parts within a stated system boundary.

Fixed cost: Cost that is approximately independent of production quantity over a defined range.

Variable cost: Cost that changes approximately with the number of parts produced.

Good-part yield: The fraction of attempted parts that meet requirements.

Break-even quantity: Production quantity at which two alternatives have equal modeled cost.

Lead time: Elapsed time from release or order authorization to delivery of acceptable output.

Throughput: Rate of acceptable output from a process or system.

Bottleneck: Resource that limits the capacity of the whole production system.

Process capability: Ability of a stable process to meet specification limits.

CNC: Computer numerical control manufacturing, generally involving programmed material removal.

Additive manufacturing: Layerwise production of parts from digital data.

Material utilization: Product mass divided by material input mass under a defined boundary.

Life-cycle assessment: Systematic evaluation of environmental impacts across defined life-cycle stages.

Functional unit: Reference quantity used to compare environmental performance consistently.

Emissions factor: Emissions associated with a unit of activity, such as kilograms CO₂-equivalent per kilowatt-hour.

Supply-chain resilience: Ability to maintain or rapidly restore supply after disruption.

Make-or-buy decision: Decision to produce internally or obtain a product or service externally.

Weighted scoring: Decision method combining normalized criterion scores and assigned weights.

Process qualification: Demonstration that a process consistently produces acceptable output.

Digital traceability: Ability to identify and control the history, version and use of digital production information.

Review Questions

Short Answer Questions

1. Distinguish fixed cost from variable cost in a CNC or AM operation.
2. Why must unit cost be based on acceptable parts rather than attempted parts?
3. Define throughput and identify two factors that can reduce effective throughput.
4. Why is process energy alone insufficient for a complete sustainability comparison?
5. State two supply-chain risks associated with digital AM production.

Long Answer Questions

1. Develop a total-cost comparison for CNC and AM, explaining the assumptions required for a break-even calculation.
2. Compare CNC and AM with respect to lead time, throughput, scalability and quality capability.
3. Explain how a life-cycle assessment can change the conclusion reached from a manufacturing-stage cost comparison.
4. Describe a structured make-or-buy decision process for a safety-critical component.
5. Discuss the advantages and limitations of weighted scoring in manufacturing technology selection.

Analytical / Application Questions

1. A CNC route has fixed cost 2,400 and variable cost 30 per part. An AM route has fixed cost 900 and variable cost 65 per part. Calculate the break-even quantity and identify the lower-cost route at 25 and 60 parts. State the assumptions behind your calculation.
2. A printer consumes 6 kW for 18 h and an associated heat-treatment furnace consumes 12 kW for 4 h. Calculate total energy and, using an emissions factor of 0.4 kg CO₂-equivalent/kWh, calculate energy-related emissions. Explain which ancillary loads might still be missing.
3. A company must choose between internal CNC, outsourced AM and internal AM for a low-volume medical device. Construct a decision matrix with at least five criteria, identify mandatory constraints, assign defensible weights and explain how you would test sensitivity to the weights.

Multiple Choice Questions with Answers

Question 1

In a linear two-process cost model, the break-even quantity is determined by which relationship?

A. Fixed-cost difference divided by variable-cost difference

B. Variable-cost difference divided by fixed-cost difference

C. Total cost multiplied by production quantity

D. Machine power divided by material utilization

Correct Answer: A. Fixed-cost difference divided by variable-cost difference

Question 2

A process produces 5 attempted parts per hour with a yield of 0.95. What is its effective throughput?

A. 4.05 parts/h

B. 4.75 parts/h

C. 5.95 parts/h

D. 95 parts/h

Correct Answer: B. 4.75 parts/h

Question 3

Which item is most appropriately included in AM post-processing cost?

A. Only CAD modeling time

B. Support removal and heat treatment

C. Raw billet purchase only

D. Customer use-phase electricity only

Correct Answer: B. Support removal and heat treatment

Question 4

What does the denominator in a machine-hour rate normally represent?

A. Calendar hours regardless of utilization

B. Productive machine hours over the chosen period

C. The number of employees in the factory

D. The number of parts in one build

Correct Answer: B. Productive machine hours over the chosen period

Question 5

Which statement best describes a bottleneck?

A. The least expensive operation

B. The resource limiting system capacity

C. The operation with the highest material utilization

D. Any operation performed manually

Correct Answer: B. The resource limiting system capacity

Question 6

In the material-utilization ratio, the numerator is normally which quantity?

A. Purchased stock mass

B. Scrap mass

C. Accepted product mass

D. Packaging mass

Correct Answer: C. Accepted product mass

Question 7

Why should environmental comparisons use a functional unit?

A. To eliminate the need for emissions data

B. To compare equivalent service or performance

C. To guarantee AM has lower emissions

D. To convert all costs into energy units

Correct Answer: B. To compare equivalent service or performance

Question 8

Which factor is directly considered by Cpk but not fully represented by Cp?

A. Process centering relative to specification limits

B. Machine purchase price

C. Material density

D. Electricity emissions factor

Correct Answer: A. Process centering relative to specification limits

Question 9

A weighted scoring model should generally be applied after which step?

A. Ignoring technical requirements

B. Mandatory feasibility screening

C. Removing all risk criteria

D. Assuming equal performance

Correct Answer: B. Mandatory feasibility screening

Question 10

A printer uses 6 kW for 18 h and a furnace uses 12 kW for 4 h. What is their combined energy consumption?

A. 60 kWh

B. 108 kWh

C. 156 kWh

D. 216 kWh

Correct Answer: C. 156 kWh